Since yesterday’s post I’ve been haunted by the question of what will drive instant messaging vendors to adopt a unified standard. I started off convinced that there must be some new feature so compelling that it would enable one client to gain a preponderant share of the market and impose its protocol on the others. But the simple reality is that IM is very straightforward: I send a message, you receive it. Without completely changing the paradigm, no feature, no matter now appealing, is likely to cause a wholesale reorganization of the market.
When I mentioned this to Cedric he looked me straight in the eye and said, “Identity.” And he’s right. If we believe that a distributed system for managing online identity will eventually prevail, this would provide a powerful motivating force for converging on a single instant messaging protocol.
Currently the profile required by IM clients is extremely limited, consisting mainly of basic personal information like name, address, maybe a thumbnail photo. This means that the pain incurred when registering for another client is minimal. This doubtless explains why people are more inclined to use several clients simultaneously than to install a metaclient that supports multiple networks.
Now imagine that your profile is linked to a whole slew of valuable information that you have built up painstakingly over a period of time: blog, photos, reputation (for trust management), consumer profile (for recommendations), community memberships, media library, etc. To the extent that this information can be leveraged by an IM application (and I believe it can), this will make it much less attractive to move between different clients.
Of course, the big online players are already attempting to lock users in by linking their services to a rich profile, as I explained in my very first blog entry. I argued that this is unlikely to succeed unless the profile is owned and managed by the user, who choses explicitly to share it with other users and services.
This gives us a much clearer picture of the future of IM. The market is not going to consolidate until you can leverage your IM account to do a lot more than just chatting. A good example of where things are heading is Google’s acquisition of Picasa. Google’s explanation for this move, namely to facilitate uploading of photos to their Blogger service, strikes me as pure hokum. Much more convincing is the theory that Google is planning to make inroads into the IM market, leveraging Picasa’s Hello.
I’m not one of those who believe that Google is going to magically become the king of IM just because they happen to have the hot brand name du jour. But the principle is valid. At the end of the day, instant messaging is a feature, not a product, and when it starts to be encompassed into a much richer environment that is open, extensible and exploits available synergies, then we’ll find a way out of the current IM morass.
Am I the only one who’s disappointed by instant messaging? Sure, it’s a revolutionary technology that demonstrates yet again the potential of the internet to revolutionize how we communicate. It’s better than email for many purposes because messages are received in real time. It’s better than telephone because it’s non-disruptive to those around you and it’s so easy to send URLs, code snippets and other text that would be unbearably tedious to dictate out loud. And best of all, it’s totally free.
Like all network effect driven technologies, instant messaging becomes exponentially more valuable the more people are hooked into the network. Unfortunately, the IM market is horribly fragmented. AOL, Microsoft and Yahoo all have significant user bases but mutually incompatible clients. As a result, IM is still far from realizing its potential impact.
It’s pretty typical for a new technology to start off fragmented, as players jockey for position in the market, hoping to emerge as the defacto standard and reap corresponding monopoly profits. There are various ways in which the market can then evolve:
- The Email Scenario - a widely supported standard emerges before the commercial players enter the market. Unless they find a way to introduce disruptive change, new entrants have no choice but to support existing norms.
- The Windows Scenario - one player manages to dominate the market and dictate proprietary “standards”. The market benefits from the existence of a common platform but innovation can suffer since a single company retains so much control.
- The VHS Scenario - two or more companies battle for the right to impose standards on the market. For any number of reasons, one player eventually gains the upper hand and a truce is called in the interest of driving market growth.
- The 3G Scenario - governments step in and impose standards in an attempt to accelerate adoption.
As far as IM is concerned, it’s clearly too late for the Email Scenario. The same can probably be said of the Windows Scenario, and it’s hard to argue that this is a bad thing (imagine if one proprietary IM client were so dominant that Microsoft, AOL or whoever could dictate standards for the entire space). Similarly, the U.S. government could theoretically mandate a standard IM protocol, but if I’m sure of one thing it’s that markets, not governments, should be the driver behind technological decisions of this type.
This leaves the VHS Scenario. On a superficial level it seems obvious that the IM space will eventually mature to the point where the major players will see that they have more to gain from standardizing on a single protocol than from engaging in a protracted land grab to the detriment of their users. The problem is that this scenario depends largely on the profit motive. Sony ended up abandoning their Betamax format because they realized they would make more money by entering the growing VHS market than by continuing to swim against the tide. IM vendors don’t make any money directly; instead they see their products as loss leaders used to shoehorn users into proprietary services like premium webmail, financial services and digital media sales. This leaves them with little incentive to open their software.
This hypothesis is supported by the recent announcement that Microsoft, AOL and Yahoo have linked their IM services — but only for enterprise customers. Obviously these companies would happily open up their protocols if this allowed them to earn money from IM, something they have been desperate to achieve since the huge popularity of IM became apparent. Charging normal consumers for IM is not a realistic (or desirable) prospect, however, so this path doesn’t seem particularly promising.
Is there hope for taxonomies? Clay Shirky doesn’t think so, and his argument that otherwise “Yahoo would be king of the hill and Google an also-ran service” is superficially convincing. Formal taxomies certainly present a daunting challenge to layman and expert alike. How on earth am I to decide whether my holiday snaps of Kilimanjaro are best filed under Regional/Countries/Tanzania, Regional/Regions/Africa/Science/Earth_Sciences, Arts/Visual_Arts/Photography or Recreation/Travel/Photos? In fact, the very premise seems absurd. Stuff simply doesn’t fall cleanly into neat, unambiguous, hierarchically organized categories.
Hence the appeal of the cutesily named folksonomy. At its heart, a folksonomy is nothing more than the age-old “grab bag of keywords” attached to some entity to make it easier to find later. The innovation is that the keywords are shared among a community of participants, with the most popular keywords displayed prominently as a guide to users. This goes a ways towards solving one of the fundamental problems of the keyword approach: the difficulty of getting people to use the same terms for the same things. When I’m struggling to decide whether it’s a floor wax or a dessert topping, it’s very helpful to see what term others have been favoring. In any reasonably focused community, this turns out to be a huge win since the distribution of keywords will tend to follow a power law (bonus question: explain why), so a fairly small number of keywords will be used very widely.
So does that mean that folksonomies are the be-all and end-all of content classification? I certainly hope not, for two reasons:
First of all, what works in a limited domain is not always applicable to the wider world. We’re in a pioneering phase of social software and online communities, so we’re still learning what works through innovation and experimentation. Will anyone want to use Baht-or-Not, your brilliant concept for a Thai counterfeiters’ community? Try it and find out. But there’s already a push to consolidate communities like the omnipresent YASNs in order to benefit from economies of scale. Once the breadth of a community expands to the point where the number of popular keywords can no longer fit on a single screen, it’s unclear whether the folksonomy approach will be as successful.
The other reason is that the Holy Grail of all semantic classification is automated processing. Folksonomies work because they leverage a very efficient natural language processing tool: the human brain. By offloading the task of disambiguation onto the user, we reduce the need for all of those fiddly niceties like hierarchy that ontologists have traditionally considered necessary. The picture changes when I want my computer to go out and find or process content for me by itself. Suddenly a much more rigid, powerful and formal classification is called for.
This is the vision of the Semantic Web, the net’s poster child for vacuous hype. More than anything, the Semantic Web has failed to take off so far because of the difficulty of getting people to categorize information in a way that machines can interpret. This is where “googlization” is going to have a big impact. The statistical techniques that Google uses to organize and rank web pages are well understood and applicable to a broad range of tasks. I don’t see why they couldn’t be used to assign keywords to content automatically. This might not yield perfect results, but it would serve a useful purpose nonetheless by bootstrapping the Semantic Web. After all, there’s no value in being the first one to add semantic classification information to your website, but the marginal utility of adding data to a huge existing corpus of tagged content would be enormous.
When the New York Times launched its online edition in January 1996, it started a trend, albeit one that lay dormant for several years. Although free, Nytimes.com required that users fill out a registration form with personal information before accessing the website. The primary purpose of this policy was to make the website more attractive to advertisers, as explained in a February 20, 2002 presentation given at a Seybold conference by none other than Arthur Sulzberger, Jr. himself, chairman and publisher of the Times:
What advertisers are also discovering is that we can help them speak to very specific segments of our audience. For instance, one of the basic premises of NYTimes.com’s online registration philosophy is the notion of user acquisition, targeting and retention. Our proprietary decision support system allows advertisers to precisely target those categories of users who fall into the demographic and psychographic profiles most relevant to their marketing objectives.
In my recollection, the Time’s decision to require registration did not provoke many objections at the time, beyond a certain amount of passive-aggressive griping on Slashdot, where any reference to Nytimes.com is still followed obsessively by the disclaimer “registration required”. Eventually other publications followed suit, with biggies like Chicagotribune.com (March 2002) and Latimes.com (April 2002) leading the charge.
The result is a bit like a virtual tragedy of the commons, where the “commons” in this case is the willingness of the public to spend time filling out the same damn web form on every news site they visit. When it was just Nytimes.com, people were prepared to put up with the aggravation, but their precious patience was quickly used up once the Poughkeepsie Post and the Boise Bulletin jumped on the bandwagon.
Another factor is the growing popularity of news aggregation websites like Google News. I used to read one or two online papers regularly (NYT and Wired News, which has never required registration), as well as hitting the big news network or newswire sites like CNN and Reuters. Now, I just breeze over to Google News and check out which headlines sound most interesting. So I end up reading papers like the Kansas City Star and Detroit Free Press which I previously never even knew existed.
The bottom line is that the need to constantly register for this and that publication and keep track of the resulting 4 zillion passwords has become unbearable. This leads us to:
Gertner’s First Law: Any sufficiently annoying problem will be solved immediately by technology, provided that there is some way to take advantage of network effects.
File sharing networks are a good example of this. All of the successive generations (Napster, Gnutella, Kazaa, Bittorrent) have been extraordinarily sensitive to network effects, i.e. the value of the network is directly related to the number of users (or rather to the square if we believe Metcalfe’s Law). As a result of this and the tangible desire on the part of consumers to have access to an open file exchange (as opposed to subscriptions or DRM-crippled content), all of the above-mentioned systems have been stupendously successful.
With little fanfare, a website appeared last November to combat the increasingly onerous online registration system. Bugmenot.com also exploits network effects by making login credentials entered into the system available to all other users. So the more people use it, the more valuable it is. As predicted by Gertner’s First Law, its popularity has risen exponentially, with over 22,000 sites now tracked according to their homepage. There are even Bugmenot extensions for Mozilla and Internet Explorer that make bucking the system even easier.
I’ve written a longish essay about the potential of Mozilla Firefox to supplant Microsoft Windows as the standard platform for software development.
Read the essay here
I discussed in a previous entry the role of TiVo in the battle with Microsoft and Apple to become the household digital media supremo. Interesting, a recent article in Business 2.0 suggests that TiVo might be ripe for a takeover by none other than Apple itself, and not just because it has that funky capitalized letter in the middle of its name, just like NeXT did.
I don’t have much time for journalists who moot zillion dollar acquisitions from the comfort of their armchairs. Merging companies is not like sticking together Lego blocks. Many is the small, nimble startup that has been snuffed out of existence following a strategic sale to a large technology corporation. And is TiVo really a “steal” at 2+ times revenues? It all depends on whether you see it as a fast-growing, revolutionary technology startup or (like many industry pundits) as a has-been on its last legs.
Nonetheless, the article is valuable in that it leads us to ponder in more detail the movement of digital media into the home. It’s certainly true that Apple is weak where TiVo is strong. Jobs may be right to trash the notion of fusing TV with computers, but that doesn’t mean that they aren’t exquisitely complementary. This is one case where I’m more inclined to concur with the views of that other tech guy. I may prefer to locate, download and organize my media sitting upright at my PC, but I want to view my movies and photos (and select music playlists) while reclining on my couch, clutching a remote control (and preferably a strawberry daiquiri).
Hence the appeal of the iTV vision: a sexy white box that sits on your TV and mediates between your Mac and your 50″ plasma screen. And it surely wouldn’t be a bad thing if Apple were to dip its little toe into the Linux pond. After all, the enemy of my enemy is my friend, is he not? Was I in too much of a hurry to reject the notion of an Apple/TiVo hookup?
Probably not, and here’s why: TiVo is the new Napster. Yes, I know. There’s an official new Napster, duly neutered and declawed by Hollywood’s cronies. But TiVo is much closer to what the original Napster represented: a wake-up call for the media industry.
I’m not a Power to the People kind of guy. I have no problem with the fact that Hollywood executives are rich as Croesus and get invited to all the best parties (though I might be a tad envious). What bugs me is that I can’t get the content that I want, when I want it, and do with it what I will once I have it. Part of this has to do with the fact that I live in Prague, where terrestrial television carries mostly American B movies that were execrable even before they were dubbed into Czech. So even though the Wimbledon website now streams already completed matches over the internet for a small fee, to see the final live I had to go to a shady sports bar with a uniformly inedible menu and watch it surrounded by drunk, pasty and incredible loud English football hooligans.
And that’s not all. Every time I go to read an interesting sounding article on Fortune.com, it informs me that I need to be a subscriber to access it. Well guess what? I don’t want to subscribe to Fortune. The only thing the world needs less than more dead trees is more useless clutter in my apartment. The subscription costs $4.95? Okay, I’ll pay the damn $4.95 for web access. Sounds reasonable, doesn’t it? Well, I wrote to Fortune customer support with a proposal to that effect, but they never answered.
An amusing post by Clay Shirky about one futile and vaguely ludicrous effort to categorize all flavors of human relationships using a handful of English-language terms. The effort even has a snappy name: XFN or XHTML Friends Network. (Note to self: blog a biting entry about the dangers of escalating acronymization. Are there any acronyms left that aren’t composed of other acronyms?)
What struck me most about this was the fact that XFN’s creators felt it necessary to find “gender-neutral” terms for all of their relationships. This had the most impact on the “Romantic” categories, not coincidentally the ones that Clay chose to pick on. The archetypical example was the choice of “date” for someone you’re dating, instead of the more natural “girlfriend” and “boyfriend”.
Someone ought to let these folks know that English is not the only language out there, and the effort to find gender-neutral words in other languages makes even less sense. In Czech, practically every term that refers to a woman uses the suffix ka. So a female doctor, teacher and police officer would translate, respectively, as doktorka, učitelka and policistka.
I encounter similar issues all the time in the course of developing AllPeers. Although we currently have only an English version, it is clear that we will want to internationalize it at some point. So I shudder when I’m forced out of expediency to write code like:
// Generate the name of the shared folder for this resource
strName = "Shared " + poResource->GetDescription()->GetName() + "s";
For non-programmers, this code says that the name of the shared folder for some type of resource should be the word “Shared”, followed by a space, followed by the name of the resource (say “Photo”), followed by an “s” (resulting in “Shared Photos”). Well, obviously this won’t work reliably even in English (consider what happens if the resource is called “Child” or “Goose”). In most languages, the problem is even worse since different types of nouns tend to form plurals in different ways (German is notorious for this) and adjectives often vary based on noun gender. So in French, a folder for shared photographs would be photos partagées, but the equivalent for books would be livres partagés, without the extra e at the end (since photo is feminine and livre is masculine).
My point is that it’s naive to imagine that we will find gender-neutral terms for human relationships in English, and even more naive to think that this will fly in any other language. Better to accept this and use more natural terms that may vary based on gender, since all of the machinery to handle this is necessary anyway in well-designed, internationalized software. The good news is that it’s a darn sight easier to generate natural language than to understand it.
Comments Off
Ah, a contentious topic. Gotta love ‘em! Yesterday’s rant on the evils of world domination by big technology and big media garnered some interesting comments:
Gandalf pointed out that:
Apple’s Steve Jobs said any DRM can be cracked and tried to persuade the labels not to use it but they insisted. Apple is in the role of middle man in the negotiations between the cartel and the consumer, iTMS is a service for their iPod customers, FairPlay was necessary to gain the content.
There is indisputably considerable tension between the tendencies to force strong DRM on customers, thereby maintaining control, and to provide DRM-less content in order to make said product more palatable to consumers. We can only speculate about the forces inside Apple that may have pushed for each approach. Certainly the decision to go with DRM was a sina qua non due to the need to appease the music industry, and the fact that FairPlay is quite low-profile supports the notion that, given the choice, Apple would have preferred to have no DRM at all.
But then what do we make of Apple’s frothy-mouthed reaction to RealNetwork’s Harmony, which reverse-engineers FairPlay so that Real’s customers can load their DRM-protected songs onto the iPod? My guess is that Apple first favored no DRM for the sensible business reasons I discussed yesterday. Having quickly realized that this wasn’t going to fly, they are now trying to get maximum mileage out of FairPlay by wielding it as a weapon of control.
mcloki mounted the soap box to promote a premise that crops up quite frequently in various guises:
There has to be some form of DRM on digital files. The system you are talking about is available right now it’s called P2P. If a musician wanted to they could put there music on a P2P network. Take your pick. Music doesn’t want to be free. It wants to be owned and loved by the owner.
This is a view that I addressed at length in the second part of my series Towards a P2P Business Model. The supposition is that, if content is unprotected, no one in their right mind would pay for it. But is this actually the case? Imagine that, against all odds, Apple had managed to convince the music industry execs to let them provide DRM-less MP3s on the iTunes Store. Does that mean that they would be selling fewer songs? I don’t see why: all of this content is already available on free P2P networks like Kazaa and EDonkey. Users are attracted to the iTunes store because it’s faster, more reliable and more convenient… not to mention that it soothes the soul to know you obtained your music legally.
Yes, a distribution mechanism that requires payment for digital content (and provides a user experience that makes people willing to cough it up) is a must if artists are to earn their daily bread. What is the significance of DRM in this regard? In my opinion, none whatsoever.
Just came across an interview in yesterday’s Business Week Online with Howard Rheingold, author of Smart Mobs: The Next Social Revolution. He echoes the sentiments in my previous entry regarding stifling of innovation by big media houses.
Too lazy to read it? Here’s the gist:
Q: Does the pushback by companies threatened by these trends, such as the record and movie companies, threaten innovation?
A: Yes.
This recent entry and its predecessor on DrunkenBlog offer some fascinating insights into Apple’s strategy with regard to the iPod, iTunes and its online music store. Although I’m not one of those who feels that Apple would have necessarily supplanted Microsoft as Ruler of the Universe if only it had licensed MacOS back in the 80’s, I have been somewhat baffled by their resistance to licensing its FairPlay DRM technology, considering that Apple apparently makes far more from iPod sales than from digital music.
The author demolishes this view, pointing out that the likely shelf life of the iPod technology is limited to a few years, at most, before a whole range of portable consumer electronics devices converge, resulting in a mobile phone that knows how to play MP3s and much much more. Apple’s true goal, according to this hypothesis, is to become the standard provider of digital media. The right way to get “lock in”, naturally, is to control the standard DRM technology. This would explain Apple’s anger at RealNetwork’s recent reverse-engineering of FairPlay.
While I find this argument very convincing, I am not as sure as Apple appears to be that DRM will turn out to be the secret sauce that lets a single player dominate the vast potential market for digital media. True, Apple has done a good job of demonstrating the advantages of controlling both ends of the distribution chain. Microsoft is moving in the same direction with its Windows Media Center Edition and planned online music store. The other big player in this space, TiVo, recently benefited from a decision by the FCC to let them go ahead with TiVoToGo, a file sharing service that lets customers share programs they have downloaded with other TiVo owners. I wouldn’t be at all surprised if they have big plans to offer pay-per-download access to digital content of various flavors in the future. Assuming this is the case, they are unlikely to use the DRM technology of their bitter competitors, so they will push their own “standard” or that of a partner.
The appeal of DRM to technology companies is thus clear, and everyone knows that big media is desperate to ensure that consumers do not have free rein to copy digital content. The spanner in the works are the consumers themselves, who universally hate DRM. Apple’s scheme, which is fairly inobtrusive, has been accepted more readily than most. But from the consumer’s perspective, no DRM is a whole heck of a lot better than a little bit of DRM.
Here’s what has big media quaking their boots: much of their profits are made not from adding any real value but from controlling the means of distribution. I don’t see DRM as an inevitable consequence of the digitization of content. Rather, it is a heavy-handed attempt by what basically amounts to a price-fixing cartel to maintain their advantage now that the marginal cost of delivering content to the masses is approaching zero.
Clinging desperately to clumsy DRM protections may well have the opposite effect. The road is now open for services that let artists (who suffer as much as consumers from the market distortion imposed by the media houses) place their content directly onto an open marketplace (think “digital eBay”) where consumers can find and purchase content without needing a centralized distributor. Granted, this would take time to catch on, but the lure of DRM-free, a la carte selection of films, TV shows and music is a powerful motivating force.
The seminal cartoon by Peter Steiner has inspired a veritable cult movement (and even a theatre production) with its humorous take on the effect of the internet on human interaction: “On the internet, nobody knows you’re a dog.” Published in the July 5, 1993 issue of The New Yorker magazine, the piece offers an admirably prescient view of anonymity in the wired age.
Besides the cute play on words, the illustration is funny because anthropomorphizing animals is funny. (I’m reminded of the old joke where two horses, having returned to their stable in the evening, engage in heated discussion about their strenous work day, at which point the farmer’s dog pipes up with some comment. The astounded horses turn to each other and exclaim, “Holy shit, a talking dog!”). But the popularity of the cartoon isn’t due only to the fact that it tickles the funny bone. The author managed, several years before the dot.com boom, to sum up the thorny issue of online anonymity in an amusing buzzphrase that perfectly captured the zeitgeist of the 1990s. Anonymity on the internet was perceived as a resounding positive, liberating us from our canine constraints to project an image onto the world of ourselves, not as we are but as we would like to be.
But it didn’t take long for the flip side of our newfound incognitos to become apparent. Most obviously, particularly in light of the turbulent world events of recent years, the internet holds the potential to hamper severely law enforcement and intelligence activities. Baddies, armed with strong encryption and speed-of-light connectivity, can free themselves from the prying eyes of those whose job it is to protect us. As with so many of life’s pertinent questions, there is no right or wrong answer as to where to draw the line between the enhanced civil liberties furnished by the net and the need to curtail these freedoms in the interest of security. It’s a judgement call bound to give rise to a broad continuum of perfectly defensible opinions, ranging from virtual anarchy to mandated cryptology chips that provide law enforcement with a “backdoor” to decipher all encrypted correspondance.
But let’s drop that hot potato. Internet anonymity raises an intriguing technical dilemma as well, one that lends itself more easily to analysis. The plain fact of the matter is that sometimes we want the parties with whom we interact repeatedly on the net to recognize us and remember certain pertinent details about us, including our relationships with others. This works today in certain microcosms, as with one-click shopping on Amazon. But we’re still forced to maintain a menagerie of address books, buddy lists and passwords that we turn to in various contexts. Cookies and password managers help, but your identity ends up tied to your web browser, and God forbid you should purge your browser cache or switch computers.
One serious effort to rectify this has come from Microsoft in the form of its .NET Passport. The idea is that you create an identity that resides on Microsoft’s server and can be beamed to whatever website you are visiting, so you only have to log in once and you never have to fill out Yet Another Registration Form. The service has encountered only mild success, to say the least, and this isn’t only because no one trusts Microsoft as far as they can spit. The whole premise is totally misguided: people want to own and control their online identity, not relinquish it to a multinational conglomerate.
