Linux and Open Source

Steven J. Vaughan-Nichols & Paula Rooney

The HTML standards process grinds on

By Dana Blankenhorn | December 29, 2008, 4:46am PST

Summary: There is a Moore’s Second Law effect at work here. As standards become more complex they take more time to coalesce. And as with other Moore’s Law artifacts this also tends to move geometrically, not arithmetically.

Back in January I wrote that HTML 5 would prove one of the big stories of 2008.

You agreed and made it the 6th most popular post at this blog for 2008.

Maybe we were both wrong. As I write this in December HTML 5 seems no closer to implementation than it was in January.

This is not an overt criticism of the W3C. I suspect this is in the nature of the standards-making process.

There is a Moore’s Second Law effect at work here. As standards become more complex they take more time to coalesce. And as with other Moore’s Law artifacts this also tends to move geometrically, not arithmetically.

The same effect helps explain the motives behind the development of enterprise open source.

As software gets more complex it takes more time to write and debug. But the value the marketplace puts on the improvements does not rise correspondingly. Not forever at any rate.

If the value of software kept growing in relation to its costs we might be paying $1 million per copy for Windows, not a few hundred dollars. Open source makes this manageable by allowing the costs to be shared, as at Eclipse or Apache.

Standards seem to work similarly. Just the cover of the standards-writing document for HTML5 is daunting. Is this code or another O.J. trial? (Rimshot.)

What seems clear in going over the contents is that, assuming it’s ever implemented, HTML5 will be as different from current Web standards as IPv6 is from IPv4.

There is a warning in the preceding. IPv6 is still not fully implemented nearly a decade after being approved. Let’s hope that doesn’t happen here.

Kick off your day with ZDNet's daily e-mail newsletter. It's the freshest tech news and opinion, served hot. Get it.

More from “Linux and Open Source”

Cygnus files first challenge to Bilski patent standard

The year of the mobile app

Topics

HTML, IPv6, Standards, Quality, Networking, Open Source, Business Operations, Dana Blankenhorn

Dana Blankenhorn has been a business journalist for 30 years, a tech freelancer since 1983.

Full Bio
Disclosure
Contact

Vendor HotSpot

Here to help you with your Document Management Needs

Read the DocuMentor blog now

Learn More »

Talkback Most Recent of 15 Talkback(s)

Follow via:: RSS; Email Alert

CSS 3 is the same . . .

On a related note, the CSS 3 standards are experiencing the same issues and are taking a long time.

One thing that is happening is that "last call"s apparently aren't. The various parts of the standard fell back to "working draft" pretty quickly and will likely stay that way indefinitely.

IPV6 is pretty much implemented universally. Everybody with a completely updated XP or Vista has it. New routers have it. It's pretty much the same story as always: Businesses move at a glacial pace, so they are the ones that usually prevent numbers from being better.
CobraA1
29th Dec 2008
- Reply to
- Flag
W3C.... Fail

The W3C seems to be too weighted down with bureaucracy and special interests to be useful any more.

The web is basically run on HTML 4, XHTML 1, CSS 2... none of which have had any significant update in years and years.

The W3Cs lack of urgency will quickly make them irrelevant. If you take a step back and look at the current standards you'll quickly realize that (X)HTML are too basic and CSS is quite convoluted...(I mean come on... margin: auto; to center something? Float, relative, absolute positioning... completely unintuitive! Then they tell you not to use XHTML tables for positioning and then turn around and put the table concept into CSS!!!)

The market won't wait for these sloths. Microsoft is coming out with Silverlight, Sun is coming out with JavaFX, and adobe is coming out with AIR. This will be the future of the web and it is gearing up to make the Netscape vs Internet Explorer wars look like kids play.... we shall see if any of the W3C standards live through that war.
mikefarinha
29th Dec 2008
- Reply to
- Flag
Is the W3C really to blame?

I don't think I criticized the W3C explicitly in my
post. But it does seem to have gotten lost in the
weeds.
DanaBlankenhorn
29th Dec 2008
- Reply to
- Flag
Well... they're "In Charge"

They are the ones in charge of maintaining and updating the standards so they bear the brunt of the criticism.
mikefarinha
29th Dec 2008
- Flag
Learn your craft

Your lack of understanding of CSS doesn't make a strong
argument.

The fail here lies with the browser developers who have
misguidedly put their weight behind the red herring that is
HTML5 instead of getting their **** together and
implementing XHTML, XForms, and getting behind evolving
the preexisting standards.
sobri
29th Dec 2008
- Reply to
- Flag
You missed my point.

The overall concept of CSS is a great concept. The implementation is awful.

HTML5 is a response to the lack of updates to the XHTML & CSS standards... I think HTML5 is the wrong solution.
mikefarinha
30th Dec 2008
- Flag
Not the future, the end.

Proprietary standards will be the end of the Internet as we know it. If you got your MBA at Yale or Harvard then you were told that to be a market leader means to fragment the market.

Your world of corporate owned standards is a bleak one indeed. Only the stockholders benefit, the end user left with needless complexity and cost.
kozmcrae
29th Dec 2008
- Reply to
- Flag
Too much ideology.

You are blinded by ideology and have a very narrow view of markets and corporations.
mikefarinha
30th Dec 2008
- Flag
Just because they're corporations

Doesn't mean they're not coming from an ideology. A business method is a kind of ideology. They can influence us to choose them or they can remove all other choices like Microsoft has. It's ok to select fewer choices as it appears you have. It's also ok for me to want more choices. And if I'm dreaming then Microsoft isn't preparing to make 15,000 layoffs later this January.
kozmcrae
1st Jan 2009
- Flag
Nutjob

Yes, everyone running Silverlight or Flash apps with real interfaces and more efficient usage of bandwidth will be "the end of the internet." Honestly, at what point did your brain cease to function? Or did it ever?
jackbond
30th Dec 2008
- Flag
Wow, I just look like a genius

Said it before, and I'll say it again.

HTML 5 = TOTAL JOKE

Why bother implementing something so completely worthless? The purpose of the HTML 5 working group isn't really to develop the next version of HTML, it's a case study on design by committee. Instead of Moore's law, why not talk about the HTML5 law: something about the inverse relation of the value of a product to the number of people who designed it.
jackbond
29th Dec 2008
- Reply to
- Flag
Why limit the discussion to HTML5?

That was the heart of my January story.

But as others have noted here, why limit the
discussion to this standard and this group?

I should do a second book on Moore's Second Law, to go
alongside my 2002 book on Moore's Law.
DanaBlankenhorn
29th Dec 2008
- Reply to
- Flag
There's so much wrong with HTML5: backward-looking, monolithic, etc.

There's so much wrong with HTML5. Here's a list of some of the biggies:

1. Backward-looking rather than forward-looking. HTML5 wants to be completely compatible with all existing web content out there, it wants to document all parser quirks, error-handling mechanisms, useless and outdated tags, proprietary APIs. Regardless of the usefulness of these things going forward, they want them to all be part of HTML5.

When Microsoft took the IE team and made them the Avalon (WPF) team, they decided to build on a clean slate rather than be bogged down by the idiosyncrities of the past, and they ended up with something much less complex, more forward-looking, and more powerful for the developer and the end-user. HTML5 will be the opposite.

2. It's one monolithic spec. There are new features in HTML5 (video tag, canvas tag, new controls, storage APIs), but they aren't segregated into their own spec like they should be.

Take CSS. CSS 2.1 is a single monolithic spec that still isn't done after what, 8 years? CSS3 is separated into modules, some of which are ready to be finalized once CSS2.1 is done. HTML5 would be much more likely to succeed if it were modularized as well.

3. The motivations and politization of the parties involved. HTML5 started as a separate effort by the "WHAT WG", which essentially was Mozilla+Opera+Safari -- in other words, the (at the time) "little people" in the browser market -- started because they saw the W3C not moving effectively with XHTML.

When the W3C decided to restart the HTML working group, the members of the WHAT WG only agreed to participate if 1) the draft of HTML5 they had worked on would be adopted *as is* as the W3C's first draft, and 2) the editor of the WHAT WG HTML5 spec, Ian Hickson (a paid employee of Google whose sole job is to write web standards), would be made the editor of the W3C spec as well.

Issue #1 (backwards-looking) above is quite obvious if you look at the members of WHAT WG and their motivations: they want their browsers' marketshare to increase and the only way for that to happen is for them to increase their compatibility with existing web content.

In all, HTML5 has a number of issues. These are just some of the main ones. It's in worse shape than the early pre-reset Longhorn project; they need to take a step back and start over and come up with something that's truly better for the long run of the web. Otherwise, as another poster noted, they'll lose out to Flash, Silverlight, etc. (Why do you think that Microsoft turned the MSHTML team into the Avalon (WPF) team back in 2001? Because they wanted to make something that was actually designed out to do the things that develoeprs want to do, and that users want it to be able to do.)
PB_z
29th Dec 2008
- Reply to
- Flag
Well said

I've been saying exactly the same, since Hickson first started this
debacle. The initial premises and goals of HTML5 are misguided, and
are setting the standards initiative back years, setting it up to lose
out spectacularly to proprietary solutions.

HTML5 is designed by and written by browser developers who're
looking to simplify their browser development burdens, to create
something that's a simple evolution of what already exists today.
Now, that would be fine if a) we didn't already have that with XHTML
and friends, and b) the rest of the industry wasn't building up
towards the next generation of internet based applications, way
beyond HTML5's timid goals.

What should have been done is for momentum to be put behind
XHTML and friends, and for those standards to be evolved towards
next generation internet applications. Instead a massive time sink,
dead end tangent has been created that's going to leave HTML
looking increasingly archaic, misguided and irrelevant over the next
few years.
sobri
29th Dec 2008
- Reply to
- Flag
RE: The HTML standards process grinds on

We can only blame one company here, Microsoft (their whole Embrace and Extend business plan) . If Microsoft had of conformed to standards in the beginning, we wouldn't be in this mess, Standards are there for a reason. Also @PB_z, who do you think this baggage is coming from, you guest it Microsoft (Microsoft being part of the W3C group). Their doing what they did to ODF and IPv6).

As for support for this new standard. IE 8, Firefox 3.1 (and any web browser that runs the Gecko Engine), Opera 11 and Saferi will have support.
ziggyfish
30th Dec 2008
- Reply to
- Flag