Dec 5

Nat Torkington

Nat Torkington

72 Hours in Silicon Welly: InnaWorks and the Future of the Web

I'm in Wellington, New Zealand's new startup capital, meeting with interesting people left right and centre. Monday night I had a great conversation with Stephen Cheng, founder of InnaWorks. InnaWorks spend 1-2 years developing a product. Their first product, a J2ME optimizer that lets you fit more functionality into a mobile device's memory footprint, is doing really well. Their second will be released early next year, is also for the mobile platform, and is terrific: Stephen's been jetsetting in talks with games companies, Internet search companies, etc. around it. Sadly I can't tell you what the next product is. I can relate a fascinating conversation that Stephen and I had about the looming battle between Microsoft, Adobe, and Sun, one that's connected to the discussion of Microsoft's web design tools going around today.

Read on for more ...

Stephen wanted my take on the mobile platform space, because he sees very interesting things happening. Microsoft's Vista struggles and the rise of the web app is creating a power vacuum in desktop apps. More precisely, desktop apps are being replaced with web apps that have offline behaviour. When you begin to think of apps in that light, you see Adobe's Apollo (a cross-platform desktop RIA environment available as a browser plugin), Java, Microsoft's Avalon, and Ajax apps as all possible successors to the Win32 throne.

Of course, this is long-term. Microsoft developers and users won't jump ship overnight. But it seems a very credible view of the situation to me, and one with all sorts of fascinating angles.

For example, at one level it's Java vs Apollo. If Flash/Flash Lite gains a lot of traction, and Adobe has made some great deals to get FlashLite onto devices, then Java loses the only significant place it's used beyond the backend and toasters. If Java's just a backend language, then it's one step away from being the next COBOL. The question becomes: will IBM and Sun let this happen?

Another level has this as a standards vs proprietary battle: Flash is not open, despite the open sourcing of the Flash engine. It's a product owned by a company and developed for commercial gain. Were the desktop to tip in Apolllo's favour, Adobe would be in a position similar to that of Microsoft today. Microsoft's extensions to the web are comparable to Adobe's--they're similarly proprietary and a point of control and exclusivity. Fans of open standards must hope that Microsoft and Adobe annihilate each other, leaving XHTML and CSS as the last technology standing on the smoking ruins of the desktop. Together they must repopulate the web sites and ... no, wait, that was a movie on the scifi channel. Back to the battle.

There's also a contrast of deployment models: Microsoft owns the browser that 90% of the world uses, but Adobe uses plugins to bypass that iron grip. If Microsoft's technology is immediately usable, and Adobe's is one click away, does Microsoft have a huge advantage? It probably falls down to who can make their own version of a non-standard web acceptable first: expect both to make partnerships with huge traffic sites in order to get the public comfortable with seeing the Apollo plugin fire up, or "works best with Windows Vista". Remember those days? Itching to get back to them? Yeah, me neither.

Stephen's obviously worried because of the impact on the mobile world. Flash Lite is the only significant challenger at the moment: the Windows Mobile stuff is an island unto itself, and I haven't seen suggestions that Microsoft are integrating desktop application development with mobile application development in any serious way. But the cross-platform nature of Apollo could shake up everyone with a J2ME business.

I think of this as the battle of Greek legends: will Apollo, Ajax, or Atlas take on the Sun for conquest of the earth and sky? I don't know who will win, though I back open standards from an innate faith in good triumphing, but I do know that it will be interesting to watch unfold.

tags:   | comments: 3   | Sphere It

Previous  |  Next

0 TrackBacks

TrackBack URL for this entry: http://blogs.oreilly.com/cgi-bin/mt/mt-t.cgi/5089

Comments: 3

  Tim O'Reilly [12.05.06 07:24 AM]

FWIW, though Flash (and Apollo) are "one click away," Adobe likes to tout the fact that Flash is actually more widely deployed than Windows, on something like 97% of all desktops, including Mac and Linux as well as Windows. Of course, HTML/Ajax is now deployed on 100% of all desktops. The point, of course, is that content types, and content-based applications, are driving adoption more than backdoor deals.

The Flash player was in some interesting ways a harbinger of the "Web 2.0" paradigm: never distribute your software, just load it automatically when the user needs it. It is a success story that could only have happened on the internet, not in the era of packaged software distribution (even though the revenue engine is the packaged software of the authoring package.)

It's also an interesting comment about "less is more." Flash didn't aspire to be the ultimate "write once, deploy anywhere" software development platform, like Java did. But in many ways, as you note, it may beat Java to that goal. It's like Christensen's account of Sony and the transistor radio: make your mark in a new market where all the features don't matter, and then move up the ladder of capabilities from there, such that ten years in, Apollo is not just a handmaiden to dynamic content but to dynamic applications.

Anyway, an interesting story. FWIW, I've seen Apollo demos, and they are really slick. There's a lot of utility there that DHTML and Ajax can't yet match. But that doesn't mean that Apollo will win (after all, re-read the paragraph above.) But as you say, it's an interesting time, with a real clash of Titans (to continue the classical mythology theme) to come.

  David [12.05.06 08:13 AM]


Interesting post. Thanks. I want to comment on just one piece of it (for now at least).

"Were the desktop to tip in Apolllo's favour, Adobe would be in a position similar to that of Microsoft today."

I disagree for a couple of reasons:

We are pursuing a much more standards oriented, open source, cross-platform, and hardware agnostic approach than Vista/WPF. Apollo is built on the truly open surce Tamarin engine and WebKit. The programming model is based the ECMAScript Standard. The PDF and SWF file formats are published and widely implemented by multiple vendors. (BTW, PDF/A is an actual ISO standard, not just a de-facto standard). Flex/MXML is just an XML abstraction on top of ECMAScript, and uses standard CSS for stying. We leverage standard XML over HTTP, standard WebServervices, etc. You will be able to build and deploy Apollo applications with any text editor, standard XHTML and JavaScript, and/or our Free Flex SDK and or PDF creation technologies free or commercial from hundreds of vendors/sources. So while we haven't open sourced every technology in Apollo, in many other aspects we leverage "standards", and either way we have fully open and free paths to developing and deploying applications.

But that isn't my main point. Even if you think we are not on the right track in terms of our moves toward standards and open source, I still think that fundamentally your statement above will prove wrong because our business model is different from Microsoft's. MSFT sells Windows and it is a massive source of their revenue. Apollo (and Adobe Reader and Flash Player on Mac/Win/Linux) is free and we do not depend on them for revenue.

For Microsoft, they make money on the desktop first and foremost from the sale of their runtime, Windows. For Adobe, Apollo is free. Development and deployment can also be free with our free SDKs and a text editor. But hopefully we can build a business model around great tools and servers and services and applications. A company that commits to building their next generation apps on Vista/WPF is making a very large committment (lock-in to a significant degree) to purchase a lot of Windows. A company that commits to building their next generation apps with Apollo/Mozilla/WebKit has not financial lock-in at all. They will buy our tools and servers if and only if we can add enough value for them to perceive the ROI, but they are not locked into paying for an OS or runtime or deployment fees.

These are exciting times for web app development and a revolution in end user experiences :)

-David Mendels


  David [12.06.06 06:16 AM]

Hello Nat, let a comment 24 hours ago and see it did not show up. ??


Post A Comment:

 (please be patient, comments may take awhile to post)

Type the characters you see in the picture above.