A few weeks ago, I attended the W3C Workshop on Web Applications and Compound documents in San Jose. Sarah Allen presented on behalf of Laszlo, with a position paper on Web UI titled The Future of the Web is not the Past of Windows.
This was my first exposure to a standards body meeting, and some of the perspectives were surprising. Attendees included representatives of Adobe (who hosted the workshop), Microsoft, IBM, Sun, the Mozilla Foundation, Opera, OpenWave, Nokia, France Telecom and more.
The meeting seemed to have been motivated by the growing interest in supporting interactive applications on the Web, and the challenges to the W3C standards posed by proprietary solutions in this area such as Microsoft's XAML.
The viewpoints expressed on the first day seemed to fall into four categories:
What I found especially surprising was the lack of acknowledgement of the reality of the browser market. Putting aside mobile devices for a moment, how is it that W3C members (including WHAT-WG) see any of these changes hitting the market?
Given Internet Explorer's lock on the browser market (and on the distribution channel for the browser), how is it that anything will change in the browser (such as implementing XForms, SVG, or some hypothetical extension of DHTML) without Microsoft deciding to make it so? Knowing that development of Internet Explorer has essentially stopped should be a major red flag to this group, yet I did not see this fact acknowledged -- to the point that a W3C representative challenged one of the Microsoft attendees to tell him "why should we listen to you?" Baffling.I see HTML/DHTML staying around indefinitely -- under no circumstance will Microsoft or anyone else truly replace it -- but being relegated over time to a document format, not a way of writing networked applications. It is difficult to imagine that the thorny issues of integrating multiple complex XML languages (SVG, XForms, XHTML) will happen in a timely manner or result in a framework which is easily learned -- particularly when there aren't widely deployed browsers that implement this hypothetical integrated standard. It is also difficult to imagine that application authors on the Web will avoid the temptation of creating XAML/Longhorn apps just because XAML is proprietary -- do Web authors avoid IE-specific DHTML today?
What will stem the XAML tide is an alternative application markup technology that is deployable across the Web, and not specific to Windows. Coming from the W3C viewpoint, this would be a difficult problem to solve considering the realities of what it would take to upgrade the browser and get it installed on everyone's desktop. I'd go so far as to say that even if the perfect XForms-SVG-XHTML browser were available right now, worked perfectly, and required only a tiny download, it would still gain only negligible browser market share. Any actual development and deployment scenario will only be more challenging.
At Laszlo, we believe we have this technology today -- an XML Web application language that can run in any browser, not just on Windows or Windows Longhorn. It is not currently itself a standard, though it incorporates a number of W3C standards. It is not Slashdot-compliant, since it relies on a ubiquitous, well-documented (though proprietary) plug-in. But it is well-specified, comparatively easy to learn, and utterly practical for deployment right now. Is it more important that a technology be approved by a standards body, or that it works the same everywhere, reliably, today?
Posted by temkin at June 19, 2004 11:26 AMIt's not just the plugin that is proprietary, but Lazlo's server backend technology. If all portions were free and unencumbered, perhaps the W3C committee would take it more seriously.
Whatever ultimately wins the cake cannot be a degraded experience when running on IE. That seems to leave two choices; either a new Microsoft technology or an open, standards-based plugin that has some way of achieving ubiquity. It would be convenient if Macromedia realized this and opened up the flash plugin.
Posted by: Anonymous at June 20, 2004 11:03 PMAnonymous,
There is a difference between an open standard (in which there is a spec with multiple implementations, some or all of which may be proprietary), and open source, which is often characterized as "unencumbered," and seems to be what you're looking for.
On a related note, the Flash file format is quite well-documented (very "open" in this respect), and it's entirely possible for someone else to implement their own Flash player. This has been done with previous versions of Flash. But this new player may or may not be open source, and it wouldn't be a W3C standard.
You say that there are two choices: MS Longhorn or a hypothetical "open, standard-based plug-in that has some way of achieving ubiquity." But that does not exist, even in the planning stage.
If you're looking for a cross-platform, realistic alternative to the XAML, you might consider opening your eyes to possibilities that are commercial, but portable. This is what I'm getting at in this post: not asking to be "taken more seriously" by the W3C (we're pretty happy with them for inviting us to present), but to show that there is a third way that is entirely practical and of genuine value today.
- David.
Posted by: David Temkin at June 21, 2004 04:51 PMThere are two problems with Lazlo as far as web standards go.
First, the use of javascript rather than static markup to embed a Lazlo object breaks scrapers, spiders, RSS readers, and caching proxies. I understand why you guys do it this way (because iframes don't do what you need), but this is a problem that must be solved before Lazlo objects can be integrated with the rest of the web.
Second, Lazlo objects themselves are opaque to the rest of the web. It's not possible to index into them using fragment identifiers and other standard tools. This is in an interesting problem that's worth chatting about at some point.
Posted by: Lucas Gonze at June 25, 2004 02:19 PMAnother item for Lazlo and web standards -- the Soundblox should use a standard playlist format. I've been taking Soundblox into account in my playlist work (e.g the XSPF XSLT at http://gonze.com/xspf/xspf-draft-2.html includes a transform into Soundblox playlist format), but in the long run not being able to use standard formats hurts Soundblox.
Posted by: Lucas Gonze at June 26, 2004 09:02 AMLucas,
Interesting points. We met with some engineers at Google last week and did discuss the problem of spidering/indexing RIAs (not only Laszlo's, but Java applets, Flash apps, and DHTML-intensive apps as well). We are very interested in working out a standard, vendor-neutral way of solving the problem, but there's a long way to go. Part of the issue is that it's not the source code you want to index (i.e., the LZX), but the dynamic data, so that it would be possible (for instance) to bookmark a product page within an RIA catalog application, or (closer to home) a song within an RIA audio player such as Soundblox.
The XSPF stuff looks interesting. It should be possible to get this format into Soundblox via XSLT or by changing Soundblox proper (note that soundblox is open source).
- D.
Your Laszlo stuff is very interesting, and I entirely agree that plugins can cause evolution, which is why IE not being upgraded to a new technology isn't relevant, you just need to get new plugins shipped not new browsers. There are 2 plugins with big enough market share to leverage the upgrades, and there's a good chance even a new one with sufficient backing could get deployed. I don't think flash is a solution, I don't think it can be re-targetted in the market, something like laszlo that runs on top of flash, maybe, I don't think having so many levels is a good thing, in the end it would depend on the tools.
Posted by: Jim Ley at June 28, 2004 01:14 AM