blog by Joe Moon

An Amateur's Lament

Wed, 30 Nov 2011 22:42:39 -0800

Amazon’s advertisement for the Kindle Fire opens with a Voltaire quote:

The instruction we find in books is like fire. We fetch it from our neighbors, kindle it at home, communicate it to others, and it becomes property of all.

Visions of the Future

Taken from a certain angle, the metaphor makes sense. There’s a Promethean ideal of content — music, movies, books — no longer constrained by the physical media of the past, and no longer subject to the old Olympian gate-keepers: record labels, Hollywood, and publishers. The promise of these devices is accessibility of content to all. The killer feature of each one is the price, which is aggressive to begin with and, if the past is any indicator, subject to relentless and continuous cuts.

Apple has a similar vision of the future but approaches from the opposite direction. It’s more an offer of luxury to the masses, a democratization of high-end design. But the result is basically the same. The emphasis is on the device itself, but it’s a vision of technological access made available to everyone.

Both utopian visions, though, hide the threat of dystopia not far underneath the surface, as utopian visions often do. In the case of Apple’s “post-PC” devices, it’s a subtle one, hidden in the very experience that makes the devices attractive.

Production and Consumption

The iPad is physically not much more than a touchscreen, which is primarily just a screen with the addition of an input system. While the direct-manipulation paradigm is indisputably more intuitive, it’s also less robust. You simply can’t get the precision of a mouse, or the tactile feedback of a physical keyboard. It’s a user-interface optimized for consumption and shallow interactions like navigation. Which isn’t to say that it’s impossible to create things on an iPad; there are plenty of examples to the contrary. But its primary purpose is for output, not input, unlike the general-purpose PC. The iPad implicitly privileges consumption over production. So insofar as these post-PC devices will replace PCs, the market- and mindshare of consumption-oriented devices will grow at the expense of production-oriented ones.

Distribution Control

Amazon’s proposition with the Kindle line is subtly different. While the Kindle (which term I use generically hereafter to refer to the entire line of Kindles, including the e-book readers and the Kindle Fire) is openly a pure consumption device, it’s also cheap enough to be offered as a supplement to PCs as opposed to a replacement. Their purpose is narrow: to remove all friction from the consumption of digital media. Amazon, as a retail company, is trying to aggressively commoditize its complements while controlling the distribution channel.

For Amazon to make money through this model, though, everything has to flow through Amazon. And with how much of the e-book retail business Amazon already controls, it’s an outcome that doesn’t seem unlikely. With a free month’s subscription to Amazon’s Prime service with every purchase of a Kindle Fire, they’re instructing consumers on how to easily stream movies from their catalog and buy music directly in addition to buying everything else, from toilet paper to power tools. Meanwhile, music streaming services are becoming more popular. The inexorable trend seems to be toward subscription services for everything. It isn’t difficult to imagine a future in which you can play a flat rate to Amazon for access to an all-you-can-eat streaming service for all content: music, magazines, newspapers, books, movies, tv shows; all on free hardware ranging from e-ink devices, tablets, set-top boxes (hooked up to televisions bought from Amazon with one click and 2-day shipping).

Centralization

In Apple’s vision of the future, you buy apps and content on iTunes, which you consume on your various iDevices. In Amazon’s, you get a free Kindle with a subscription to Amazon Prime, and stream all your content. The real price of the ease and ubiquity of access offered by both of these companies is centralization of control. These visions represent a step backward into our recent past of broadcast media, with a few powerful entities acting as gatekeepers to our collective culture. Which may seem inoffensive if you prefer curated, professionally produced content anyway. But then you might realize that Apple plays an actively censorial role in its App Store. You might find that Amazon has bizarrely tone-deaf and comically ironic behavior in its recent past, like remotely deleting purchased copies of George Orwell’s 1984 from Kindles.

The most frightening and subversive part of Amazon’s strategy to me is the Silk Browser. While offering the proposition of a better browsing experience, Amazon has begun to insinuate itself into our great decentralizing force: the World Wide Web.

It is perhaps simply the historical cycle of a new technology’s decentralizing force dissipating into subsequent centralized control described by Tim Wu in The Master Switch.

Lamentation

My lament, then, is for the dying bazaar. For the crushed dream of the blogging revolution, replaced by closed and proprietary systems like Facebook. For the notion of putting the means of production into the hands of the masses, replaced by ubiquitous and unimaginably slick consumption devices. For a flourishing chaos of freely exchanged ideas, curated, ranked, and sorted by people, organizations, and algorithms that are themselves curated, ranked, and sorted. For a more participatory World Wide Web.

An Alternative Vision of the Future

Not that it’s over. As I’ve noted before, Google is a huge and powerful engineering monastery that has a symbiotic relationship with the Web and DNA consisting of decentralized production and distribution. And there are plenty of smaller efforts to make the Web a competitive endeavor, as well certain types of participatory depth and richness of experience that only the Web can provide. And we’ve only begun to see the new frontier of possible decentralized business models for content production afforded by the open Web.

Looking around the Web, though, it’s not that hard to understand why people might seek refuge in closed systems. A lot of it is unavoidable. Some places on the Web will always be cesspools. But there are some matters of convention we can change by consensus to make the Web a better, more participatory place that encourages deep, serious thought and collaboration as well as frivolous ephemera. There are ways to optimize reading, civil discussion, and community-building. And the beauty of a consensual, conventional system is that I just have to convince enough people that this is worth doing.

We Are What We Choose

Wed, 09 Nov 2011 14:20:00 -0800

In this post on Cyborgology, PJ Rey convincingly speaks to the project of dismissing the concept of digital dualism. He argues that as technology becomes sufficiently advanced and ubiquitous, we begin to think of it as natural extensions of ourselves.

As such, is [sic] not hyperbolic to claim, for example, that Facebook is a piece of equipment that has become an extension of our very consciousness. As equipment, social media fundamentally alters who we are.

Rings true for me. But more than anything else it emphasizes the importance of thinking carefully about the ways in which we are extending ourselves.

Noam Chomsky provoked a great discussion in the last few weeks by offhandedly dismissing new media as “extremely rapid, very shallow communication.” Nathan Jurgenson responded in a piece in Salon, talking about the great benefits of new media, e.g. its utility in popular movements like the Arab Spring and “Occupy” protests, as well as the biases inherent in Chomsky’s somewhat crotchety dismissal.

Chomksy, of all people, ought to take note. When he defends his form of communicating (printed books and periodical essays) with claims that tweeting/texting lacks depth, he is implicitly suggesting that nonwhites and those in the Third World are inherently communicating less deeply than their white and first-world counterparts. He doesn’t seem to know enough about the reality of social media to examine his own assumptions.

But ultimately I agree most with this post by Mike Plugh.

Chomsky’s critiques, while they seem to have some overtones of morality, read in a particular way, aren’t all that far off the mark. The shallowness of social media forms, like Twitter, is both the strength and the weakness of their nature. As a medium, Twitter is largely superficial, shallow, and evanescent. One can only be so ‘deep’ in 140 characters or less, and so the superficiality of the messages prompt higher involvement by the audience to derive meaning (a largely positive characteristic), while at the same time leaving out important depth behind the communication.

…

Jurgenson rightly points out that, “To some degree (and we can debate how large this is) social media has enabled a mass manufacturing of dissent.” However, dissent is a contrary position, and a position fueled by dissatisfaction and emotion. Immediate communication of a type less-considered and nuanced is largely emotional in nature. Again, this is a strength when outrage and action are the prescription for intolerable circumstances, but do nothing little for producing solutions or the important enlightened perspectives that must accompany any change for it to be lasting and serve humanity most broadly. Dissent as a state of moral indignation is useful and important, but is also the same fuel that ignites the mob or the riot when lacking in substance or focus. Manufacturing consent, as Chomsky famously observed, is a process of propaganda which relies on appealing superficially to people’s gut emotions and less deeply considered convictions.

Both posts are entirely worth reading.

I take from this discussion that not only do different digital tools have different strengths and weaknesses for different activities, but they each encourage certain types of behavior. To quote myself:

[The] quality of the connections on Facebook is very poor. And this is because the signals are low bandwidth by design. Input fields for text updates are small and restrictive. Photos are easy to share but poorly presented, the system implemented not to share rich, expressive photos, but low-grade, mundane, documentarian ones: I was here; I did this; I was with these people. Ubiquitous interaction is reduced to comment fields that encourage short, ad hoc reactions, or—worse—to the elemental, further irreducible unary piece of information: the “like” (though I suppose the “poke” contains even less information). Facebook casts an extremely wide net into the waters of the collective mind-space of its users, but that net penetrates only into the shallowest of depths.

Twitter provides accessibility with a similar sacrifice of bandwidth. I’ve always felt like Twitter’s great innovation was not the arbitrary character limit, but the frictionless interface that collapses the separation between input and output. The system itself doesn’t judge how noteworthy your expression is, leaving that to the network at large in as close to a democratic way as is really conceivable. But the limitations, also similarly to Facebook’s, enforce a shallowness of sentiment, encourage pith over comprehensiveness, and discourage real discourse, or any sort of conversation at all.

And the web itself sometimes seems inimical to depth of thought.

That we are subsuming technologies into ourselves seems trivial compared to the point that what we are subsuming encourage us to think only in a certain way. If we examine carefully what we are becoming, I don’t think we’ll like what we find.

And some, like Nicholas Carr, have already, and they don’t like what they find at all. He, and, presumably, Chomsky, would advocate, then, a retreat back to the old media. Full-length books are what encourage the deep, sustained concentration necessary for truly worthwhile thought. I think “Digital Fasting” movements (restrictions of internet/screen time) also stem from a sentiment that these technologies are having a pernicious effect on us.

But a retreat to old media can’t be the correct answer. Not only because of the limits of physical media—like relegation to passive consumption and the lack of communal experience—but also because the old media have assumptions of their own built in that are just as pernicious. There is an authoritarian centralization of power inherent to the old publishing model, for example, as well as limited viable formats. Ideas of a certain size, too long for long-form, but too short for full-length books, simply had no way to exist in old media. They had to either be cut down or inflated to fill viable formats.

New media have no such restrictions. In fact, our new digital tools are the most versatile, plastic, and accessible that we’ve ever had. We should be honest and clear about what our tools are doing to us. But we should then focus our energy on creating the tools that we want, instead of trying to stop the inexorable tide of change.

In Defense of the New Google Reader

Wed, 02 Nov 2011 16:43:14 -0700

Brian Shih, a former PM for Google Reader, wrote a scathing take-down of the recent changes to Google Reader on Monday.

And plenty of his criticisms are well deserved. It’s certainly true that the visual changes seem to prioritize consistency with the rest of Google’s visual redesign over the practical considerations of the product.

I think the visual criticisms are ultimately pretty minor, though. There were plenty of complaints about the preview of Gmail’s recent user interface changes as well, but in the official release, Google seems to have addressed most of them. And Shih himself agrees that Google should be trying to visually unify its products. Finally, let’s not kid ourselves: the old Reader was uglier, and not terribly more usable. It was just what we got used to.¹ Will Google eventually find the right balance between visual consistency and actual usability? I’ll reserve judgment for now.

The real meat of Shih’s criticism, though, is for the new sharing model. Again, he agrees that Google should integrate Reader with Plus. But he goes on to make some criticisms that don’t fully take into account the differences between the two products, and others that simply get the facts wrong.

First, he complains that the new sharing flow makes sharing items harder. Fair enough. But, as I wrote in an earlier blog post, this is purposeful. It adds friction at the point of sharing, makes you pause to consider who the audience for this particular item is. Which is how it should be.²

Second, he complains that you must publicly +1 a post in order to share it. This is incorrect, as he notes in an update.

Third, he complains at the lack of a transition from the old sharing model to the new one under Google+. And this seems to be what many people seem to be most angry about. Google Reader users are widely upset that the communities that have organically grown in their feed reader has disappeared. What these people forget is that the old sharing model and the communities that formed in it were opaque, irrational, and hard to control. Discoverability was basically nonexistent, loosely based on your Gmail contacts. It was never very clear who could see your posts, and it was never clear who could see the comments you make on posts.

The new sharing model, under Google+, is perfectly clear about who can see your posts, who can see your comments, granular about who you can share with, and gives you more control over comment moderation. In short: it’s better in every way.

I agree with Shih that it’s unfortunate not to be able to consume shared content from Reader. In fact, I’d go so far as to say I’d prefer to consume, +1, and share all my content through Reader’s interface, and not have to visit the G+ site ever again.

The one glaring omission from the new sharing model to me is the omission of an important point on the spectrum public-ness: public but not broadcast. Under G+ now, if I want to share an item publicly, I am also thrusting it in front of everyone who has me circled. But sometimes I don’t want to do that. Sometimes I want to share something about bipedal robots, for example, that I don’t care if it’s public, but I also don’t want to stuff in everyone’s feed who knows me, for the same reason that I don’t care if anyone knows about my fear of bipedal robots, but I also don’t bring them up in conversation with everyone I meet.

What G+ needs is publicly subscribable Circles. These would be feeds that you can post to that are public, but only people who have chosen to subscribe to them can see. It fits nicely with the granularity of G+, and adds the organic growth model of the old Reader in a much clearer way.

Google is notorious for releasing new products that seem half-baked. Only a few days after the public outcry over changes in Reader, there is a new one over the new native Gmail client for iOS. Google does this a lot. But it also iterates on these products at a furious rate. And it appears to do this in reponse to public feedback. Again, the new Gmail redesign has addressed almost all of the initial criticisms as far as I can tell. Which means the public outcry is good. It tells Google what to fix. I just wish the outcry was over stuff that matters instead of what strikes me as simple aversion to change.

There were plenty of browser extensions and user styles that ‘fixed’ the old Google Reader appearance along many of precisely the ways Shih complains about here. ↩
I do also lament the removal of keyboard shortcuts, but I assume they’ll be added back in soon. ↩

Intimacy is Performance

Wed, 19 Oct 2011 15:56:00 -0700

I was surprised to find a lot of interesting discussion around my last post about Intimacy and Performance on Facebook. Some of it was about the distinction between ‘intimacy’ and ‘performance’ and a lot was about mediated experience. Both of the discussion tracks got me thinking about our relationship to communication technology.

We can measure the progress of communication technology in two ways:

The reduction of physical limits on social interaction.
The increase of signal fidelity of digital communication to pre-technological social communication.

Inventions like writing, telegraphy, telephony, etc. break down the limits of physical interactions geographically and temporally, but each introduces new limits. Telephony, for example, broke down the barrier of distance, but introduced limits of one-to-one and voice-only. As telephony and its replacements improved, these introduced limits disappeared. With group video chat technology, we can now have fairly high-fidelity social experiences across long distances without those early limits, and we can safely expect the fidelity to continue to increase.

Intimacy is Performance

At least part of the reason I posed a distinction between intimacy and performance in my last post is that that’s one of the limits that our technology imposes on our social interactions. Offline, intimacy is much easier to understand as a subset of performance: performance in a smaller context. In our analog social interactions, we have many different contexts that overlap in complex ways, and we perform slightly differently in all of them. In the context of close friends and family, that performance is called ‘intimacy.’

Limits

Earlier communication technologies were easier to understand and adopt because they were simpler, and they more easily mapped to physical interactions. Letter correspondence and telephony were easy to understand as versions of a face to face conversation. Email, SMS, and chat are similarly easy to understand. But the increasingly complex world of online social networking does not map to the offline very well.

Chris Poole, founder of 4chan, provided one interesting criticism at Web 2.0 conference, which @dannygilligan sums up nicely on Twitter: “Sharing on the web not a question of who you share with but who you share as.” But that’s only half-correct. The fact is, offline these two concepts “who you share with” and “who you share as” are really the same thing: context. It’s only online that they’re separated artificially, and this asymmetry is one of the limits of the digital social medium.

Many of these limits are ones of sharp distinction: between private and public; and between one-to-one and one-to-all. While things like email lists made the representation of different contexts possible, they didn’t make it easy or frictionless. Facebook and Google+, with their respective list features, each take a step toward removing some of the friction by offering more granular contexts, but these steps are not big or particularly inspired ones.

Even with Google+ Circles, to take an example, the asymmetricality and nonobviousness of the social context places friction on every interaction. When you post to a Circle, you decide what your context is. But when you comment on someone else’s post, it takes some work to figure out the context, and thus, your behavior. That’s friction. This is what I was trying to get at with an earlier post on ‘place’ as a more intuitive metaphor for social networking.

Context is something that might be hard to think about when implementing a social network, precisely because it’s so ubiquitous, obvious, and unconscious in an offline setting. But these are reasons why it should be a prime consideration for any online social interaction service.

Of course, in the digital medium, there are many novel concepts that simply can’t be analogized to offline interactions easily. The fact of a permanently and publicly accessible record of interactions is the most obvious one. The spectrum of publicity is another. I would argue that the spaces on that spectrum that digital media open up to us are genuinely enriching and useful. It means, for example, that @mention conversations on Twitter, which aren’t broadcast (don’t appear in followers’ timelines with exceptions) but are publicly accessible if you look for them, provide access to an entire category of information on the internet that was never available before.

I think these are the type of technological innovations that are drastic enough, i.e. don’t map well enough to previous paradigms, that they represent inevitable shifts in the culture. It’s hard to recognize calls for a new ‘forgetfulness’ on the internet as anything other than nostalgic Ludditism, at least for me. I think making social context clear through user interfaces will mitigate some of these effects. For example, a permanent record of statements you make within a specific context will only ever be available to others who share that context, so the eternal memory of the internet doesn’t matter as much as the eternal memory of the public internet. But some of the digital age’s affordances will be drastically culture-changing, not features or bugs, but simply facts of digital life.

Presence

There are also more ambiguous artifacts of digital representation. The combination of ubiquitous access and time-shifting capabilities afforded by the internet blur the old distinction between real-time presence (face to face, telephone) and correspondence (letters, email, SMS). There is now more of a spectrum of synchronous to asynchronous communication, with some technologies, like Twitter and chat, existing on a broad range of it.¹

Collapse

It seems pretty clear to me that as digital identity and analog identity become increasingly collapsed, the notion of presence will resolve away from binary states and toward a constantly fluctuating analog amount. We will always be available for digital communication at some level, but we will be more or less digitally present depending on what we’re doing and what kind of access we have at any given point. This is already true enough that the idea of going completely ‘off the grid’ for some amount of time is increasingly popular and we are already starting develop social norms on behavior related to immediate level of physical-digital presence.

Nathan Jurgenson, who writes about “digital dualism” on his blog, emphasizes that the distinction is artificial.² I sort of disagree. I think the virtual is currently another context with varying degrees of overlap with other contexts, just as ‘work’ and ‘home’ and ‘friends’ are different contexts. And somewhat paradoxically, I think the distinction of ‘digital’ as a separate context will fade and collapse as our digital tools more accurately represent the distinctions of our offline contexts.

But I don’t think publicity-pushing policies like Facebook’s help to close the gaps of digital dualism. Facebook’s goal is to obliterate context online by defaulting everything to public. If anything, this abuse of users’ trust is driving them to compartmentalize their digital lives more. It drives them toward more conscious and more active management of their digital identities, which is the exact opposite of what Facebook wants.

The problem left to solve is removing the friction of cognition involved in placing yourself in and managing your social context.

Intimacy and Performance on Facebook

Tue, 27 Sep 2011 20:22:05 -0700

Facebook launched three interesting new features last Thursday: Timeline, Ticker, and Open Graph.

Timeline

Timeline is a replacement for the profile page that collects and displays your Facebook status updates, photos, links, etc. in a clean, easily perusable way. One of the first things I thought of after seeing the announcement, was this “Dear Sophie” ad from Google.

Superficially, they’re very similar. Both showcase the use of technology to capture meaningful events and share them. Even the types of relationships and events are similar. But the different presentations betray an important distinction in purpose between the two sets of services.

The “Dear Sophie” ad shows technology in service of creating records and sharing moments in an intimate way, and ends with the sentiment: “The web is what you make of it.” While the Timeline promotional video shows similar vignettes, Mark Zuckerberg’s presentation closes the video with: “Timeline is the story of your life. It has three pieces: all your stories, your apps and a new way to express who you are.”

In other words, while “Dear Sophie” is an expression of an intimate set of messages from one individual to another, the Timeline is an aggregated set of shared moments, collected in the service of public display, an artifact of self-expression and identity. And what I find jarring about this formulation is the same thing that bothers me about the alarming trend of weddings in which the photographers and videographers have free reign, even during the ceremony, in order to get the best, most cinematic record of the event, at the expense of the event itself and everyone participating. It’s a conflation of the record of the event with the event itself, or even a privileging of the record over what gives the record its meaning and power. At the same time it (ingeniously) adds to the pressure to record all meaningful events on Facebook in order to make sure it becomes part of your identity.¹

Ticker and Open Graph

Open Graph adds a way for Facebook apps and Web sites to make updates for you in various ways. For example, if you give it the proper permissions, the newly integrated Spotify app will automatically share each new song you listen to. But these updates don’t go into your main newsfeed, they go instead into the Ticker, a new area on your Facebook page that shows a running update of your friends’ auto-shared activities online. Other examples of Ticker activity include: reading an article, watching a video, playing a game.

Zuckerberg refers to this as “frictionless sharing,” a dream of a kind of meta-panopticon with which everyone can see what everyone else is doing in real time. There are parts of this that I actually find kind of compelling. The Ticker is a nod toward “ambient intimacy” the idea that you can approximate the low-level intimacy of occupying the same space as someone, in a digital way.

But, by being public, the Ticker fails at achieving intimacy. Because if being in the same room with someone creates intimacy, being in the same room as everyone creates the opposite. It turns all of your activity into performance. And what many have hinted at is that removing friction from sharing just displaces that friction. If everything I do on the web is under the public gaze, I have to reflect for a moment before I take any action — before I listen to a song, watch a video, play a game, or click on a link. It simply moves the friction from sharing onto the activity, in the worst kind of self-censorial way.²³

This is especially true in combination with the Timeline, which aggregates Ticker activity into “a new way to express who you are.” Not only does the Open Graph and the Ticker turn your online activity into a performance, it then turns that performance into your very identity.

Intimacy vs. Performance

Both intimacy and performance are important parts of what makes social networks like Facebook compelling. Initially, the performance part was the profile, where you listed your biographical information, your likes and dislikes, your expressions of identity (in the narrow ways that Facebook allowed). The performance part was important for discovering new people and expanding your network, for actually forming the connections. The intimacy part was pretty much everything else: your status updates, messages, comments, photos, and videos. It consisted of the kind of stuff in the two videos above. It was what injected meaning into those connections that you’d made.

At every turn, Facebook seems to have subverted the intimacy of social experiences by turning them into public performances. Not only has the intimacy of what was once private slowly eroded into the public, but more and more of Facebook users’ online activity is being drawn into the performative identity. If an anthropomorphized Facebook had a Facebook profile, its Timeline would show a clear progression of updates that moves from mostly private toward all public, all the time.

The worst part is that I don’t think this has the effect that Facebook wants. I don’t think a frog-boiling style of slow erosion of privacy means people just continue to share in the same way except in public. It means that the people who understand what’s going on become wary, stop trusting, and eventually stop using the service. And people who don’t understand what’s happening will eventually hit situations in which something doesn’t work the way they thought it did (often embarrassingly), and the uncertainty of their mental model will result in less usage and make that usage more tentative and more careful.

Ultimately, it just means less intimacy. Less signal. Less of exactly what this kind of technology is supposed to enable.

"You are the product."

Tue, 06 Sep 2011 22:42:13 -0700

In a talk last Friday at the dConstruct conference, design luminary Don Norman repeated an increasingly popular refrain about Google: “[In] fact, the advertisers are the users and you are the product.” It’s the kind of glib soundbite that certain bloggers can’t resist, but collapses on close scrutiny.

The argument goes something like this: You are only truly a company’s customer if you pay them directly for products or services. Google users don’t pay Google directly, therefore Google users are not Google customers, but products that Google sells to advertisers. The logic that underlies this argument is that a user must be a customer or a product, never both, and the implication is that it’s bad to be a ‘product.’

The sentiment wouldn’t be compelling if it didn’t have some kernel of truth. It’s certainly true that any advertising-based business model has a different relationship with its users than a more traditional one that sells products directly to customers. But the dichotomy between customer and product is a false one, and it’s precisely the incentive structure of the ad-based model that shows it to be false.

Accountability

Ultimately, the argument is one about accountability. Apple, to take an obvious example, is accountable to its customers because customers pay them. Their incentive is to make the highest quality product so that customers keep paying for them. Google, the argument goes, is not accountable to its users, because the users don’t pay Google. It takes only a moment’s consideration to realize that this line of reasoning is facile at best, and actively dishonest at worst. Because of course Google is still accountable to its users, just not through direct monetary transactions. Google is accountable to its users because advertisers won’t pay if they don’t have users. Google is also accountable to advertisers, but that doesn’t negate its accountability to its users. It still needs happy users to make money.

Alternatives

It’s really an argument against advertising in general, and ignores the fact that advertising has historically enabled the production of all kinds of media, from network television shows to newspapers, to magazines to ahem tech blogs,¹ that would otherwise have been impossible. It ignores the fact that Google, specifically, has given the world access to a mindbogglingly useful utility (its search engine) for free, supported by ads that are often actually useful. And that’s not to mention all the other amazing products that Google offers for free.

Incentives

There are certainly trade-offs. Because Google can target ads better with more information, it has an incentive to gather as much data about its users as possible. But this incentive has always existed for media companies like newspapers and magazines. Information about the subscribership was always more valuable than the subscriptions themselves, because publishers as a matter of course sold their subscribers’ information to advertisers. At least Google just uses the data internally to target ads. This data-gathering incentive will always leave Google at a point on a spectrum of privacy that some people will consider invasive and others won’t mind.²

But it’s not like a direct transaction business doesn’t have some bad incentives. Apple, as a vendor, wants to extract as much money as possible from its customers. Most of the time this incentive translates to creating high-quality products. But sometimes it translates to hindering competition through artifical extra-market means like patent litigation. Apple also has the incentive to produce its products as cheaply as possible. Most of the time this incentive translates to making its operations as efficient as possible. But sometimes it translates to sacrificing worker conditions. These are both controversial issues, but my point is that there are good and bad incentives in any business model, and there are plenty of examples of abuse in both.

Currency

I think it’s easy for rich people to say things like “you are the product.” But I don’t think most people, even having considered all the implications, consider themselves products. I certainly don’t. We just think of ourselves as paying for the products with alternative currency, and we’re glad that we can avail ourselves of these products in exchange for resources that we have in great supply: our attention and our information.

Meaningless

The incentives are real. Some are good and some are bad. They can be useful to point out as explanatory forces for real practices, but without evidence of real malfeasance, sentiments like “you are the product” are just empty rhetoric.

via GigaOm

The most dishonest part about someone like John Gruber making this argument is that Daring Fireball is, of course, wholly ad-supported. ↩
To a large degree, I find Google’s gathering of my information — my personal data, my browsing habits, my interests — to be mostly inoffensive. Keeping that information secret doesn’t do me much good, but Google can aggregate my data with that of others and make it hugely useful. That’s pure wealth creation as far as I’m concerned. ↩

Google's Open Hand and Closed Fist

Mon, 29 Aug 2011 23:46:27 -0700

Paul Buchheit coined Google’s unofficial motto - “Don’t be evil” - early in the company’s existence. But Google has only been pushing the vague notion of “open” in the last few years. The notion is vague at least partly because it’s so all-encompassing and partly because of Google’s penchant for using it in whatever way is most convenient at the time. But that sword cuts both ways, as it makes it easy for Google’s critics to call the company out when they don’t live up to an equally vague standard. So, what exactly does Google mean by ‘open,’ and how open is Google?

Open

As laid out in this this blog post, Google refers to a couple of different things with the term: “open technology” and “open information.” Open technology, in turn, consists of: open source; and open standards and APIs. Google is unequivocally open on these two technology vectors with respect to certain products: Chrome, Chrome OS, Wave, Buzz, Pubsubhubbub, etc.¹

Closed

On others, they’re decidedly closed. Google’s search and advertising businesses are very opaque.² Which is certainly understandable, since Google makes the vast majority of its money on search and ads. It can’t really be open with these because their entire business depends on them, and I think most reasonable people can understand that.

Many have also rightly criticized Google for its notoriously bad customer service, but I don’t think this qualifies as ‘closedness’ as much as poor execution. From what I understand, this opacity isn’t a result of any desire to keep secrets so much as an unwillingness to regard the problem as a non-engineering one and devote adequate human resources.

I’m not sure if Google’s level of secrecy in general is less or greater than the average large corporation, but my sense is that it’s not. It’s a little beside the point, anyway, because this isn’t one of the things Google refers to when it claims to be open.

Android

The venue in which Google has taken the most criticism, Android, is a complex and ambiguous one. There are several vectors of criticism here, too. Part of the problem is that there are two distinct parts to Android. There is the completely open source part, the source code to which is available to anyone who wants to download it and run on any device they feel like.³ Then there’s the ‘Google experience’ Android that ships with Google’s closed source apps: Market, Gmail, Maps, Navigation, and Voice; and only with explicit approval from Google. For clarity, I’ll refer to them as AOSP (for Android Open Source Project, as its generally referred to in open source circles) and Android, respectively.

AOSP

AOSP has taken some heat from the Open Source Software (OSS) community for two reasons. First, there’s been some discussion of Google possibly violating the GPL license, but there’s mostly disagreement on the topic and, from what I gather, even if true it’s a borderline violation. Second is AOSP’s somewhat unconventional operation. Unlike most OSS projects, even other Google ones like Chrome and Chrome OS, Google develops most of the software internally with little to no transparency or input from the community, only periodically dumping the releases into the public. This has been the Android team’s modus operandi since the beginning. In this respect, Google is more closed with AOSP than most OSS projects, even other Google ones. With Honeycomb, the latest version of Android, this has gotten even worse. The delay between Google’s release of the source code to its partners and its release to the public is the longest to date.

But let’s put this into perspective by considering the total spectrum of open to closed. From this vantage, the difference in openness between AOSP and other OSS projects is basically neglibile compared to the difference in openness between AOSP and any other major player in the industry. Because there are no other OSS mobile operating systems.⁴ So is Google open with AOSP? I would say yes, even at its most closed.

Developers

Another vector within Android is in relation to developers. Here, Android as a whole is as open as it gets. The Android Market is available to anyone who wants to develop for it.⁵ Developers can use any tools they want. And app distribution isn’t even restricted to the Market; i.e. you can install any compatible app from any source, including simple internet downloads. Amazon has even released their own app store for Android.

Consumers

A third vector is in relation to consumers. Here, again, Android is completely open. On the devices that Google was directly involved with, like the original G1, the original Droid, and Google’s flagship Nexus devices, the hardware is unlocked and root access is readily available, which means you can install third party mods or ROMs, i.e. custom modifications to the operating system. Some of these third party mods, like CyanogenMod, have an active, thriving open source development community, and are adding features faster than Android proper.

Manufacturers and Carriers

The caveat here, of course, is that the majority of handsets are locked down by the manufacturers, probably at least in part at the behest of carriers.⁶ And so here we come to a vector on which Google is not completely open. Google’s policies toward vendors in the early days of Android were quite open. And the vendors took full advantage, by locking down the hardware and adding extensive user interface customizations on top of stock Android in an effort to differentiate their brands. They also began to pre-install apps that you couldn’t uninstall. It even extended to such liberties as making Bing the default search engine with no option to switch. The result was a fragmented Android ecosystem as well as a generally poorer experience for the end user.

But Google is increasingly using control of its closed source application suite and access to the Android Market to exert pressure on the manufacturers and carriers away from these practices. Some of the details that came out of the Skyhook⁷ trial even indicate that Google now requires explicit approval over each model before it ships.

So what we see here is Google stuck between opposing vectors: open to consumers vs. open to vendors. It can’t be completely open to both. They seem increasingly to be choosing openness for consumers, which necessarily means being more closed with the vendors. So is Google being less open in some ways with Android? Yes. Do I care? Not a single iota, because it means they’re being more open along a different, more important vector. In fact, I encourage it.

Mostly Open

So is Google perfectly open in every way? Of course not. But overall, if we can consider openness on a spectrum, then, yes, they are more open by a ridiculous margin than any of their competitors in any business except search and advertising. In that business they are exactly as open as everyone else (which is to say: not at all). Is that enough to legitimately claim to be ‘open’ as a PR/marketing tactic? I say yes. I obviously think there’s a lot of nuance that they elide over, but I don’t think it’s ultimately disingenous. I would appreciate if they addressed this nuance more, but I have no idea if it would be practical. (Because does the general public know or care about any of this? Probably not. Of the people who do care, will any of them be swayed? Seems unlikely.)

“Don’t be evil.”

So, is Google open out of moral sentiment or out of self-interest? Well, here’s what’s special about the company: both. Google has managed to invent a business model wherein their incentives align with consumers’. Their promotion of an open, decentralized, interoperable Web accessed through commoditized (maximally available) hardware and software serves Google and the world at large simultaneously. It also sidesteps the adversarial nature of selling things to customers. And despite what I think is a fundamental conflict, Google does this by making advertising useful.

Are there also some incentives that are bad? Yes. Google has a strong incentive to collect data. This is mitigated by their need to keep their user base happy. Which means they have to weigh data collection against privacy intrusion. But the incentive exists. How is Google doing on this front? Not perfectly, but certainly better than others, and certainly not as bad as they could probably get away with without getting in trouble with the general public.

And I think this is where the company’s moral founding principles emerge. “Don’t be evil” may be glib and too broad by far. It may lack the nuance to address some real issues. But it propagates, in a blunt way, from the founders and leadership down into important decision-making, recruiting, etc. Google has made compromises, certainly. The partnership with Verizon on net neutrality and its complex history in China are the two examples that dismayed me the most. But Google’s public and open espousal of such a simple and powerful principle as “Don’t be evil” leaves them open to criticism. It makes them accountable for their actions to more than just the shareholders in an important way.

Google+ is a notable exception to this list. Google hasn’t said much about it, but my working theory is that, after the disastrous failures of both Wave and Buzz, their plan with Google+ is to try to offer a solid, compelling user experience first to gain significant adoption, and only then start to open up the technology for people to start creating their own interoperable social networking services. I think some of its inexplicably draconian policies, on pseudonyms, for example, make sense from this perspective. ↩
Though, they do talk about it. ↩
This is the Android that Andy Rubin was referring to in his infamous tweet. ↩
Though Nokia did partner with Intel to create an open source mobile operating system called Meego, it never grew out of vaporware status. ↩
Though the ability to sell paid applications on the market is restricted on a country-by-country basis, with full access only slowly being rolled out. ↩
This means it takes time for third party mod developers to find the necessary exploits to get root access and bootloader access before they can install custom mods. While this often happens before the devices are even released, it also often takes weeks or months, and sometimes never happens at all. There’s been a push away from these restrictive policies, however. HTC recently released a bootloader unlocking tool that gives you full access to the hardware, though its use voids the device’s warranty. ↩
More here. ↩

A Different Kind of GUI

Tue, 23 Aug 2011 22:47:00 -0700

“In the beginning was the command line.”

But then the command line became graphics and dwelt among us. From its earliest days, Apple privileged the lay user over the technical one, regarding the need to understand technical details as friction. With the Lisa and the Macintosh, Apple picked up where Xerox Palo Alto Research Center left off. The guiding principles of intuitiveness and discoverability led Apple to replace the command line interface (CLI) with the graphical user interface (GUI), featuring windows, menus, icons, and a pointing device (WIMP).

For these historical reasons, the GUI has always been pointer-driven, at least with respect to computer operating systems. Apple has only recently begun the transition to something new: iOS’s direct manipulation interface; still a GUI, but no longer driven by the same windows or menus or a pointing device. With this transition, Apple achieved its oldest dream more successfully than anyone anticipated. And it now seems to be pulling the rest of its product line into that dream. To use Steve Jobs’ terminology, they’re slowly transitioning out of the “truck” business.

The Mouse

It’s enlightening to consider what made the mouse such an indispensable companion to the GUI for so long.

It’s intuitive. It’s easy to grasp the basic concept that the cursor is an extension of the hand, that its motion corresponds to the motion of the mouse.
All interactions are built on a minimal set of axiomatic actions: move, hover, click, release, and drag.
It provides highly precise and unrestricted motion of the cursor, as opposed to a joystick, which limits the cursor’s direction and velocity.

With careful application of the WIMP metaphor, these features made every aspect of a GUI accessible to the user. But this was only the minimal set afforded by the technology that was available at the time. With touch screen technology and the direct manipulation paradigm, Apple has not just removed an entire layer of abstraction, but also reduced that minimal set of actions even further to a more intuitive level: tap, swipe, and pinch. The resulting interface is qualitatively more intuitive and trivially easy for new users.¹

But restricting the minimal set of actions that far has implications that I think Apple is only starting to realize. With such a limited interaction bandwidth, for example, any given complex action has to take a correspondingly large number of sequential inputs. Apple’s way of mitigating this effect in iOS appears to be gestures, some of which, like five-finger pinch, are frankly ridiculous. Traditional desktop GUIs mitigate this by overloading the basic controls with incrementally added extra functionality that breaks metaphor: double-click, right-click, the scroll wheel, any number of additional mouse buttons. These are all extremely useful additions (because they increase the interaction bandwidth), but this phenomenon is actually part of what makes modern desktop GUIs inaccessible to novices and is why I imagine Apple has traditionally been so resistant to these kinds of innovations. But I think the underlying problem isn’t simply the tendency for cruft to accumulate, but the severely limiting nature of the metaphor to begin with. iOS, just a few years old, already contains more undiscoverable, out-of-metaphor inputs (gestures) than discoverable strictly in-metaphor ones.

But even aside from the unintuitiveness, there are a lot of reasons not to like the mouse:

The arbitrary motion of a mouse cursor makes the effort analog, as opposed to the digital motion of keying. Commands, even repetitive ones, can’t be relegated to muscle memory the way command line commands or keyboard navigation can.
Because of this, mousing is a conscious process. Hitting precise targets takes a lot more cognitive effort and close attention than keying.
Moving the mouse hand back and forth between mouse and keyboard is often an annoying cognitive task-switch and that breaks flow.
The requirement of moving back and forth between mouse and keyboard creates friction for typing. And since typing is one of the main ways of producing content, this friction is a particular type that only applies to production and not to consumption.
It’s slower, or feels slower. (According to Bruce Tognazzini, Apple R&D found that mousing is in fact faster than keyboarding. This strikes me as highly dubious today not only for all the reasons given above, but also because the article is from 1989 and the research is presumably from before that. I believe the research was also done on non-expert users.
While the high degree of precision afforded by the mouse is a boon to applications that benefit from it, as the dominant mode of interaction, it forces that level of precision on every transaction. For every dialogue box with two simple choices, for example, the user still must traverse an arbitrary distance and hit a cursor target that is tiny relative to the rest of the screen, which is, for the transaction, completely unused. So to make an input that’s essentially a binary 0 or 1, the user in real terms has to make a huge, well-calibrated, cognitively expensive analog input that’s largely wasted.

With OS X 10.7 Lion, Apple presents a clear and well-documented iOSification of OS X.² A prominent sign to me is the seeming move toward deprecating the mouse, replacing and adding functionality with touchpad gestures. I empathize with the desire to leave the mouse behind. But where some of the iOS gestures are quite silly, the decisions Apple has made about Lion’s touchpad gestures are strange in a different way. I’ve discussed before Apple’s attempt to render the direct manipulation interface of iOS on the fundamentally indirect interface of the desktop. But aside from that, these gestures are no more efficient or discoverable than keyboard shortcuts and they do make you move your hand off the keyboard. While they are arguably a bit more intuitive and reminiscent³ of a real touch-based direct manipulation input system, it’s at heart really just a case of trading one set of undiscoverable metaphors for another. At least with keyboard shortcuts, it’s possible to expose functionality because the keys have a conventional labeling system built in. Gestures do not. I don’t see any real way to reconcile the inherent and fundamental difference between the two paradigms (direct and indirect) with the ergonomic constraints of different types of computing. Maybe Apple’s solution will be to further marginalize the needs of the technical user and simply move to a touch-only interface, with touch-screen laptops as their physically largest products. It’ll be interesting to see.

Building Blocks

But if this does mark the beginning of the end of the pointing device-driven model for consumer computing, it seems like a good time to rethink the technical user’s computing experience, unburdened by the need to cater to the novice or non-technical user. While the CLI seems to be regaining prominence and popularity among some subset of users, I don’t think a return to the command line is practical. The GUI has introduced some concepts and workflows that are very powerful. As real computers with dedicated input devices become more and more relegated to technical users, maybe a more robust but less intuitive interface becomes more viable. In fact, maybe we can abandon ‘intuitiveness’ as the primary motivator or at least stop defining the word as minimal interaction bandwidth.

Some starting points:

A purposefully high-bandwidth interaction model, i.e. a large but consistent set of initial undiscoverable actions to learn, that, at best, isn’t meaningfully less ‘intuitive,’⁴ but instead simply has a learning curve with a different shape.⁵
The mouse is too powerful to get rid of entirely. There are many applications for which it is an ideal or at least very appropriate input device, such as graphics rendering, photo manipulation, and first-person shooters. So, to minimize travel between mouse and keyboard, functions of the mouse and the mouse-hand side of the keyboard should overlap as much as possible. These probably include: motion and selection.
I’ll revisit 10/GUI again, for some of the fantastic insights therein:
- Single axis of windows (Con10uum).
- Different levels of interaction.
Direct manipulation interfaces suggest that cursors are unnecessary. In our thought experiment, they would probably exist only in certain applications, specifically mouse-based ones.
We can draw inspiration from keyboard-only UIs.
- Vim is a venerable text editor that I’ve recently converted to. Its interface is modal, and its main mode of operation is through a command-based console.
- Windows’ keyboard access of menus, with visual indicators.

Fresh Start

My initial idea doesn’t stray too far from traditional WIMP systems. We’ll retain the windows, icons, and menus, but relegate the pointing device to only when necessary or appropriate. Let’s start with windows:

I’ll steal Con10uum’s single axis for windows with no allowance for vertical resizing, since I agree that the extra axis really only adds to the clutter and complexity.

Without a cursor, we navigate by using four directional keys on the keyboard, which I’ll just call [left], [right], [up], and [down], to select and highlight whole objects e.g. windows, more like navigating console game menus:

In this single axis system, z-order is really the same as horizontal order. So we’ll use z-order in a different way, by adding nested elements, and two more directional keys: [in] and [out].

With any element selected in the UI, we can hit [menu] to invoke a consistent menu of possible actions to take on the element, with a corresponding key or sequence of keys for each action.

This makes possible command sequences that can become unconscious, relegated to muscle memory in the same way typing words can be. It makes movement sequences equally unconscious, as any Vim user will tell you. The benefits of making these into unconscious processes are many. Not only are they potentially faster and less interruptive to cognitive flow, but they can accumulate in a way that analog pointer-based UIs cannot. Once you learn the sequence for a particular menu command and use it enough to internalize it, you no longer have to think about it and can then spend that cognition on learning a new sequence. Learning to use the interface becomes an almost linguistic exercise. Pointer-based UIs can never reach this level because every action, every invocation of a menu or icon is a conscious act that demands much more attention.

Vim gives us other directions to explore with a primarily keyboard-driven interface. The basic principles of fewest number of key presses and proximity to the home row keys seem like good ones. Also, shortcuts to specific locations, e.g. the [os] key would take you directly to the operating system level and the [application] key to the application level. We can also take Vim’s concept of iteration, so [5], [left] would take you left five times. While Vim lives on the far end of a brutal learning curve, it’s a rich source of insight into how powerful a user interface can be when discoverability and intuitiveness are cast aside.

Starting with a high-bandwidth interaction model at the OS level has the added benefit of leaving fewer UI decisions to applications. A user can access more actions through a consistent menu/key sequence throughout the system, obviating the need to learn new keyboard shortcuts and interactions for each new application.

This is just a rough draft of an idea.

It’s an attempt to start thinking in ways unbound by the conventions we’ve been living with for decades. I imagine there are many different directions one can take, starting from the new assumptions we are allowed post-mouse.

The implications of a system like this are probably too big to patch onto the systems we have now. There would be a keyboard-driven interface analog to Fitts’s Law that would have implications for application design as well as operating system design. E.g. an application designer would always want to minimize the number of key presses it would take to move from any location to any other location, which would require a careful balance between breadth and depth of element distribution.

Anything in the near future will probably have to look a lot more like what we already have, and be filled with the compromises that we see in Lion and early demonstrations of Windows 8. But recent advancements provide us with a new context in which to fundamentally reconsider human computer interaction. There’s never been a time when it was more possible to try something truly new, or to more significantly change the landscape of future computing. My fear is that the trend toward simpler, more tightly controlled, narrower user experiences optimized for passive consumption will dominate. My hope is that we use the opportunity to create user experiences that encourage more active consumption, more substantial production, a generally richer, denser, deeper world.

Google, Motorola, and Patents

Tue, 16 Aug 2011 20:38:00 -0700

Google is buying Motorola Mobility for $12.5 billion. It’s Google’s largest acquisition ever. And it’s clearly about patents. But before I go into the issue of patents more broadly, let’s take a quick look at what Google is buying exactly.

Motorola Mobility (MMI), which was the handset and cable box division of Motorola, spun off early this year. Motorola obviously has a long history in the mobile industry (they own the now expired patent on the cell phone) and have a deep and presumably strong patent portfolio. This includes about 14,300 patents, and 6,700 patents pending. According to this, in July they also had $5.5B in cash and deferred tax assets.

So, without going into the issue of patent quality, we can broadly compare this deal in terms of patent acquisition with the auction of the Nortel patent portfolio that Google lost in July to a consortium that includes Microsoft, RIM, and Apple.

Nortel: $4.5B / 6,000 patents = $750k per patent.

MMI: ($12.5B − $5.5B) / 14,300 patents = $489,510.49 per patent.

Plus 6,700 patents pending.

Plus a hardware company.

Plus, as Nilay Patel notes, a very strong patent creation engine.

So a cursory examination suggests it’s a pretty good deal for Google.

Motorola

From MMI’s perspective, it’s hard to see this as anything less than a huge win. Despite its success with its Droid handsets, MMI hasn’t been profitable and has in fact been looking for a buyer since at least last year. They clearly saw Google’s (specifically Android’s) recent patent problems as an opportunity and seized it. MMI CEO Sanjay Jha’s recent public statements about suing Android manufacturers over IP were probably part of the negotiations.

The recent escalation of the tech patent war was quite a boon for MMI. Ars Technica estimates that Google could have bought MMI last year for $6B less.

Android

What does this mean for Android?¹ Google has somewhat contradictory things to say about the acquisition.

“Together, we will create amazing user experiences…”

And:

“We will run Motorola as a separate business.”

While maybe not directly self-contradictory, the announcement of the deal exposes the inherent cognitive dissonance in being a company that both makes hardware and licenses its software to other hardware makers. There are directly conflicting incentives set up now, which can’t help but cause friction.

One rising sentiment is that this is good because now that Google owns the software and hardware, they can make a vertically integrated product of much higher quality, able to compete more directly with Apple. But it’s not that simple. Apple products aren’t good just because they’re vertically integrated. And Google has no experience making vertical integrated products.

It’s also misaligned with Google’s ultimate incentive of commoditizing everything that comes between a user and the Web.

For the other Android manufacturers, HTC and Samsung being the most relevant, it’s a bit complicated.² On the one hand, it must certainly worry them for Google to be joining the hardware business. On the other, the growing litigiousness of the industry was starting to make Android licenses more costly both in terms of IP royalties and the litigation itself. So Google joining the patent fight on the side of Android may be a welcome development. Seems like a net wash to me. And it’s not like there are many alternatives to Android, anyway. Windows Phone 7 is faltering in the market, while Android has become the leader in terms of market share.

Patents

Ultimately, Google is only in this position because of patents. Just in the last few weeks, about $17B have changed hands over tech patents. That’s a lot of billion dollars.

That’s money that could have gone into research and development. Not that all of that money just disappeared, but there must be a significant transactional cost. It’s also that much more money that’s now in some sense ‘invested’ in the patent system. If the patent system were to be abolished immediately, all that would immediately become a wasted investment. As the industry players continue to pour money into patents, the system only becomes more entrenched.

Which puts Google in an especially crappy position. Google believes patents are a tax on innovation. And Google wants to encourage innovation, if only because that further eliminates barriers between users and the Web. But in the current environment, Google can’t not play the game and hope to get anywhere. So it has to work both sides. It has to acquire patents to defend against patent litigation, while at the same time fighting on the PR front and lobbying in Washington for patent reform.

Not only that, but Google handicaps itself by only using patents defensively. To my knowledge, Google has never used patents offensively. (Yet.) You can interpret that as self-interest or as admirable ideological consistency. I see it as both.

And further, the mechanics of the patent system favor large companies with large patent portfolios, because it both gives them the ability to sue new upstart companies for patent infringement and protects them from small companies with ‘legitimate’ patent suits. Incumbents love the patent system. Since Google’s business model disproportionately benefits from universally increased innovation, it alone among big, wealthy, influential companies has any incentive to fight for patent reform.

It’s Google versus the World on two fronts.

World War G

On the surface it seems like a good deal for Google. But I think it’s a sign of bad things coming for Google, for innovation, and for the world. I think what people lose sight of is that some (breathtaking, well crafted, seamlessly designed) innovations change the world for rich people (that’s you and me). Others simply change the world.

A class of people this is great for? Lawyers.

In the short term, this is probably good for Android users. As a Motorola Droid X owner, I’ve had to go to great lengths to remove Motorola’s frankly terrible customizations of Android. It’s easy to imagine that Google will nix all that and maybe even use the acquisition to push the entire industry into a less fragmented ecosystem. ↩
Their public statements are basically meaningless PR speak, which journalists seem to be interpreting according to their respective prejudices. I don’t think there’s any information you can really get out of them. ↩

Google's Existential Crisis

Wed, 13 Jul 2011 16:24:16 -0700

Google’s success was founded on the execution¹ of two innovations: 1. Leveraging the hyperlink to create the best index (i.e. crowd-sourcing); and 2. Targeting advertising to search results to monetize that index. Apparent diversification into new areas like content (YouTube) and operating systems (Android and Chrome OS) can be understood to exist in service to the two founding innovations. These driving principles are Google’s DNA, and as successful as the company has been, if there are fundamental problems with these two principles, it could mean trouble for the company, and trouble for the web at large, to the extent that Google is good for it.

The Algorithm

The search engine’s ranking algorithm, PageRank, measures ‘relevancy.’ It began with the simple insight that the more links there are to a particular page, the more important it is, but the algorithm very quickly became increasingly sophisticated, and it now considers “more than 500 million variables and 2 billion terms.”

Google’s relationship with the web is symbiotic. The web became useable in no small part because Google made it possible to find relevant things. As the web became more important, so did Google. And as Google became more important, it also became a bigger target. Search engine optimization became a prominent and diverse industry that ranges from qualified professionals exhorting legitimate and sanctioned best practices to cargo cultists offering opaque rituals to increase search result rankings. Orthogonally, companies like Demand Media began to mine real-time search data and hire cheap labor to create abundant shallow content, exploiting holes in the web to make money on ads. Recently, it even seemed like such ‘content farms’ could rise to become an existential threat to Google by degrading search results into meaninglessness.

This shows Google’s symbiosis with the web to be profoundly weird in a Heisenbergian way: Google cannot observe the web without changing it. Search engine optimizers and content farms both respond to Google’s algorithm. SEO changes the way publishers create and present data, as well as what content they produce. Content farms form a more directly parasitic relationship, because as search traffic data exposes demand for content, they produce it. But as Google’s relevancy algorithm increases the measure for quality, the quality of content changes to fit those criteria. This puts Google in the strange position of being able to invoke content by defining what is of high quality. Which presents a problem analogous to the one of standardized testing: it shifts incentives to ‘teach to the test’ instead of to teach what’s important. The simple solution is that if you can make a test good enough, the incentives will align properly.²

As the test-maker, Google becomes in no small way the arbiter of truth. The tension that this forms between Google and the content farm industry describes a sort of Turing arms race: the growing ability of content farms to cheaply produce content in an increasingly automatic fashion pushes against the escalating precision of Google’s quality algorithm. The inevitable conclusion of this process is automatically generated, demand-invoked content that’s indistinguishable or better than anything people can provide.

Not only that, but in December 2009, Google began to personalize search results according to users’ locations, search histories, social graphs, etc. As the search profile becomes more extensive, the personal search algorithm becomes more intensely personal, and an ‘objective,’ One True Search ceases to exist.

As I’ve discussed before, Google’s incentives are such that it wants to increase use of the web, broadly. This is good, as long as the web is good, however you define that. But as its power to create the web (or at least influence it) waxes, so does its responsibility. It’s not hard to imagine a dystopian future where the entire web devolves into a broad, shallowly addictive ad-ridden hell (this already describes some sectors of the web). Google would have to have an extremely long-term outlook to decide to serve as superego instead of as id to the web. But maybe that’s not so hard to imagine either.

Google famously makes the vast majority of its money on advertising, specifically with its AdWords program, which auctions keywords to advertisers. The key to the program’s success is its highly targeted nature, i.e. the relevance of the advertisements.³ The weird tension here is that the quality of the search results is in conflict with the quality of search ads. If the search results are perfect, and exactly what the user wants every time, then there’s no need for her to ever bother with advertisements. Conversely, if the product being advertised is maximally relevant, then it should be her first search result anyway. Advertisements only offer value to a user if their relevance has parity with search’s. But they only offer value to an advertiser if they are exposed where they wouldn’t be organically, i.e. where they are irrelevant, or distracting.

The other tension is one that stems from something I briefly touched on in another post: Google wants to encourage quality in the web and provide a good user experience so that people continue to use the web and its services. Simultaneously, Google wants to extract the most value from the user by exposing her to the most ads, i.e. distracting her. Minimally distracting ads would be nonexistent ones and maximally useful services would be ad-free. Conversely, maximally value-extracting services would be all ads all the time. The two incentives of quality and distraction are directly opposed. The result of this cognitive dissonance is what we see when we use Google products: generally tasteful, useful, and understated ads.

So when Google places relevant ads, advertisers are in some sense paying Google simply to do its job. But when Google places irrelevant ads, it undermines the quality of the web, its search results, and its user experience. And maybe the scariest part of this to me is that the quality incentive is one conceptual remove farther than the distraction one.

Maybe it’s no more perverse than any company’s tension between making the highest quality product in order to invoke desire and spending the least amount of money possible in its production. And maybe Google is culturally constituted to thinking in uniquely long-term ways that ultimately make everyone better off. The recent battle with SEO-driven content farms that resulted in the ‘Panda’ update to the search algorithm would seem to indicate so.⁴ But I imagine it’s easy to think long-term when you’re succeeding. We’ll see what happens when the success inevitably slows.

Google’s ranking algorithm was based on earlier ideas: citation analysis and Hyperlink-Induced Topic Search (HITS). ↩
I’m not competent to speak to whether that’s possible in either case. ↩
The auction part, which helps Google maximize its profits, is also really important, but I would say secondarily so. ↩
More cynically, you could call it a response to growing criticism about search quality. ↩

A New Metaphor for Social Networking

Wed, 29 Jun 2011 22:21:00 -0700

Yesterday, Google launched Google+, which, according to Steven Levy’s Wired article, is less of a product and more of a new direction for the search company, a redefinition of its approach to its core mission of “organizing the world’s information.” Treating a social networking product as an extension of this mission is simultaneously reassuring and slightly creepy. Creepy because of the way Google constitutionally seems to treat people as just another category of information to be indexed, and, well, reassuring for the same reason, because it seems like a way for Google to finally get it right.

There don’t appear to be any truly new features included in Google+. Instead, there’s a clean and uncharacteristically whimsical user interface to a core set of features that many social networks have done before: sharing, photos, group chat, group video chat, etc. One word that’s conspicuously missing from the announcement, though: “open.” No mention of open source, APIs, protocols, federation, or interoperability. Which is sad, but perhaps understandable after the abject failures of Wave and Buzz.¹

The feature I’m most conflicted about is “Circles.” Circles is Google+’s take on friend/list management, and works a lot like Facebook’s “Groups.” Where Facebook’s feature is a tacked-on afterthought, though, reportedly ranging from barely useable to horribly broken, Circles, as a tightly integrated core feature, looks a lot friendlier, faster, and more fun to use. What’s conflicting about it is that lists, while the most obvious and seemingly flexible solution to the problem of selective interaction, are actually quite limited and deceptively complex.

Two Problems: Ambiguity and Asymmetry

It seems straightforward enough. Jill has different lists. When she shares something on one list, everyone on that list can see the item. Simple. It quickly gets more complicated, though. For example, what if Jack wants to comment on that item? Who can see the comment? In Google Reader, anyone who was in the original list can see the comment. Which is potentially confusing for Jack, who has no idea who can see his comment, and for Jim, who can suddenly see a comment from someone he doesn’t know on an item in his feed. This is normal on the open web, where there is an explicit expectation of total publicity. But in a social networking context of commenting on a friend’s item, it can easily feel like overexposure. An alternative way to do it might be for only Jill to be able to see comments on her shared items, but this would simply resolve into many one-on-one conversations between Jill and anyone who comments on her item. A third way to handle this would be to only expose Jack’s comments to people in Jack’s friend list. But which of Jack’s lists? And should it be only the overlap between Jack’s list and the original list Jill shared it to?

Straightforward lists start out simple, but once you make them bi-directional, quickly gain an order of complexity. And not only are the individual ways to do it somewhat hard to define and explicate, but so is even the question of which of these is in use. The way Facebook Groups and Google+ Circles address the issue is that everyone involved can check to see who is on the list of people included in the share of the particular item. Which may seem like it solves the problem, and perhaps does to some degree, but it fails to address what the issue really is, which is social context. An example: Jack is friends with Jill both at work and outside of work. Jack shares something to his “Work” Circle. Jill sees the item, but mistakenly thinking she’s in a non-work context, makes a crude comment. While this system does give you access to the current social context, it doesn’t make it immediately obvious. I.e. Jill could conceivably check to see who will be able to see her comment each time she makes a comment, but the cognitive load and general uncertainty will push Jill to undershare, rather than have to manually check the social context each time she makes a comment.

It also does not address the problem of asymmetry between my circles and my friends’ circles. Let’s say Jill logs onto Google+, clicks on her “Friends” Circle, and begins to peruse her friends’ posts there. When she clicks on one of Jack’s shared items to comment on it, she might have the expectation that her comment is going out to the Circle that she’s in at the moment, which is different from the Circle that Jack posted to in the first place. This expectation is reinforced by the fact that in her user interface, she’s still in her Friends Circle. But her comment will actually be going to Jack’s, of course. Which is potentially pretty confusing. The fundamental problem here being output lists are different from input lists, but they’re called the same thing, and the conceptual difference isn’t easy to keep track of.

Ultimately, the problem is that the way information flows in any of these situations is non-obvious, complex, hard to model coherently in a simple user interface, and there’s no metaphor that you can map the interaction onto.

“Places”

Here’s a better idea than Groups or Circles: Instead of or in addition to lists, imagine a Place that represents a semi-permanent grouping that’s easy to create and archive, centered around anything from topic to group to actual physical location or event. A Place could be public or private² (meaning no one outside of the Place could see what happens there), and could be closed to or open for anyone to join.³ The metaphor solves both problems of ambiguity and asymmetry, by transparently conflating the input and output lists.

The primary way to use this sort of system would be to visit each Place serially as separate feeds, such that the user would have a clear idea of “where” she is as well as in what social context.⁴

So when Jill clicks on her “Work” Place, she always knows more or less exactly who’s there and, maybe more importantly, she understands the social context. Likewise, Jack also always knows this, and when Jack comments on Jill’s post in Work, he goes to the same place in his UI (Work) that Jill went in hers to make the original post, which eliminates the confusion about how the information flows. It’s also the same list of people for both of them (and everyone else in the Place), so there’s no confusion about who’s going to see the post or the comment; it’s always obvious.

Not that it would always be obvious. If Jill is a part of Jack’s “Bar People” Place, but doesn’t really know everyone else there, she’ll have to check to see who’s on the list. But importantly, she won’t necessarily have to in order to understand immediately what the social context is, and to have a general idea of how familiar she is with the Place.

Note that I am not proposing anything functionally very different from the list systems that Facebook and Google+ already have in place. Places are just slightly more restrictive and present the information in a different way. They give the user a mental model that makes it much easier to immediately and intuitively understand how her information flows, who or at least what kind of people are involved in the conversation, and what sort of tone would be appropriate.

Places also add conceptual use-cases, like event-centered streams. Imagine Jill creates a Place called “Birthday Party” and invites everyone who will be there. Then all the planning, conversation, and photography can take place there, automatically aggregating all the relevant data and simultaneously creating a record of the event in an organic way. This could easily scale to even very large events, including conferences, classes, and concerts. It’s hashtag 2.0.

While I had originally been thinking about this concept in the context of open, interoperable social networks and self-hosted identities and social graphs, where current social networks would be what I call Places (Twitter for shouting pithy aphorisms, LinkedIn for ‘professional networking,’ Flickr for sharing photos, Facebook for… I’m not sure), I thought it could be illuminating to see how well it would apply inside of a social network. Pretty well, it would seem.

But, really, another (albeit prettier) data silo doesn’t do much for me. The fact that it’s from Google makes me immediately interested, but I’m disappointed that it’s not a more open initiative. I’m still hopeful that that part is in the works.

Though I really wouldn’t attribute the failures of those products to their ‘open’ nature. ↩
It seems important to me to give clear indication of the privacy setting in each Place in the UI, perhaps by changing the color scheme: e.g. dark background with light text for a private Place, vice versa for a public one. ↩
I leave up to the reader to think about who sets the privacy setting, whether and how the privacy setting can change, who can invite people, who can control who can invite people, who can have write access as well as read access, and who controls that, and the different mechanisms that would have to be involved for each of these decisions. I’d like to point out here that Google+ seems to subscribe to some of the basic content relationships I wrote about here. ↩
It’s not that hard to imagine an alternate view that collapses all the different Place streams together, but gives a clear indication of which Place each individual post belongs to. ↩

Project: Depth

Tue, 14 Jun 2011 19:30:32 -0700

I don’t particularly enjoy traditional social settings like bars or parties. Partly this is because I can’t hear very well, especially when there’s a lot of background noise. But probably a more important reason is that I feel like a lot of the conversation that happens in these settings isn’t productive. Not productive in the sense of making me money or advancing civilization, but in the sense of making us understand each other better on a more than superficial level, or in the sense of enlightening us or engaging us to some significant depth. From my perspective, most of the idle chatter that happens in these settings are basically overhead. Not all of it. And I understand that some people prefer superficial conversation, so it’s not overhead for them. Some people enjoy eating bread by itself; to me, bread is for holding meat.

I prefer and seek depth in my endeavors. I’m starting to think it’s my overarching project. I think it’s what all my various little projects have been about: Breakfast, the various incarnations of Since the War, the party chat, NMBC, EMDN, and now EMSN.¹ It’s also what motivates my thoughts on social networks, browsers, etc, as well as on user interfaces and the rationality engine.

Depth is my project.

And since this blog is basically a brain-dump, and that’s what I spend most of my idle cycles on, I guess that’s mostly what this blog’s about.

There are two separate vectors of depth that I think about:

Depth of experience. I refer to the immersive experience, in which you are so engrossed in your artifact that when you are interrupted, it’s as disorienting as waking up in an unfamiliar place from a vivid dream. This experience can come in many forms. For me, it’s usually reading (novels, essays, and long form journalism), but often films and sometimes video games (Ico comes to mind).
Depth of connection. Conversations can be engrossing, too. They can also enlighten (sometimes mutually), expand (in the transitive sense of the word), and increase intimacy. This is true both off- and online. And online, the internet equivalent of idle chatter is shallow interaction—snarky comments, likes, pithy aphorisms, how-to articles, etc. (This is not to say that shallow interaction has no value. It does², but again, I like meat.)

There are oceans to explore, but we stay in puddles.

The problem of the proliferation of shallow communication is more fundamental than the implementation of currently popular social networks. I’m starting to see it as three parts:

Human nature. People have short attention spans and are easily distracted. They are also incredibly susceptible to Skinnerian novelty-seeking behavior. (Also, many or most people genuinely prefer shallow breadth, or even just shallows.)
The architecture of the web. The web is made of hyperlinks. Any given page generally points to many other pages. If you imagine that the goal of reading a web page is to read it all the way through from start to finish, the page itself fights you by giving you many orthogonal avenues along the way.³
Currently viable online business models. Web pages make money by distracting you, either by getting you to look at an ad or by getting you to click on one. As long as advertising drives the web, and advertisers measure success in page views, this isn’t going to change.

Look forward, not backward; upward, not forward.

There are huge, Lovecraftian commercial forces at work, with a vested interest in keeping our attention spans short, and our feedback loops shorter. These forces feed on ‘eyeballs’ and ‘clicks’ and measure us in aggregate. It seems unlikely that we can change that.

But there are countervailing forces. Some commercial ones, like Flickr and Stack Exchange,⁴ harness the power of communities to dig deep mines of richness on the web. Others are non-commercial, like Wikipedia. Many come from people scratching their own itches: Readability and Instapaper, made by and for people who wanted to be able to read web writing without the now conventional distractions of modern web design; any number of anti-procrastination applications that block your internet connection for specified periods of time to improve task focus; and services like longform.org and delivereads that encourage deep reading of long form journalism. Physical devices that enable depth are becoming wildly popular: the Kindle is an obvious example, as is the iPad, which, despite my misgivings, enables an unprecedented immersion experience by presenting an intuitive and emotionally satisfying interaction model. And of course there are others thinking about the same problems.

We are all naturally insular to some degree. We are all anxious, novelty-seeking apes. We are buoyant, in the working metaphor. And there are two ways to dive deeper: learn to hold our breath longer; and make tools to so we don’t have to. Can’t hurt to try both.

I’m not trying to take all the credit for all of these, just explicating my underlying reason for participating. ↩
I find the concept of “ambient intimacy” to be an interesting one. ↩
There’s a lot more here, like the fact that the internet is basically a custom tailored novelty aggregator and interruption machine, but others have covered this exhaustively, I think. Probably the canonical example is Nicholas Carr’s The Shallows: What the Internet is Doing to Our Brains. ↩
Probably not deep on either of the vectors I’ve enumerated, but Stack Exchange was built as a direct response to the shallow cesspool of technical knowledge that was available on the internet. It maybe represents a third vector: depth of information? ↩

The Personal Cloud

Mon, 06 Jun 2011 23:38:00 -0700

Things seem inexorably trending toward cloud models of computing.¹ Which is great, right? You can have access to your data, apps, accounts, and settings from anywhere. Somebody else does all the administration, including security updates, backups, etc. And it all only gets better as the internet does, i.e. as connections get more ubiquitous and more ubiquitously faster.

But if these things are cyclical, what does it look like when the pendulum swings back? Certainly there are reasons to want it to. There are the obvious privacy implications, which I don’t actually care that much about. But there are also related legal issues, like law enforcement agencies getting access to your data through the cloud hosting services you use. It also creates a single point of failure for your entire web life: your login. There are numerous horror stories of people’s accounts being hacked or revoked, featuring Kafkaesque experiences of trying to prove account ownership, or even just trying to find someone to prove ownership to. All of these basically boil down to: the centralization of information, and the relinquishment of control to the centralized authorities.

Carry Capacity

I spent approximately 75% of my enlistment in the Marine Corps complaining. A lot of that time I was complaining about how much gear we had to carry. It was a ridiculous amount (I would say to anyone within earshot) just barely not enough to completely immobilize us. With all of modern technology at our disposal, why was it that we had to carry so much weight?

What I didn’t realize at the time (Nick pointed out to me later) was that the amount of weight that soldiers have carried has stayed pretty much constant since soldiering first became a thing. The range and capabilities of an individual soldier has increased as technology made gear lighter, but the fundamental point is that you still want to pack a given soldier with as much stuff as he can carry.

Consumer computer hardware has historically worked in a similar way. People have always spent generally the same amount of money on computers, and as computers got faster and cheaper, we bought better computers with the same amount of money, because they were never really fast enough. But at some point, they sort of did get fast enough and we stopped spending as much as we possibly could on computers. It was as if soldiering technology got so good that a soldier’s range and capabilities were unlimited. You didn’t have to load him up to capacity any more.

So we started spending some of that money on mobile devices, which have different constraints like size and power consumption, that make the calculus of capacity limits very different. But the general principle remains, and at some point in the near future, mobile devices are going to be fast enough that we won’t have to spend the maximum available resources for acceptable performance, especially as cloud computing becomes dominant and presents different bottlenecks (mostly wireless bandwidth).

Eating and Having Cake

If the pendulum is going to swing back toward decentralization, then, the imminent computing paradigm has to offer the best of personal computing and cloud computing: the personal cloud. The term is kind of nebulous right now, with some vendors using it to refer to their web-accessible network storage devices², and others confusingly using it to refer to personal use of cloud services. What I have in mind might be better described as the self-hosted cloud.

Under personal computing, the canonical data lives on the client device (because there basically was no other place for it). Under cloud computing, the canonical data lives on the hosting service’s servers. Your Gmail data, for example, lives on Google’s servers, and it gets synced to your browser or your mail client. When you take an action like reading or sending mail, your client sends that information to the server, the server takes the action and updates itself, then sends the new canonical data to all of the clients. Under the personal cloud, canonical data will live on a server you own, that sits in your home or office. To access the data, you simply go to your home server’s url from anywhere with web access and log in. All the convenience, full control.

More interestingly, you’ll be able to install applications to your server, ideally with one click from the web. A webmail service, for example, that you have complete control over and access to from anywhere, including APIs for access from client devices and third-party services. Calendars, notes, photos, music, movies, etc. with granular control over access. Publishing platforms are basically already there. Office productivity apps are too, more or less.

The operating system and applications could be silently and automatically updated, like Chrome/Chrome OS. I would imagine that home servers would be modular enough that to add capacity, you could simply buy more, faster, larger-capacity boxes and add them to the network. You might still use proper cloud services, but probably strictly for encrypted backups of data that the hosts have no access to. No more privacy/legal concerns. No more account revocation/customer service nightmare.

The implications of widespread adoption of this model are interesting to think about. I’m not sure how the trade-off math would work for simple web apps. On one hand, the infrastructure needs are reduced to distribution of the application; no more user data, server administration, or even much user account administration. On the other, maybe more end user support? More profoundly, this would make decentralized, interoperable social networks much more viable. You’d have all of your canonical social data on your own server, with full ownership of your social graph, and granular control over who gets access to what. It would also make spinning up and spinning down of private or transient social networks much more viable, since everyone has a server.

Apple or Google? Neither.

There are already approximations of this model in existence. Some people VNC or ssh into remote machines to work using their device effectively as a thin client. But the user experience would obviously have to be a lot easier and more seamless for this model to be at all viable. It would have to be at least as easily deployable as a regular Windows PC is today, and I think the necessary client and server technology is only now becoming fast/cheap enough for this to be really productized for the consumer.

Apple does seem to be gesturing in this direction with their release of Mac OS X Lion Server for $50. But Apple’s core competency is in the narrow user experience of the robust client and, again, the company only seems to be interested in using cloud technologies to make that experience as great as possible.

Google on the other hand, wants to commoditize everything between the user and the web, more specifically, between the user and its own products, where they can expose the user to ads. These products, like free Android and ChromeOS on cheap devices to use Gmail and Google Docs and other web services means no one has to pay for expensive software suites anymore.

Microsoft got double-disrupted by Google and Apple. Apple’s invention (basically) of a viable mobile computing market means people barely even need computers anymore.³ As everything, including enterprise, goes web and mobile, Microsoft is getting squeezed out of the market. It’s going to take a while, but it’s happening. They’re certainly still profitable and will be for a while, but the market trends are pretty clear.

As Nick points out, Microsoft is moving toward a full cloud business, Azure, with Windows 8 as a stopgap. But the model isn’t Microsoft’s core competency the way it is Google’s (or even Amazon’s), and it seems unlikely that Microsoft will be able to price their services competitively. But maybe I’m totally wrong about that.

But selling and supporting home server software on commodity hardware seems exactly like what Microsoft has always been good at. They’ve already been doing something very similar in the enterprise market for a long time, and they have plenty of experience in the consumer market (unlike, for example, Oracle).

If Apple has a flaw, it’s a neglect of the web. If Google has a flaw, it’s over-centralization. I think Microsoft has a chance at disrupting its disruptors, at riding the pendulum swing to long-term relevance in the computing industry.

Even Apple joined in yesterday by unveiling their “iCloud” service, which is an interesting take on the cloud: peculiarly (but perhaps predictably) device-centric, in an increasingly web-centric world. I.e. iCloud leverages cloud technology to make your experience with Apple devices much more seamless, but it doesn’t offer access to cloud data from the web or enable any sort of collaboration, at least as far as I can tell. “Google’s frame is the browser window. Apple’s frame is the screen,” as John Gruber put it, rather succinctly. ↩
E.g. Tonido, Iomega, and Pogoplug. ↩
With the addition of over-the-air syncing, OS activation, and updates, this is now more true than ever. ↩

Windows 8

Sat, 04 Jun 2011 18:03:00 -0700

Microsoft unveiled a demo of their next version of Windows on Wednesday, and it appears that they’re doubling down on Windows Phone 7’s “Metro” user interface. Which is great, because it (the UI) is the most novel, useful, and genuinely interesting thing Microsoft has done in the consumer space in a long time. And it’s a surprisingly and encouragingly risky move, considering Windows Phone 7’s limited success so far.

There are some things I really like about the Metro UI. I love its departure from the WIMP paradigm, and in a direction other than the grid of icons. The idea of rich, interactive tiles is great,¹ as is the idea of making shortcuts to discrete functions of apps as opposed to the apps themselves and the concept of “hubs” that collect data from across applicable apps.

Some of the ideas in the demo that extend Metro for Windows 8 are fantastic, too. The implementation of multi-tasking (actually having two apps on the screen at once, not running background apps) is basically what I’ve been wanting for a while.

Maybe the most interesting part, though, was the exposure of legacy Windows 7 programs running alongside the new Windows 8 interface. While it’s superficially jarring to see two such different UI styles juxtaposed, the really interesting idea that it presents is a convertible device that exposes an intuitive, simplified, pared-down interface in one orientation, and another more robust, complex, and capable interface in the other.

The obvious application for this is a tablet that you can use as an iPad-like consumption device for casual use that you can dock to a mouse and keyboard for “real work” and maintain all the functionality of a PC. There are some instances of something like this already: tablet/laptop convertibles that dual boot into Android and Windows 7; as well as the Motorola Atrix, an Android phone that you can dock into a laptop-like terminal device and boots into a linux-based web browser environment. Windows 8, though, forks the abstraction at a different level, which makes it much more compelling by making it almost a matter of separating function and data from presentation. Ideally, the two modes would offer potentially the exact same data and (selective) functionality in two separate but consistent experiences, instead of the weird split-brain experiences of current offerings.²

In much the same way that the Metro UI’s introduction answered the question of mobile interface in a refreshingly novel way, this dualistic (Platonic?) paradigm skirts the conceit of the iPad, and addresses the inherent compromises in a qualitatively different manner. While the iPad manages expectations by offering a completely separate and different experience that’s tailored to the device, Windows 8 offers full functionality but in a crucially context-sensitive fashion.³

It’s a really fascinating idea, but Microsoft has a very difficult execution ahead of them. Unless they’re prepared to seriously rethink the dominant mode of robust input,⁴ they’re going to have to figure out a way to seamlessly integrate and reconcile a fundamentally touch-based direct-interaction interface with a fundamentally more abstract, indirectly manipulated one. Additionally, they’ll have to address whether developers are going to have to create two separate interfaces to all of their apps.

This exposes a more important and fundamental challenge, though. The problem that Microsoft appears to be trying to solve with Windows 8 is that of being able to take all of your computing with you wherever you go, and have the full experience available to you, and its solution is predicated on a particular confluence of technologies becoming mature and robust enough for this to be feasible. The implementation of iOS was similarly predicated on battery life and low power consumption processors finally becoming powerful enough to run a serious operating system on. But as Moore’s Law’s continuing influence makes hardware/power/bandwidth even cheaper, another solution to this problem that’s being worked in parallel is becoming even more feasible, interesting, capable, and solves another set of problems at the same time: the cloud.

Which renders a lot of this dicussion moot. As web technologies become more powerful and users realize more and more that they don’t need everything that traditional desktop applications have to offer, more of the things that sit between the user and the browser becomes overhead.⁵ After the web hits a certain level of robustness, these differences between operating system user interfaces are ones simply of chrome preference. They become commodities.

To reiterate: I like a lot of what Microsoft is trying to do with Windows 8, and I applaud their direction of energy and ambition. But it seems like a solidly middle-term goal that doesn’t relate to a long-term goal in any real way. Maybe it’s a great way to maximize productivity and extract value from a technological valley, but it in no way addresses the mountain on the other side, which is steep and high enough to poke through the clouds.

They are, however, not exactly original. Android’s widgets offer the same utility, and more. Metro’s solution is considerably more elegant and unifying, which brings up something of a UI paradox: It looks so much cleaner and prettier when everything is consistent and uniform, but it does make it that much harder to differentiate/distinguish the different tiles/icons, especially quickly. ↩
I imagine even the desktop mode and desktop applications will get a UI revamp to make them more Metro-y, aesthetically at the very least. ↩
As John Gruber notes, the touch-based, direct-manipulation UI model is only one part of how iOS simplifies the user experience. Other ways include: hiding the file system, single-tasking full-screen apps, and platform-specific apps built from the ground up with the specific interface in mind. (Gruber also notes other differences like battery life that I don’t think relate to the discussion at hand.) ↩
Even if they do, how will legacy apps integrate into the new system? It’s an important question because legacy support is one of the reasons Microsoft can’t just do a complete reboot or a completely separate operating system for tablets or simply extend Windows Phone 7 for tablets. ↩
That’s obviously a sort of Google-centric way of seeing things, but even Microsoft acknowledges the trend with its own web-based offerings of Office, and with its announcement that Windows 8 development will be done in HTML and Javascript. ↩

Rat Holes

Mon, 16 May 2011 21:18:00 -0700

Rat holes I’ve fallen into:

Programming Languages.
Programming books.
Text editors.
Keyboards.
Version control systems (Git).
Password management.
Backup.

Rat holes I’ve got one foot in, but haven’t fallen all the way into, but might still:

Web stuff. (Still sort of have one foot in this one.)¹
Linux/Unix.

I eventually settled on Java, so I can work on Android. I eventually settled on buying a specific Java book. I eventually settled on Vim. I eventually settled on a keyboard (this is a minor one). I eventually settled on a version control system (and I don’t even have a real use for one yet).

Rat holes are things that I think I have to think about before continuing my main project (learning to make software). They are probably not things that are necessary to continue my project, but I fall into them anyway. Some of them take a really long time to decide on because they are often in some way religious or philosophical in nature and the amount of research to be done is insanely deep and sometimes leads in circles.

A positive, and partially true way of thinking about some of these rat holes is that I want to use the best tools available to me, and I want to learn to use these tools well before I start making something with them. Chabo described it to me as depth-first traversal as opposed to breadth-first traversal.

I think obsessive-compulsive completionism is one factor. The truth, though, is probably that it’s easier and more fun to read about and learn about this stuff than it is to get about the business of learning to program. It’s peripherally useful procrastination, which is potentially more harmful than plain goofing off because it can feel productive because solving these more immediate concerns are more deeply satisfying than making relatively small progress on the big project.

There is some value in rat holes. If you spend all day using a keyboard, you should have a good one, just like if you spend all day sitting, you should have a good chair; the value proposition simply dictates it. The hard part is knowing how much value is in a given rat hole, and how much work it will take to extract that value. But for someone new to a domain, there isn’t a really good way to gauge the value proposition, because the research necessary is part of the work of extraction and the uncertainty of that position and the fear of path dependence can be paralyzing.

Some people seem to be more or less immune to rat holes. Maybe the distinction is simply between constitutionally pragmatic people and constitutionally idealistic people. Or maybe it’s a distinction between short-term thinking people and long-term thinking people. Either way, the immune ones tend also not to worry much about what kind of keyboard they have, or which text editor or IDE to use.²

Rat hole avoidance heuristic:

Be aware of the phenomenon.
Self-monitor. Periodically ask myself if I am entering or have entered a rat hole.
Determine whether I am in a blocking or non-blocking rat hole. (Can I delay this decision until I know more about the domain?)
If non-blocking, skip. Else do only enough research to make a quick decision, and only invest enough time to not be blocked and move on. Remind myself that I can change my mind later.
Periodically ask myself whether and how this rat hole advances the big project.

I’d like to move my blog to a possibly self-hosted statically served CMS like Jekyll. Why? Good question. ↩
I think this is especially true where it pertains to aesthetics. Generally, I think immune ones value function much higher than form, and don’t worry about what their hardware or software looks like. ↩

Holding the Web in Your Hands

Thu, 12 May 2011 21:59:00 -0700

I’m typing this on a Cr-48, Google’s pilot hardware for Chrome OS. On Wednesday at the second keynote address of Google I/O, their developer conference, Google announced that final retail hardware will be available this Summer from Samsung and Acer. The day before, at the first keynote, they announced Android Ice Cream Sandwich (no version number yet), which will unify the OS across phones, tablets, and Google TV.

While at first glance it may seem like Android and Chrome OS target the same space, the respective tones of the two keynotes and the pricing models of the coming Chromebooks indicate the following break-down: Android for consumers; Chrome OS for enterprise.

That axis is interesting in certain respects, but it’s not the one I find most interesting: Android for touch interface; Chrome OS for separate hardware inputs.¹

Since the original iPad’s introduction, I’ve been ambivalent about the form factor. On one hand, its direct manipulation interface paradigm (and especially Apple’s groundbreaking execution of it) is intuitive, powerful, and genuinely fun. On the other, it’s restrictive: the interaction bandwidth is simply less than the venerable mouse + keyboard combination; and the metaphor itself, while making certain interactions—like swiping and pinch-to-zoom—natural, narrows the available input options, relegating anything more than the first-order gestures to hidden, basically undiscoverable gestures that bear no relation to the metaphor.²

Text input is another problem. The on-screen keyboard—even with automatic—real-time error correction, is no substitute for a real hardware keyboard.³ It’s transient and stultifying, which can’t but inform the text that passes through it to some, however vanishing, degree. The form factor itself, furthermore, ergonomically discourages text input.⁴

Technology as Philosophy

My biggest gripe about the iPad is that it’s a reversion to pre-internet media values. If implementations of technology are founded on philosophical assumptions, the old broadcast media systems like television and radio are founded on authoritarian, heirarchical ones. A (more or less) centralized, consolidated authority produced the content that the rest of us used to sit back and consume. In this respect, the personal computing and internet revolutions—and blogging, specifically—were intrinsically democratizing ones. The iPad is a betrayal/reversal of this trend, a device that’s optimized for passive consumption, the interface optimized for interaction with the device, instead of interaction with other people.⁵

The Chromebook, conversely, is an open embrace of the web.⁶⁷ It’s an instantiation of the promise of the removal of every barrier between us and a place where we can congregate, communicate, and collaborate freely, as well as an investment in a future where all of our computing is accessible from anywhere, from any device.

I submit, too, that the laptop form factor is a more democratic one. In lieu of a claustrophobic software keyboard, it dedicates half of the visible hardware to input, metaphorically elevating the user’s production to the same level of importance as her consumption. It says something about what the purpose of the device is: to participate in the global conversation, in the exchange of ideas.

This ad introducing the Chromebook asserts that it’s not a laptop, or even a computer. Instead, “it actually IS the web.” It’s a beautiful dream.

Tentative plans for a Chrome OS tablet have not come to fruition, at least not yet. ↩
Even Apple has had to resort to things like three- and four-finger swiping gestures that map to commands that have no intuitive mapping to the the gestures. I’ve written more about the limitations of a the direct manipulation paradigm here . ↩
The tactile feedback of a physical keyboard is difficult to replicate. I’d hoped that Swype or other software keyboard technologies would succeed at routing around this problem, but in my experience you still have to pay attention to both the keyboard and the text you’re producing which is a lot of mental overhead that the tactile feedback of keyboards obviates. Blind type seems like it’s close to a solution, but I haven’t gotten to try it yet. ↩
This is a big enough problem that Bev puts down the iPad and grabs her iPhone to write emails. ↩
I think this bears out, too, in Old Media’s enthusiastic embrace of the iPad as a publishing platform, despite Apple’s fairly draconian restrictions and price structure. ↩
Of course, Google, as a company, constitutionally embraces the web. ↩
I do have misgivings, though, about the “cloud” model, and giving up control of our data and identities to corporations. ↩

The 2-Foot User Interface for the 10-Foot Screen

Thu, 07 Apr 2011 23:17:00 -0700

Television user interfaces could be so much better. I thought that with the introduction of set-top box internet video solutions, user interfaces would make menu and playback navigation and content management easier, especially now that we have ways to interact with these boxes using our smartphones.¹ If anyone, I thought Apple—with their customary attention to user experience and their vertical integration—would have an elegant solution to interaction with TVs. But when I hooked up our new Apple TV2² and tested out the remote application on the iPod Touch and the iPad, all I got was a simulation of the physical remote control.³ Everything else I’ve tried is pretty much the same, to include: Boxee and Windows Media Center (couldn’t even actually find an Android or iOS app to work with this, unsurprisingly).

The 10-Foot User Interface

Used to be that all you needed was a few buttons on your remote: power toggle, channel up and down, volume up and down; plus some optionals: numbers 0–9, mute toggle, channel recall, etc. Once you got used to it, you didn’t even need to look at the remote, which was nice. But as television sets became more complex, with more options, settings, and configurations, remotes got loaded with more buttons. Peripheral devices added even more buttons, usually on a separate remote, which you could integrate into a single universal remote if you wanted to and had the technological wherewithal to do so.

At first, with simple sets, this really was fine. And even as TVs got more complicated and loaded with more functionality, you still needed a way to control all the additional functionality. There wasn’t any other place for this than the remote and there wasn’t any way to do it on the remote without adding buttons.

Meanwhile, from the other direction, desktop computer interfaces were simplified and enlarged for use on the TV screen (from 10 feet away). Window management was (naturally) discarded for a simple environment optimized for doing one thing: watching video.

As these two TV paradigms converge in devices like the Apple TV,⁴ Boxee Box, and Google TV, as well as in TVs themselves, the functionality starts to overwhelm the UI bandwidth: UI elements increasingly clutter up and obscure the screen, menu hierarchies become deeper and more complicated, and the fact is that a TV is not a computer monitor and it sucks to do computer monitor things on a TV screen.⁵

Reboot

No one’s really been in a position to be able to rethink TV UI until very recently. Even now, one cannot make the assumption that a TV watcher has a peripheral with which to interact with the TV⁶, so the system designers still have to provide an on-screen UI by default.

But I think it’s getting closer. Watching TV with a laptop, tablet, or smartphone in hand is becoming more popular, as people idly multitask, google, or participate in social media as they watch.⁷ And if you take it for granted that your user has a secondary screen, you can start to really change the way TV UI works.

As a starting point, you could relegate all UI elements to the secondary screen, leaving the primary screen completely over to what it’s specifically designed for: playing video. The secondary screen, meanwhile, because it’s a 2-foot or 1-foot UI device, can present a lot more information, and present navigation that is faster, more informationally dense, granular, and intuitive than the on-screen navigations that we’re used to.

You could also add more features, like playlists or queues,⁸ app integration,⁹ games, and multiple video streams.¹⁰

Looks like this might solve all my complaints. I’ll have to try this out.

Update

Looks like this might solve all my complaints.

And iPod Touches. ↩
Bev got me one for Valentine’s Day. ↩
At least text entry was much easier; hitting a text field would invoke the touch screen keyboard. ↩
The ATV isn’t actually a convergence device, since it doesn’t take input from traditional signals like analog broadcast, cable, or peripheral physical media players, but it sort of will be if IPTV ever becomes a reality. ↩
This is why WebTV failed, or at least why it sucked. ↩
Obviously because not everyone has one and shipping an iPod Touch equivalent as a remote is certainly cost-prohibitive. ↩
Also, some high-end universal remotes cost as much as iPod Touches. ↩
Which becomes more important as short web clips become a ubiquitous, first-class medium. ↩
Twitter with automatic hashtags for whatever you’re watching? ↩
Like picture-in-picture, but more robust. ↩

Rationality Engine

Tue, 29 Mar 2011 22:02:00 -0700

Paul Graham’s essay on disagreement, the discussion C-Rob started here, an interview with Jeff Atwood I heard recently, and a comment by Jake on GReader got me thinking about debate, though really more about how debate is sort of an implementation of rationality. Rationality, I think, is sort of self-evidently important, but not easy. Humans have natural cognitive biases that are hard to recognize, especially in oneself. Also, so much of our collective brainpower and discussion is spent/wasted on faulty arguments and corrections thereof. We could spend so much more of our cognition and time more effectively if we could cut the argumentational fat.

So, how to learn to think more rationally? I idly peruse sites like Less Wrong and Overcoming Bias and I often find the articles to be interesting, but they’re very often dry, difficult to digest, and even more difficult to apply.

I think part of the problem is that while these places (and, I’m sure, others) provide explicit/declarative knowledge, they don’t provide implicit/procedural knowledge.

Proposal: Rationality Fu¹

I would like to see a rationality engine, a collaborative, pedagogical tool to encourage rational thinking. I think it would consist of the following components:

Khan Academy-style classes to teach the explicit concepts to include:
1. Symbolic logic
2. Logical fallacies.
3. Cognitive biases.
4. Statistics/statistical analysis.
Exercises to test and enforce understanding.
A game system² to include:
1. Experience/karma and rank (or belts).
2. Achievements/badges.
A discussion engine like the one C-Rob and I discussed to include:
1. Symbolically logical statement templates in which you can plug words to make an argument.
2. Threaded conversation with configurable/visualizable views.

Thinking in terms of symbolic logic can facilitate emotional disengagement from a discussion, so that we can recognize our mistakes and biases more clearly, and concede wrongheaded points more readily. I think knowing about logical fallacies and cognitive biases doesn’t necessarily help us avoid them and I believe that thinking is action, and rationality is habit. Having a system of discussion that exposes the underlying mechanics of ideas can make our everyday thinking more clear and more precise.

Higher-performance climbing shoes can improve your climbing but getting stronger and learning technique are more pertinent and effective. Similarly, we have many tools available to us to augment our thinking, but more effective and more important would be for us to improve the generative power: thinking itself.

I think we’ve harvested the low-hanging fruit of our cognitive symbiosis with the internet. Time to climb the tree.³

Think fu? Brain fu? ↩
In the above-mentioned interview, Jeff Atwood, co-founder of Q&A site Stack Exchange talks about how a little bit of gamification encouraged the collaborative creation of an incredibly rich (the richest?) resource for programmers, both as a static repository for knowledge and as a place to get help from other users of the site. ↩
Obviously a very small subset of the population at large would be at all interested in this idea, but I think it largely overlaps the set of people who like to think about difficult things to think about, as well as the set of people who like to debate things online. ↩

Climbing Rocks

Mon, 21 Mar 2011 23:04:00 -0700

I had been to a climbing gym once or twice, but only started climbing seriously while I was stationed in Okinawa in 2005. A friend of mine, Jay, took me outdoors and got me on a top-rope. I was hooked instantly. It was exhilirating, challenging, and focused my attention both mentally and physically in a way I’d never really experienced before. Admittedly, the gear was no small part of the attraction at the beginning, and sport climbing involves a considerable amount of gear.

“Free-climbing,” or class five climbing, is defined as climbing a vertical or near-vertical face under your own strength, without the aid of special gear in your upward motion (though most people use safety gear in case they fall).¹

Free-climbing is further divided into “traditional,” or “trad,” climbing, “sport” climbing, and “bouldering.” Trad involves climbing bare rock while placing cams and nuts into cracks as protection from falling. Sport routes are generally bolted, meaning bolts and anchors are drilled into the rock at certain intervals so you can place protection.²

Like most, I started out sport climbing on top-rope, i.e. with a rope already set up at the top of a route, but Jay very quickly got me on lead by not telling me how much harder and scarier it was. Which worked pretty well. (His other trick was to refuse to let me down until I’d finished a route.) And I really enjoyed it and stuck with it. I continued to buy gear: higher-performance shoes, carabiners, quick-draws; and I was preparing to make the leap to trad, an intimidating prospect considering how much more money I’d have to spend on a rack of cams and nuts, as well as the substantial new body of expertise I’d have to absorb: protection placement.

But then a funny thing happened: I went in the opposite direction, I started bouldering. All it took was an invitation to join a couple of acquaintances from the University climbing gym on a bouldering trip to Squamish.

Bouldering, of course, is the practice of climbing relatively low, but generally much more difficult routes, or “problems” in bouldering terms. The practice started as just that: practice, something for rock-climbing pioneers to do in Yosemite Valley, between ascents of big walls, to train for strength and technique. It soon became an activity unto itself.

I had bouldered before the Squamish trip, of course, but only idly, mostly between sport routes or when I couldn’t find a partner. It had always just seemed like sport climbing, except shorter. But that trip to Squamish unlocked an understanding about the very divergent nature of bouldering for me that’s a bit difficult to articulate.

There are all the obvious differences: bouldering involves less gear, just a pair of shoes and optionally a chalk bag and crash pad versus shoes, chalk bag, rope, harness, belay device, a rack of quick-draws, a helmet, and (importantly) a partner. And as I mentioned earlier, while the equipment was initially a draw for me, I later found that it and other meta-climbing periphera like learning strategy, clipping techniques, knots, belay techniques, etc. served to distract me from the business of actual climbing.

Bouldering also generally involves more strength and intensity whereas sport climbing emphasizes endurance and economy of motion. It was bouldering that opened my eyes not only to the incredible range of what the human body is capable of, but also to a greater understanding of how seemingly unrelated factors contribute in a holistic fashion to success. Sport climbing necessitates strategic thinking, but bouldering narrows focus onto specific technique: everything from minutely detailed adjustments in foot placement,³ to frequently counter-intuitive body-positioning, to the all-important engagement of core muscles.

In these and other general ways, bouldering is a more abstract and more internally pure experience. The distinct values inherent to the separate disciplines are right there in their respective objectives. The sport climber’s goal, to scale a natural formation, is decidedly more external, because altitude is the ambition and measure of achievement, whereas the boulderer’s goal is better described as performing a somewhat arbitrary, artificial, perhaps contrived series of movements.⁴ This distinction isn’t airtight, by any means: the sport climber does not simply choose the easiest ascent of a particular crag or peak or face, and neither does the boulderer generally have completely arbitrary starting and finishing points.⁵ But the distinction remains.

Where sport climbing is a struggle to conquer nature, bouldering is a struggle to conquer one’s own weakness.

The two disciplines are complementary, to be sure; each makes you better to a large degree at the other. But bouldering teaches you more about your own body, about your personal biomechanics,⁶ and about how to most effectively wield them. And it is for me an ultimately more spiritual experience, in that for long moments at a time, instead of all the nearly omnipresent internal voices chattering about which holds to use and how far the next clip is, there is, internally, only silence, only pure, focused will, a transcendently complete collapse of mind and body.

Somewhat paradoxically, bouldering is also more social than sport climbing. Whereas sport climbers generally work in pairs, alternating between climbing and belaying, boulderers, with notable exceptions, tend to gather in groups and work problems and problem sets together, with mutual encouragement and instruction.⁷ I’ve definitely met more people bouldering than I ever did sport climbing.⁸

I’ve been climbing for about six years now, and find myself in a bit of a dilemma. As I get stronger and more capable, and climb harder problems, I put disproportionately more stress on certain joints, and my aging body⁹ meanwhile is more susceptible to injury, and needs longer recovery times. Certain knuckles are more or less permanently inflamed now, and tendons in both wrists strained. And at the same time, improvement that was steady for years has slowed to a trickle.

All of which is merely to note that the low-hanging fruit of youthful vigor and easy improvement have been harvested, and the real work is simply yet to begin. It means I have to be a better steward of my body, and that any further improvement is going to be hard-won. It means I’ll actually have to spend time on preventive weight-lifting and exercises, focus training sessions on specific strength and technical objectives, cross-train for endurance, generally pay more attention to what my body is telling me.

As with any serious endeavor (I suspect), the more I learn about climbing, the more I become aware of the humbling vastness of what I don’t know. After a period of soul-searching and asking myself if the point of climbing was to get better at climbing, or to enjoy myself, I realize, even as I write this, that the two are intertwined, and that my own struggle with weakness is only now beginning.

Class six climbing is called “aid climbing,” and involves ascenders and other tools to essentially create artifical hand- and footholds. Class four, or “scrambling,” is climbing relatively gently sloped surfaces with only the occasional need for hands.↩
More extreme (read: dangerous) forms of climbing include “free soloing,” and the recently pioneered “free basing.” Free soloing refers to climbing routes alone sans protective gear, with a fall generally resulting in serious injury or death. As far as I know, Dean Potter is the only person in the world to combine free soloing with BASE jumping in what he calls free base climbing: he free solos wearing a parachute that he can deploy if he falls. This allows him to push the limits of free soloing, extending the range of what he can attempt without any other equipment. Also, he generally jumps from the top of a route he’s climbed.↩
Both on footholds (using toes, edges, and heel-hooks) and in support positions (flagging).↩
To speak in broad generalities, in my experience, sport (and, naturally, trad) climbers tend to think and talk a lot more about nature and their relationship to it (adversarial or otherwise) than boulderers, emphasize outdoor climbing a lot more, and get bored and tired of indoor climbing much more easily.↩
Either way, it’s an absurd activity, rock-climbing, though I think bouldering is more accepting of this inherenet absurdity.↩
People have a fairly astounding variety of shapes and sizes. Height is an obvious differentiator when it comes to climbing (though, somewhat counterintuitively, excessive height is often a liability when it comes to difficult bouldering because the strength to weight ratio is a lot harder to maximize) but so are wingspan versus leg versus torso lengths, flexibility, hand/finger size, and muscle density. Two climbers of equal height and strength may have vastly different climbing styles due to different limb lengths and flexibility.↩
A friend once made the comparison to skateboarding, which I’ve never done, but the comparison definitely seems apt.↩
The majority of our Portland friends are either climbers or people we met through climbers.↩
Not that I’m that old, but I never had a single injury and could climb more energetically and for longer intervals with shorter recovery times just a couple of years ago.↩

Touch UI is not the future of everything.

Sun, 27 Feb 2011 18:47:00 -0800

Apple beat even the most optimistic expectations and sold over 15 million iPads in the first year of its release, and it looks like the second version is just around the corner. The competitors are lining up as well: the Motorola Xoom, running Android 3.0 Honeycomb, the HP PalmPad, and the RIM Playbook all look like interesting offerings.¹

Many have pointed out the sad irony of the fact that Microsoft, who has been pushing tablet PCs for over a decade, seems to have by far the least compelling offering, and no apparent serious road map for future products that address the deficiencies of desktop Windows on a tablet form factor. Historically, Microsoft’s tablet and even smartphone operating systems have been desktop Windows pared down (and barely, in the case of tablets) for use with a stylus. I had a convertible laptop, in fact, and though I was excited about the idea, it never really worked that well, for reasons I was never able to articulate.

In the aftermath of iOS, the reasons are pretty obvious.

There were many aspects of the iPhone that revolutionized and redefined the smartphone market. The iTunes and iPod ecosystem certainly helped. So did apps (though third party apps didn’t come until later, so didn’t really contribute to the initial popularity), and so did Apple’s traditional focus on hardware design. But the single largest and most revolutionary factor of the iPhone’s (and the iPod Touch’s as well as the iPad’s) success was the user interface. Apple designed the UI around a new paradigm of touch and direct manipulation, shining a light on how elegantly simple, refreshing, and intuitive human-computer interaction could be. The dominant paradigm shifted and now the measure of a smartphone UI basically boils down to how well the paradigm is executed.²

Apple’s touch interface allowed the dismissal of many of the abstractions that people found difficult about computing: Buttons took a step away from metaphor and toward the physical as did navigation and scrolling. Diminishing these abstraction-distances made the idea of transforming the device into the specific running application more feasible and more emotionally satisfying.

While it makes complete and and perfect sense to collapse the distinction between input/control and display for a mobile device like a smartphone because by their nature these devices are constrained by their form, I would argue that it makes less sense as the device gets larger. Larger devices (including, arguably, current iPad-sized tablets) are fatiguing to use for long periods of time, heavy to hold up and ergonomically unsound. Occlusion can also be a problem.³ As this UI concept video demonstrates, these problems get bigger with the device. On a real workstation, these problems become intractable.

With OS X 10.7 Lion, Apple wants to bring many of the things that make iOS appealing to the desktop. For certain aspects, this is great: saving application/document states, abstraction of the file system, simplification of application management, etc. For other aspects, the results are ambiguous: disappearing scroll bars and reversing the direction of two-finger scrolling, for example. To take the latter example, Apple is attempting to bring desktop multitouch gestures closer to the iOS counterparts. In 10.6 dragging two fingers down on the trackpad scrolls down a page. In 10.7, the same gesture scrolls up, just as would happen if you were to touch an iPhone’s screen and drag down. There isn’t necessarily a correct way to do this (it reminds me of the change of the default in mouse-look direction in first person shooters), but it’s weird to change an established convention in a way that’s not really any more intuitive. The intractable problem here is that a direct manipulation paradigm is fundamentally different from an indirect one, and trying to force one onto the other creates a bad experience.

Still other paradigms Apple is trying to bring from iOS to the desktop are just bad. Full screen applications, for example, are perfectly sensible on a handheld, size-constrained device (and maybe even on smaller laptops), but on a world of large desktop screens, it only restricts the interface artificially. It’s also really strange to see Apple embrace this concept now when it makes the least sense after so many of years of denying it completely.

Just as Microsoft tried to force the desktop’s indirect manipulation UI paradigm onto devices where a direct manipulation UI would have been more appropriate, and suffered for it, Apple seems to be trying to do the exact opposite. There’s no denying that Apple is the uncontested thought-leader in computer UI, which is why I’d be much more interested in seeing them offer new ideas in desktop UI with the very different inherent ergonomic constraints of the desktop environment in mind. The above-linked 10/GUI concept is still the most compelling I’ve seen with these considerations in mind. Especially with successive generations growing up with ubiquitous technology, I really don’t think dumbing down our interactions with computers is helpful.

Notes

I’m not particularly interested in tablets because I always feel restricted even on laptops. To me, the tablet is an even more restricted laptop, one that’s built much more for consumption than production. For a portable, I’m much more interested in the MacBook Air. ↩
Android’s primary influence before the iPhone was the Blackberry. WebOS and Windows Phone 7 are widely regarded as excellent UIs (at least conceptually), while Blackberry OS’s half-hearted attempt to convert to a touch paradigm is widely panned. ↩
Good app design can mitigate this, but not eliminate it. ↩

blog by Joe Moon

An Amateur's Lament

Visions of the Future

Production and Consumption

Distribution Control

Centralization

Lamentation

An Alternative Vision of the Future

More Reading on Apple

More Reading on Amazon

More Reading on “Post-PC”

More Reading

We Are What We Choose

In Defense of the New Google Reader

Intimacy is Performance

Intimacy is Performance

Limits

Digital Augmentation of Social Interaction

Presence

Collapse

Further Reading

Intimacy and Performance on Facebook

Timeline

Ticker and Open Graph

Intimacy vs. Performance

Further Reading

"You are the product."

Accountability

Alternatives

Incentives

Currency

Meaningless

Google's Open Hand and Closed Fist

Open

Closed

Android

AOSP

Developers

Consumers

Manufacturers and Carriers

Mostly Open

“Don’t be evil.”

A Different Kind of GUI

The Mouse

Building Blocks

Fresh Start

This is just a rough draft of an idea.

Further Reading

Google, Motorola, and Patents

Motorola

Android

Patents

World War G

Google's Existential Crisis

The Algorithm

Ads

A New Metaphor for Social Networking

Two Problems: Ambiguity and Asymmetry

“Places”

Project: Depth

There are oceans to explore, but we stay in puddles.

Look forward, not backward; upward, not forward.

The Personal Cloud

Carry Capacity

Eating and Having Cake

Apple or Google? Neither.

Windows 8

Rat Holes

Holding the Web in Your Hands

Technology as Philosophy

The 2-Foot User Interface for the 10-Foot Screen

The 10-Foot User Interface

Reboot

Update

Rationality Engine

Proposal: Rationality Fu1

Climbing Rocks

Touch UI is not the future of everything.

Notes

Proposal: Rationality Fu¹