Mobile version (Display Regular Site)

Skip to: Navigation | Content | Sidebar | Footer


Weblog Entry

WHAT's Next

June 09, 2004

WHAT’s Going On? — Simon Willison analyzes the new WHAT Working Group, composed of representatives of three of the major four browser vendors.

On a broader level, what’s brewing in browser space is a war over what’s next. The web as a document viewing platform appears to be as good as it will get for a long time. The future is in web applications, and everyone has a different solution. Microsoft is betting heavily on its proprietary XAML and the forthcoming Longhorn technologies it will integrate with. W3C concerns are more interested in integrated solutions of XML, SVG, XForms, etc.

The reality is that right now, the web still revolves around HTML. The W3C has been trying to put it to bed for ages as it moves on to new, non-backwards compatible XML-based languages. But millions of sites and applications depend on it today. They’re not all going to find the money to upgrade to the Next Big Thing, when it seems the only thing coming next is fragmentation.

So the minds behind Opera, Mozilla, and Safari have formed a new working group called the Web Hypertext Application Technology Working Group, or WHAT for short. The focus is to develop solutions that are extensions of vanilla HTML, in order to co-operate with today’s browsers instead of relying on non-existent future implementations.

What of XHTML? What of CSS? It’s increasingly obvious that while it’s possible to use them for applications, they’re not the best tools for the job. It’s unclear whether they’ll play a key part in any of the new strategies, or will exist only to complement in document-based settings. Most indications at present lean toward the latter.

The next chapter of the ongoing saga of the web is shaping up to be a brand new epic. What makes this battle different is that the major player is a well-established company with deep pockets, and the standards body that should be releasing definitive and practical technologies appears to be wrapped up in theoretical future technologies and ignoring the current problems web authors face.

If a group of alternative browser vendors can make a difference, when their combined total market share is still in the single digits, then the forming of WHAT might be a very important milestone one day. For now it’s still anyone’s guess where this battle will take us, although given the players, it’s not hard to draw unpleasant conclusions.


Reader Comments

1
James Z says:
June 09, 02h

I don’t quite see what your getting at about “unpleasant conclusions”. I believe that “given the players”, Safari, Mozilla & Opera, all being heavily involved in very compliant standards accepting applications we can at least expect something that makes sense.

Besides that I believe XHTML and CSS to already be too widespread for the constituent players to throw it to the wind.

2
Micah says:
June 09, 02h

It sucks to be a forward-thinking web developer these days. It rules to be 18 years old, because I know by the time I’m 50 that all these problems will be solved!

3
Dave S. says:
June 09, 02h

“I don’t quite see what your getting at about ‘unpleasant conclusions’.”

Sorry, maybe that wasn’t clear – I was basically placing Microsoft and its market share in one corner, and the M/O/S consortium and theirs in the other.

The unknown third is whatever the W3C is doing, but it would seem a lot of people are trivializing that work as being ‘unworkable’ or too mired in thinking about the far off future to be practical to today’s developers. Daniel Glazman has some interesting thoughts on this that analyze the reality of today’s web – http://webperso.easyconnect.fr/danielglazman/weblog/dotclear/?2004/06/08/362-future-of-html-and-the-web

4
June 09, 12h

I agree that the future is in web applications, but I believe XHTML and CSS *will* remain the key ingredients. For the upcoming release of Visual Studio.NET 2005 Microsoft folks promise that server controls will generate fully compliant XHTML code (and I’m trying to coax them to say it out loud). If they fall short of their promise you’ll see a lot of bitching about it at my site. I won’t settle for less. As to WHAT… I think it’s great to have this initiative. It’s a good move.

5
Ian Firth says:
June 10, 01h

How interesting is it that the site is broken in IE6 ?

6
jgraham says:
June 10, 02h

> Though I would much prefer that they use XHTML and put their stuff in a namespace, rather than breaking HTML 4 validity…

I think you’re missing the point a bit.

The new tech won’t break HTML 4.(01) validity because it won’t /be/ HTML 4.01. It will be a new spec which will have a new name (HTML 4.2 or extended-HTML or something). Any document authored against the new doctype will, of course, not be valid HTML 4, but would be expected to validate against the new doctype. Of course, the specifcatioon will be similar enough to HTML 4 that HTML 4 UAs will be able to provide an acceptable rendering of the content.

Using XHTML would entirely defeat the point. XML as a document format on the web hasn’t taken off. The most widely used browser doesn’t support it at all. The browsers that do support it generally do a better job of HTML 4. Most importantly, authors don’t want it. The XML ‘error handling’ rules mean that most existing CMS’s would have to be junked or undergo a major upgrade. Authors would have to get really anal about well-formedness when most authors can’t write a well-formed static HTML page. All this for very very little benefit - most people don’t have a pressing need for MathML - which is the only widely implemented XML-requiring technology in web browsers.

If WHAT went with XML, it would be a faliure. As it is, it has some chance of success.

> And once they’re done, I encourage them to submit it to W3C.

Out of interest, why? I’m interested to know why the W3C believes that it should be the default home for all web specs - what benefit does submitting to W3C have over just producing actual implementations?

> we deserve to have a decent Web application platform, and we all know by now HTML ain’t it.

Why not? I mean, HTML has a rather good track record - being, as it is, responsble for basically 100% of all web applications at present. Sure it’s not perfect, but why do you consider it so bad that it should be thrown out?

> Brendan Eich blocked XForms development on Mozilla back in October.

” if someone wants to step up, implement XForms well, and own the code, Mozilla will take
it as an extension. A plurality of drivers would even include it by
default if it were sufficiently small [TBD, certainly 30K on top of
common XML and HTML forms infrastructure is small enough for me]”

http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&c2coff=1&safe=off&selm=4096B118.3050903%40meer.net

> I think it’s a good idea to start demanding what Web developers are going to need 3 years from now.

3 years from now? IE 6 will stiill be the dominant browser, though probably IE 7 will have reasonable traction. IE still won’t support XHTML ( http://ln.hixie.ch/?start=1086158925&count=1 ) and so HTML 4 and technologies compatible with it will still be the defacto standard for web applications. Businesses won’t have spent the time and money needed to deal with XML error-handling and will have no plans to do so since their apps already work. Xforms and SVG will be implemented in Mozilla and will be as widely used on the web as XUL and MathML are today. Opera will implement SVG but not Xforms. Plugins will still be too complex for most users to deal with. The browser will not be in any danger of being replaced by some sort of plugin runtime. Some websites will start using XAML or some other Microsoft technology since that will reach 95% of desktops and will offer functionality they can’t get any other way. The W3C will be developing SVG 2.0 which will have a passing resembalance to something that was once a vector-graphics spec.

> I don’t want to be using HTML 4.01 to create Web apps 10 years from now, no matter how much gets slapped into it.

And I guess that, in 1970, everyone was hoping that in 10 years C would be irrelevant and everyone would be programming in Lisp.

Once technologies reach a critical mass, they die very very slowly. HTML is going to be a fact of life for the next 50 years at least, maybe longer. With a little luck, it might not be the dominant format then, but I’m confident that any attempts to throw it out entirely and replace it with something significantly more complex (anything XML based qualifies here) will be a faliure.

7
Matt May says:
June 10, 09h

I presented at the workshop, on behalf of W3C/WAI.

I think that, if Hixie et al. want to create an HTML-based format to fix forms and submit it to W3C, then they’re doing the right thing now. I would _love_ to have stuff like standard editable HTML textareas, for example. And once they’re done, I encourage them to submit it to W3C. (Though I would much prefer that they use XHTML and put their stuff in a namespace, rather than breaking HTML 4 validity…) If they’re planning on being done with it this year, there was really no point in bringing it to W3C, since so much of it is still in the requirements-gathering stage.

If it goes this way, I think it’s a good idea to start demanding what Web developers are going to need 3 years from now. These are useful hacks to hold us over until then, but we deserve to have a decent Web application platform, and we all know by now HTML ain’t it.

I do take issue with Hixie’s claim that XForms is going to take years to implement and nobody’s going to use it. He and Brendan Eich blocked XForms development on Mozilla back in October.

http://bugzilla.mozilla.org/show_bug.cgi?id=97806#c71

What we have here is a set of vendors saying they want to write stuff that’s easy to absorb into their existing engines. Which is all well and good, but ultimately, short-sighted. I don’t want to be using HTML 4.01 to create Web apps 10 years from now, no matter how much gets slapped into it. There comes a time where we have to move past HTML and into something tailored to creating actual applications. What we have now isn’t cutting it.

8
June 11, 03h

Dave S Wrote: “I was basically placing Microsoft and its market share in one corner, and the M/O/S consortium and theirs in the other.”

In the red corner we have three of the best: Mozilla - emerged like a phoenix from the ashes of Netscape; Opera - what a performance! And Safari - well… I’m yet to explore that one. Then, in the blue /e/ corner, we’ve got Internet Explorer - the undisputed champion of the infamous 20th century browser wars; and ironically the worst, yet most popular user agent there is. Not only is IE single-handedly holding back the development of the web into a rich, accessible, multi-media and application platform; but it is also dragging everyone, and everything else down with it as it buckles under the pressure of new interoperable standards.

From what I can gather, it seems the overall theme for the reasons behind this backwards step to improve the virtually obsolete HTML 4.01, is mostly because of the complete lack of support from Microsoft for W3C standards such as XHTML, CSS, and the more recent XForms recomendation, despite the other three major browser vendors having successfully implemented them relatively well (except XForms).

Is it really time to throw in the proverbial towel, and concede that Microsoft will never conform to standards? And that IE will remain popular for many more years to come? Should we really give up on the *future* that W3 attempts to bring us? Should we focus solely on developing technologies that will be compatible with non-standard user agents that we have now? I agree that all standards should be interoperable, but When did the term “interoperable implementation” get expanded to “internet-explorer-operable implementation”? This is exactly what WHAT seems to be doing, driven partly by Hixie’s comments on backwards compatibility.
http://ln.hixie.ch/?start=1085764602&count=1

Although I agree with the WHATWG’s principle of creating seperate language enhancements specifically for creating web applications, I disagree with making these enhancements backwards compatible with HTML4, and current UAs. I think these enhancements should be specifically for compatibility with XHTML, perhaps with the idea that Web Applications Markup Language (WAML), and other related proposals will eventually evolve into a standalone markup language (with it’s own media type - application/what+xml, or something) that can be used in conjunction with XHTML on the web; using XHTML for documents, and WHAT’s Web Apps for applications.

9
June 11, 06h

James wrote:

”> Though I would much prefer that they use XHTML and put their stuff in a namespace, rather than breaking HTML 4 validity
I think you’re missing the point a bit.
The new tech won’t break HTML 4.(01) validity because it won’t /be/ HTML 4.01. It will be a new spec which will have a new name (HTML 4.2 or extended-HTML or something).”

There’s no reason why there can’t be *both* an HTML 4.2, incorporating Webforms *and* a WebForms module for XHTML 1.1.

Indeed, the “forward-looking” web-designers will demand it. Being reluctant to part with their spiffy new forward-slashes, they will be equally adamant about getting to use the new WebForms Specification.

Absent a WebForms module for XHTML, there will be much gnashing of teeth in the land, as Zeldman marshals the troups to abandon XHTML 1.x for HTML 4.2.

10
June 11, 06h

Indeed, both WebApplications and WebForms (don’t know about WebControls) are explicitly targetted to be both a revision to HTML 4.x *and* a module for XHTML 1.1.

As to WebControls, since DOM support in XHTML is such a rat’s nest currently, I don’t think revising the HTML_Document DOM can result in any further brokenness.

11
jgraham says:
June 11, 07h

> There’s no reason why there can’t be *both* an HTML 4.2, incorporating Webforms *and* a WebForms module for XHTML 1.1.

Sure. But my point was really that it can’t *just* sit around in its own XML namespace because that would defeat the purpose of the project.