Wak Oentoenk

One of the things I like about the English language is the number of words that have multiple meanings. Take the word hysterical. It can be used to describe someone who is reacting irrationally, with an extreme excess of emotion. As in, for example, all the stock traders who heard that AT&T had activated only 146,000 iPhones in the last two days of its fiscal quarter, concluded that iPhone sales must be far below expectations, and beat Apple's stock down 6% in a single day (link).

Then the next day Apple releases its own quarterly report, and it turns out the company sold 270,000 iPhones in the same period. This is leading to all sorts of convoluted explanations for why the numbers from AT&T and Apple don't match (link).

So now everybody says that it was all a misunderstanding and the iPhone is actually selling great, but there's another aspect to the story. None of the US mobile operators are very good at tracking what has sold through their retail stores. Unlike a consumer electronics chain, they do not keep tight track of their inventory because they care a lot more about selling service plans than they do about selling hardware. So although I'm sure Apple knows exactly how many iPhones it shipped to AT&T, I doubt anyone knows exactly how many actually sold through to users.

Our industry, and those of us who hang out on the web in particular, has a hunger for quick information and snap judgments. Real, high-quality data is slow to gather, and usually too expensive to give away free (link). Often it doesn't exist at all. So we make sweeping conclusions based on anecdotes, and we're far too quick to judge successes and failures.

The bottom line: The iPhone is probably selling well, but lots of phones sell well to enthusiasts in the first month and then fall flat. We won't really know how the iPhone is doing until we have a couple of quarters of sales data to look at. Anyone who tells you otherwise is trying to sell you a stock picking service.

Which brings me to the other definition of hysterical: A situation that is very funny. I think that one applies here as well.

"Every time we ship product to the market, whether it's in Japan or here in the U.S. or in Europe, it sells out in a matter of days.... If you see one, buy it. Don't assume that you can come back later and find one." --George Harrison, SVP of marketing and corporate communications, Nintendo of America, quoted by Reuters

For a marketing guy, Nirvana is when the world gives you permission to say something utterly outrageous, without anyone questioning it or even believing that it's marketing. Nintendo reached Nirvana two weeks ago.

Coincidentally, that's also when I wandered into the local Wal-Mart, on the off chance that they might have a Wii in stock. I had been looking for one for eight months, ever since my wife shocked me by telling me that she wanted a Wii for Christmas. (This from a woman who has traditionally had about as much interest in video games as I have in quilting. Her only explanation: "It looks like fun.")

Anyway, I walked into Wal-Mart, and sure enough, there was a single Wii box locked into the glass display case. Did I follow Mr. Harrison's "advice," and buy it on the spot? You bet I did, especially after the clerk told me someone else had called the store and tried to reserve my Wii. Tough luck, buddy. Wal-Mart doesn't do reservations. Try Toys-R-Us.

So I bring the Wii home, and now I find that I'm going through stages in my feelings about it, much like the Kubler-Ross model of the stages people go through when confronting grief (link). The stages of grief are denial, anger, bargaining, depression, and acceptance. The stages of Wii (so far) are delight, disgust, and anticipation. I have no idea what'll come next.

The delight stage was all about discovering the controller. There has been plenty written about the Wii's wireless, motion-sensing controllers, but until you actually play with them it's hard to understand how much fun they can be. I set up the system after everyone else was asleep, feeling that it was my responsibility to do some thorough testing that night, so there would be no risk of the family being disappointed if the system didn't work properly.

The first thing I tried was the Wii's tennis game, which was astonishingly easy, and fun. You just swing the "racket" at the right time, and you'll hit the ball. The controller has a vibrator in it, so you feel a little jump when you make contact. But the best feature is that there's a speaker in the controller too, so when you hit the ball you hear the familiar pock sound from your hand, rather than the television. I didn't notice the speaker when I set up the Wii. The first few times I hit the ball, I kept wondering how Nintento had manipulated the TV's stereo to make it seem like the sound was coming from my hand.

The tennis game's graphics are embarrassingly bad (it's like playing tennis against a salt shaker with a head on it), but the gameplay was so fluid and immediately rewarding that I didn't care.

Great new technology products give you a rush, a feeling of empowerment as you realize that you can now do things you simply couldn't do before. The first time I used a Macintosh was all about that. HyperCard was the same. And WordPress came pretty close. The Wii fits in that company because it opens up a whole new paradigm of gaming.

After testing the tennis game tennis thoroughly, my arm started to tighten up, so I decided to try some different games. That was when I entered the second phase of Wii discovery...

Complete disgust. There have been plenty of press reports about people accidentally throwing the Wii's controller through a window when they got too enthusiastic, but I may be the first person who almost did it intentionally. In contrast to the tennis game, some of the Wii's games are infuriatingly bad. The Wii's golf game is ridiculously difficult to control -- I couldn't even get my golfer to point in the right direction, let alone control a shot accurately. Fortunately, the game requires you to give up on a hole after 20 shots; otherwise, I might still be playing. Even a simple whack-a-mole simulation became an exercise in frustration as I tried in vain to position the hammer on the screen.

The common denominator of these games is bad use of the controller. The golf game has several modes, in which you choose direction for a shot, elevation, and so on. The controller's just not accurate enough to make it work. In Whack-a-Mole, the problem is that you're supposed to move the controller like the tip of a joystick -- forward, backward, left, right -- like the slider on a Ouija board. That's OK on a table, but when you're holding the controller in midair your arm gets tired really quickly. Human arms aren't designed to move like that on an ongoing basis.

The ironic thing is that those games actually work pretty well on a traditional joystick.

There's a lesson here about the strengths and weaknesses of integrated systems design. The controller on its own would not have been successful -- it's terrible for controlling traditional games. But the games on their own would also have failed -- bowling was one of the worst games on the Nintendo GameCube, but it's one of the best on the Wii. To get a breakthrough, you have to design the hardware and software together as an integrated system.

But that same integration also presents a lot of challenges to game designers. The Wii requires a thorough rethink of how a game is structured and what you can do with it. You can't just take an existing game, port it directly, and expect it to work well. At a minimum, the whole interface has to be rethought. But really what we should be doing is rethinking what sorts of things you can do in a game. What about a game in which you draw images on the screen using the controller, or conduct an orchestra? I don't know if either of those would be entertaining, but it's the sort of stuff we should be thinking about.

Which brings me to the third stage of Wii discovery...

Anticipation. I have one word for you: lightsaber. Like every boy who grew up watching the Star Wars movies, I've always had a secret desire to play with a lightsaber. Not one of those plastic things they sell at Toys-R-Us, I'm talking about a real lightsaber that makes that buzzing noise and can cut through steel like butter. I'm not sure what I'd use it for -- it seems a bit like overkill for tree pruning -- but I know I want one.

With the Wii, we finally have a device that can make it happen, at least in simulation. Supposedly there's a Lego Star Wars game on the way for the Wii, which will let you control your lightsaber directly. I am both impatient to get it and dreading it. The dread comes because this is a port of an existing game rather than a redesign. Some reports say you won't really have full control over your lightsaber (link).

The disappointment could be crushing, so I'll have to test it before I let the family try it. To protect them.

One thing's for sure -- if it works, my wife's not going to be wasting the Wii on tennis anymore.

The argument over the viability of Web 2.0 applications misses the point -- most of the applications on any new computing platform die. What matters are the innovations and new business models that we learn from them (link).

Last time in this series I discussed what we're learning from Web 2 about managing a community online (link). This time I want to talk about the role the Internet is playing in the creation of new forms of media.


Is the internet a new medium?

I should start with a definition of what a medium is. Webster calls it, "a channel or system of communication, information, or entertainment" (link). I want to build on that a little. To me, a medium is something that moves information and/or entertainment between people. Movies are a medium, newspapers are a medium, oil painting is a medium. So is the telephone call, when you think about it. Each medium has its own distinct usages, economic model, and audience.

A lot of people have written about the Internet and/or the Web as a new "medium." A quick online search will give you thousands of articles and weblog posts on the subject. But there's something funny about the articles -- although they all call the Internet a medium, they define that medium in many different ways. For example...

--The Internet is a medium for mixed-media communication.
--It's a medium for online music broadcasting.
--It's a medium for making politically-motivated attacks. (And an unregulated medium at that. Heaven forbid we should practice unregulated politics.)
--It's "a perfect medium for the sale of software and other digital products."
--It's a medium for interactive, moving content.
--It's a "new medium for business communication."
--It's "a medium of news dissemination."
--It's "a new medium for design."
--It's a new medium for video.
--It's a new medium for communication by individuals.
--It's a new medium for socializing.

I think that in reality the Internet is not a new medium for anything. It's a transport mechanism. It is to data what a road is to eighteen-wheel trucks. And the Web isn't a medium either; it's a set of protocols for accessing and delivering data. To abuse the road analogy, it's the warehouses and truck stops that load, unload, and service the trucks.


The Internet is a meta-medium

When we talk about the Internet as a medium, we're confusing the delivery mechanism with the goods being delivered. This is a crucial distinction, because if you think of the Internet as a medium you won't understand its real power. The Internet is a meta-medium. It's a medium for creating new types of media; a general-purpose mechanism that spews new media as quickly as people can think them up.

And spew it does. As I hope you know if you've been reading this weblog for a while, I am not a fan of hype and overblown predictions. But I think the evidence shows that the Internet is enabling an explosion of new forms of media at a faster rate than ever before in human history. I believe this is one of the most revolutionary effects of the Internet, but we're so close to it that we don't think about it much.


Freeing media from the distribution mechanism

In the past, each new form of media was generally tied to a unique distribution infrastructure, technology base, and economic model. For a new medium to arise, you generally had to create a whole new production and distribution mechanism for it. For example:

Novels required the development of the printing press, a distribution infrastructure consisting of publishers and bookstores, and an economic model in which the reader pays and revenue is shared with the publisher and distribution chain.

Radio serial drama required the invention and sale of millions of radios, the construction of studios and transmitters, the creation of production companies and networks, and an economic model in which advertisers paid for the programs.

Movies required not just the creation of motion picture cameras, but also studios to produce the films, modified theaters to show them, a distribution system to deliver the reels of film, and an economic model in which ticket revenue and in-theater food sales combined to pay for the whole thing.

The huge effort and investment involved in creating these distribution chains severely limited the growth of new forms of media. For example, it took about 20 years from the invention of television and movies until either of them reached broad commercial distribution.

In contrast, new media proliferate on the Internet as fast as people can visit new websites and install plug-ins. (Obviously, this applies only to media that can be distributed electronically. But that still covers a lot.)

This chart gives you an idea of how the pace of change has accelerated.


This chart was based in part on a fantastic media history here.

Some people would say that most of the Internet media types I listed on the right edge of the chart aren't actually new media; that they're just a tweak on existing media. For example, Henry Jenkins argued in a great article for MIT Technology Review that you have to differentiate between media, genres, and delivery technologies (link):

Recorded sound is a medium. Radio drama is a genre. CDs, MP3 files and eight-track cassettes are delivery technologies. Genres and delivery technologies come and go, but media persist as layers within an ever more complicated information and entertainment system.

I think he's right from the perspective of classifying things analytically, but if you follow that thinking religiously then it's almost impossible to create a new medium any more, unless smell-o-vision or machine telepathy comes along. I think in practical terms, you have a new medium as soon as you create a substantially different set of audience and business dynamics, because those are the changes that create meaningful new economic opportunities for creative people and businesses.

Here's the test: if you can't take material created for some other medium and replay it unchanged, then I think you've invented a new medium. CDs were not a new medium because they were created and sold in the same way, to the same people, as vinyl LPs. But radio drama was a new medium, because it had its own distinct audience and rules. You couldn't just take a stage play and turn it into a radio drama unmodified.

By this standard, the Internet is spawning new media forms faster than bunnies breed in Australia.

Of course, not all of these new types of media will be successful long-term. But it's exciting to see so much experimentation happening so quickly, and I believe it will have a profound effect on the ways we communicate and entertain ourselves in the years to come.


The revolution in front of you

Okay, so that's the theoretical foundation on what's happening. Let's discuss some examples -- three new forms of media we're creating, the rules and opportunities they create, and what comes next.


Online video

Oh, man. This one's so complex that you could write a book on it. The term "video" includes a huge variety of different things -- music videos, TV shows, animation, movies, video clips from amateurs, even commercials. Each one appears to have a different online audience and different financials.

Some of them have already run through a cycle of excitement and disappointment. For example, some people speak of an "internet animation era" that came and went at the start of the decade (you can read more about the expectations here). Usually the culprit for the disappointment is the failure to find a sustainable business model.

The hottest area in online video today is obviously short clips like the ones you see on YouTube. The ironic thing is that this form of video had virtually no traction prior to the Internet. Meanwhile, movies and TV shows -- which everyone predicted would move onto the Internet quickly -- don't have nearly as much momentum online.

Why YouTube is successful. Using YouTube is like eating potato chips ("crisps" if you live in the UK). When you're bored, it's great to browse short video clips looking for things that are funny or amazing or just plain weird. The brilliant aspects of YouTube (in my opinion) are that the video loads fast (can you imagine eating potato chips if you had to unwrap every chip individually?), and that the YouTube site links you to lots of other related videos, so it's easy to wander. If one video is boring, you're only moments away from something else.

This instant gratification factor turns the rules of traditional video on its head. In traditional video, quality and an immersive experience are king. To suck people into a television program or a movie, you use incredibly high quality images, editing, and sound. (If you want to know how important this is, look at all the enormous amounts of money the industry is spending to move to high-definition broadcasting and higher-capacity DVDs.)

That's why short online video is a different medium. Rather than immersion, the goal is instant gratification.

But how do you make money? The problem with short online video is that no one's sure how to make money from it. You pay to see a movie. You watch ads on television (well, you're supposed to, unless you use TiVo). Many companies are trying to attach commercials to online videos, but the result is often extraordinarily annoying to viewers.

That's not intuitive to the broadcast folks. Depending on what country you're in, to watch free TV you'll typically watch nine to 20 minutes of commercials in order to see an hour of programming (link). That's a ratio of between 15% and 30% commercials.

Apply that same ratio to a short online video, and you're watching a 30 second commercial to see a two minute video clip. Sounds reasonable, right? It's actually borderline intolerable to viewers because it breaks the instant gratification cycle. The whole idea is to beat boredom, not generate it.

Remember, this is a new medium. It has its own rules.

Maybe the answer will be very short ads, but no one knows what's short enough, and if those short ads will even work. Or maybe the answer is putting print ads on the website alongside the video. But unlike search, you don't know what topics a video viewer will be interested in, so it's much harder to target the ads. How will you individually track the demographics of people viewing more than six million separate YouTube clips? You'd basically have to build a database on the individual thoughts and behavior of every Internet user. That, I presume, is why YouTube was a good strategic investment for Google. It's also why I'm deeply skeptical about the high-profile efforts by entertainment companies to create sites competing with YouTube. Without Google's demographic and ad-targeting infrastructure, it will be hard for a competitor to monetize its videos.

And oh by the way, it's not clear that even Google can make this whole video thing work financially.

So let's classify short online video as an emerging medium: Proven audience, unproven economics.

Video in the mobile world. This is the current Flavor of the Month in the mobile data world. (Or maybe it was last month's flavor, and this month is GPS.) Anyway, there are a lot of people predicting that video is going to be very hot in the mobile space.

As was the case with PCs, you have to ask what sort of video you're talking about. The most intuitive use is short video. We know people use mobiles as boredom-busters, and short video is almost custom-made for that. But we run into the same economic problems as we have on PCs, only more so. It's not clear how many commercials people will tolerate in their mobile video.

Broadcast video, viewed on mobiles, is becoming popular in Asia. But by my standard that's not a new medium -- it's just building a television into your phone. And it bypasses the Internet, so it's not relevant to this discussion. (I recently wrote a long article on mobile video; if you missed it you can read it here.)


Virtual Reality as a Medium: Second Life

Most people think of Second Life as a game, or maybe a cult. But my Rubicon colleague Bruce La Fetra recently wrote an article (link) making the case that it's a new medium, and I believe he's right. Think about it. Here's the test of a new medium:

--Facilitates interaction between people. Second Life certainly does that.
--Has its own distinct audience. Double check. That's why some people look at Second Life as a cult.
--Has its own economic model. Triple check. This one even has its own currency.

A virtual meeting place. Second Life is so flexible that it's very hard to say what it'll turn into ultimately. It's already a meeting space for some people, and the upcoming addition of voice should improve that dramatically. Supposedly Cisco is providing pre-built avatars for employees, and a number of tech companies are using it for meetings (check out the slightly breathless but eye-opening article here).

Second Life is a tool for holding three-dimensional visual conversations...I know some people can't hold a serious business conversation without a pen and paper to draw with; Second Life is made for those people....One day, you'll be able to import sales data from an Oracle database, create a three-dimensional diagram of that data that changes in near-realtime, and hold a meeting of top corporate executives all over the world in Second Life to discuss the results. --Mitch Wagner

Prototyping the physical world. Another clear use for VR is allowing individuals and corporations to create interactive experiences for others. For example, as Bruce points out, hotels are starting to test lobby layouts using Second Life. Brands like GeekSquad are using Second Life to reach out to customers, giving them another way to engage (read more about it here).

Some of this commentary is so enthusiastic that it reminds me of the commentary we saw in the bubble period. Second Life is definitely a geek playground, but I'm not sure how many "normal" people will want to mess around in virtual reality. We won't know until we try.

Is it a business or a standard? The ultimate business model for Second Life is still up in the air. Land owners pay real dollars for virtual real estate and corporate avatars, giving Linden Lab a revenue stream. However, the company is in the process of open-sourcing its server code. This will make it possible for anyone to create their own "land" without paying Linden Lab, and dramatically increases the likelihood that Second Life's technology will become a generalized standard for virtual reality. That's very healthy for the medium, but leaves Linden Lab without an obvious business model. There's an interesting discussion here.

The process of moving from a captive platform to the base of an open ecosystem is incredibly tricky. I think Linden is right to do it, because otherwise an open standard for virtual reality would have eventually emerged, pushing Second Life completely out of the picture (think of what happened when AOL went up against the Internet). But now Linden will need to find some parts of that open ecosystem where it can provide valued services. I think managing the virtual currency is a good start, but I haven't been able to find any clear statement of what the company's long-term financial model will be; please post a comment if you find one.

So the status of Second Life is similar to that of online video: Definite audience, unclear financials.

Virtual reality and mobile. Virtual reality thrives on large screens and fast processors. I think it's probably safe to say that it'll be limited to PC-sized devices for a long time (at least until we get flexible screens and fuel cells powerful enough to drive high-end graphics processors in a mobile). Until that day, I wouldn't be investing heavily in creating a SecondLife client for Nokia S60.


Feeds

Actually, these are several new media that I have grouped together for convenience: Text feeds, audio feeds, and video feeds. Plus more types of feeds to come.

Different feed types have different audiences. Steve Olechowski of Feedburner gives a great speech summarizing the feed world and what's happening in it. One of the interesting tidbits he gives out is that different types of feeds tend to be dominated by different subjects. Text feeds most commonly focus on technology, while audio feeds are most often about music, social issues, and religion ("Godcasts"). So different forms of communication -- text vs. recorded speech -- attract different types of creators and audiences. I suspect that video feeds are going to be different yet again, although it's probably too early to judge today. You can hear one of Steve's speeches here.

The thing I like about feeds is that they're efficient. Rather than going to a website to read or listen, you can bring the content to you and access it on your terms. A lot of people use online feed readers like Feed Burner, but my favorite is Feed Blitz, which consolidates all your feeds into a single daily e-mail. That lets me scan about a hundred articles a day in a matter of minutes.

Text feed vs. weblogs. One problem with text feeds is that they take readers away from your weblog, meaning they won't see the ads. That creates a lot of concern for weblog authors who rely on advertising. So they do things like putting only article summaries in their feeds, or embedding ads in the feeds, neither of which are popular with feed users.

Olechowski argues that authors shouldn't worry -- that the people who read feeds are different from the people who read websites, so there's little cannibalism. He says that providing a full-text feed from your weblog actually increases visitors to the site, rather than reducing them.

He has an incentive to say that, since his business is distributing feeds. But I think he may also have a point. Let's use Mobile Opportunity as an example: About 80% of the readers coming directly here are referrals from other websites and web searches, not returning readers. I think the general pattern for readers is that they come here from a web search or other link, and if they like the content then they subscribe to the feed. That's why I put extensive introductory information and links to previous articles in the sidebar on the right side of the page. If a web search visitor is interested in the sort of things I write about, I want to make sure they can determine that quickly so they'll either bookmark the page or subscribe to the feed.

The feed readers never see the sidebar, but they don't need it because they know what I've written about before. People who read via feeds have a different set of special needs. Chances are they use a feed reader that consolidates a lot of different feeds, which they then skim quickly. That makes it very important to use self-explanatory headlines for articles, and clear sub-heads within each article so people can skim easily. Web links are a special problem -- because they're colored and underlined, they stand out from the text. But they're not usually the things you want people to skim, because they don't summarize the content. That's why I've started putting links at the ends of sentences, rather than embedding them in the flow of the sentence.

I'm not trying to make money from this site, but if I were, I'd have to think very hard about what sort of ads go on the web page vs. in the feed, and where they get placed.

The bottom line: you write a little differently for a feed than you do for a weblog, and the financial model is subtly different as well. So it's a slightly different medium.

Status of feeds: Text feeds are quite well established, and audio feeds took off rapidly once they were enabled on the iPod. The financial model (to the extent that there is one) appears to be advertising, but I haven't seen a good discussion of the economics of advertising within feeds (please post a comment if you know of one). Presumably Google's recent purchase of FeedBurner is intended to allow them to stream ads into feeds, so we'll probably see more activity there. The dynamics of other types of feeds (video, etc) are still to be determined.

Feeds and the mobile world. Feeds are a spectacular fit for the mobile world; actually a much better fit than browsing. In general browsing is something you do live, while feeds can be fetched in the background, cached on the device, and then read or listened to whenever the user wishes.

A text feed is also much easier to reformat for a small screen. In a lot of ways, it's designed to be reformatted.

If I were working on a mobile data device today, I'd push this feature very hard -- figure out who my target customers are and what feeds they'd be most likely to enjoy, cache the top ten our so automatically, and give a great discovery mechanism so people can easily find more. Feeds are a commodity in that you can get them for free, but easy navigation and discovery of feeds is potentially a very attractive area for innovation.

I know third party developers are already doing this; if I were at a mobile hardware company I'd be making it a standard feature in every device.


What comes next?

What other media are emerging? Many more new forms of media emerging than I've listed here. I'm very interested in your ideas -- what do you think are some others to watch, and what's special about them? One I'd love to investigate more is the rise of casual games -- quickie games, usually based on Flash. Games like this were very popular in the early days of personal computing, and they seem to be making a comeback on the web. You can find some nifty ones on sites like Kongregate (link; check out Fancy Pants).

The transcendent need for a billing mechanism. When I said that the Web is a tool for creating new media, I left out an important detail. It's three-quarters of the tool. We have a great delivery system, and Google is well on its way to dominating the advertising part of the financial model. What's missing is a standard mechanism for people to pay for content that's not supported by advertising. Some types of content work fine with ads, but I think some other types are better when paid for. Novels, short stories, music, and research reports all qualify. Creators and readers would both benefit from a system in which people could easily pay a few dimes or a few dollars directly to the author, but today we generally have to fumble with credit cards and awkward systems like PayPal. And credit card vendors strongly discourage small payments.

Minipayments vs. micropayments. The Web community chewed over this issue and spat it out several years ago. They believe that micropayments are dead, and the subject is closed. You can find examples here and here and here and here. Wikipedia has a nicely balanced discussion of the debate here.

This is one of those cases where the groupthink tendency of the tech industry is a liability. It reminds me of MP3 players before the iPod -- a lot of people have tried something, nobody's gotten it right yet, and therefore it must be impossible. It'll continue to be impossible up until someone does it right, at which time everyone will suddenly agree that it was inevitable.

(Quick aside: Whenever everyone in the tech industry agrees on something, bet against them. A perfect consensus is a sign that healthy questioning has ceased, and there's bound to be a blind spot.)

In this case, I think the blind spot was that people predicted the wrong role and features for micropayments. Some people made it a payment vs. advertising debate (link). It's not -- some types of media are good for advertising, some good for payment. We need both, with a creative tension between them.

Another problem is that some of the advocates of micropayments envisioned a very fine-grained payment system, in which people would pay hundredths of cents for all sorts of content, like the way natural gas or water is metered. That sounded logical, but it didn't work in practice because gas and water are predictable commodities; you don't mind metering because you know exactly what you'll get. You don't know how good a website will be until you've visited it, by which point you have already paid if you're metering. We need larger payments for content that people can preview and read reviews about before they pay. Apple has proved decisively that on the wired Internet a payment system that charges about a buck for discrete chunks of content can indeed succeed.

Call it minipayments.

We desperately need a generalized minipayment system for content on the web. Because people have to trust it, it needs to come from a major vendor, and it should be exposed to developers as a web service so it can grow rapidly. Ideally, it should be tied to a lot of existing content with an easy discovery mechanism (again, like iTunes). Yahoo would be the perfect company to provide this service. Microsoft could do it too. Unfortunately, a lot of companies are focusing a huge amount of their energy on the almost hopeless task of beating Google in search advertising, when the better opportunity is owning a different piece of the infrastructure, one that doesn't have a dominant vendor yet.

Other companies that could do it include Amazon, Apple, eBay, and even Linden Lab. Google could do it too, of course, but it appears to be more interested in stealing PayPal's customers than in building something new.

I'd put this service on the list of computing products I want desperately, right after the info pad. Somebody's going to do it eventually. When they do they'll get a great business franchise, and the explosion of new media on the Web will accelerate even further.

I can't wait.

Next time: The Web as a software development platform.

There are so many interesting things going on in the industry that it's frustrating, because I don't have time to write about them all.

Jerry Yang is now in charge at Yahoo, which in my opinion means a lot because a founder is often much more willing to revisit old assumptions and make radical changes than is someone who came in after the fact. (I know the stereotype is that founders resist change, but I've found that the exact opposite is often true, especially if the founder is moving up after spending time lower in the management chain.)

Google bought Grand Central, which underlines their interest in providing client software for mobile phones. It's a significant change for Google because up to now they have focused mostly on providing mobile versions of their existing web apps, like Maps. Grand Central is different; it's a call management system that embeds Google deeply in the life of a mobile user. It implies a much tighter relationship between Google and the user than most other Google products, and it's not something that you can easily monetize through advertising -- which makes me wonder whether Google is planning to run it standalone or integrate it into something bigger.

But the strangest recent development was Apple's decision to port its Safari web browser to Windows.

It is not easy to port a browser to a new platform. There's a huge amount of programming involved -- to do the actual port, to debug it, and to maintain and upgrade the code as people identify small incompatibilities and ask for new features. I lived through PalmSource's effort to get a good browser for Palm OS, and talked with the Be veterans about their browser work. The quick summary: it's a huge pain in the butt.

What's Apple hoping to get? The engineers at Apple who are spending their time on Safari for Windows could be creating new features for the iPhone, or helping to finish the next version of Mac OS X. Although Apple is rich enough to hire a lot of engineers, the supply of really good ones is limited, so Apple's definitely paying a price to do the port. And for what? To get people to use an alternate browser, you have to give it away for free. So there's no immediate benefit to doing the port.

A lot of Apple enthusiast sites have asked what's going on, but I'm not persuaded by most of the answers they came up with. For example, a site called Apple Matters gave four possible motivations: for bragging rights, to show Windows users what it's like to use a Mac, to give iPhone website developers a tool to test their sites, and to get revenue from search referrals to Yahoo and Google (link).

Apple Matters seems like a very good site, and to give them credit, even they were skeptical about some of the possible explanations. None of them work for me. Apple doesn't need more bragging rights, a browser is a very awkward way to show off the Mac UI, iPhone developers can buy an iPhone to test their sites, and the search referral fees from Yahoo and Google can't be all that big or everyone would be writing browsers.

I think the motivation runs deeper. It turns out that Apple didn't just port the browser to Windows; it ported the browser, the underlying Web rendering engine, and the Mac OS X programming frameworks that the browser relies on. In other words, Apple ported an entire OS layer onto Windows, and the browser is riding on top of that (link).

Now that's interesting. Apple is backing into the cross-platform OS layer business. Maybe the OS layer is just a convenient way to do the browser port. Or maybe the browser is just a trojan horse to get the OS layer on a lot more systems.

Add to this situation Apple's other recent strange announcement -- that it's "enabling" iPhone applications development by supporting Ajax web software on the iPhone. The problem with Ajax/Web2 applications is that they rely on a constant network connection in order to work. They're just thin clients to a server on the Web. Considering the iPhone's lack of true 3G connection speed, and AT&T/Cingular's well-documented data coverage limitations, Ajax-style development is about the worst thing you could do on the iPhone. What the developers wanted was the ability to create native Mac OS X applications, and Apple blew them off.

Why piss off the developers, and why put such a huge handicap on people supporting your critical new product?

Maybe the iPhone is so screwed up internally that it can't support third party apps. Sure, and maybe Apple wants to port Safari to Windows just for ego.

If you want a single idea that explains both actions, it's this: Apple realizes that in the long term, the development platform that matters is not the OS on the hardware, but the software layer that the web apps run on (I believe that; you can read more here). Apple realizes that this layer will eventually become good enough to displace native personal computer apps. Web apps then become both an opportunity and a challenge for Apple. The opportunity is that they're a way to take down Microsoft. The challenge is that the same process that obsoletes Windows obsoletes other PC operating systems, including Mac OS.

This makes it vital for Apple to create its own Web apps layer, so it can control its own destiny and increase its power. That goal would be so important that Apple would be willing to handicap iPhone apps development in the short term in order to make developers focus on the web apps platform in the long term.

If that's Apple's thinking, then the next thing to watch for will be Apple gradually adding more features to its OS layer, in the guise of browser APIs and feature enhancements. Those features will be deployed at the same time on the Mac, the iPhone, and Windows Safari. And Apple will start evangelizing web app developers to use them.

The war to come. This could set up a brutal competition in software layers, between Adobe Apollo, Microsoft Silverlight, Sun's revised Java, Firefox's platform, and Apple. Google fits in there somewhere as well, but it's not clear if they'll try to create their own platform or work with several other players.

I think this is where the most interesting action's going to be in applications development in the next few years. Stay tuned.