Problems of Standards Implementation

Steven Pemberton, W3C and CWI, Amsterdam

Chair, W3C XForms Working Group

Chair, W3C HTML Working Group

W3C: not really a standardisation body

Really: Open distributed software design

Starts with problem identification, then working group creation, who write requirements

Then working drafts with requirement to respond to public comment

'Last call'

'Candidate recommendation' with exit criteria

Finally: Recommendation

In other words: consensus based, tested for interoperability before release

Some standards

HTML4, December 1997: Successful unification attempt; 4-5 years to acceptance; still flaws in some implementations.

CSS1 Dec 1996: Successful attempt to do styling right. 5 years to acceptance. Initial objections from 'traditional' Netscape; Microsoft saw it as a way to get ahead after internet U-turn. Now widely implemented. Widely used, especially after demise of Netscape 4. 'Bugwards compatibility' problems.

XHTML1 Jan 2000: XML based HTML. Widespread adoption. Not though by 'traditional' Microsoft.

XForms: New. Most implemented W3C specification at release. Objections amongst traditional browser builders. XHTML2 follows.

Problems

Vicious circle: no demand - no implementation - no demand

Established browser builders tend to be conservative

Bad implementations: means less interoperability; creates bugwards compatibility problems; lack of certification.

Unfortunate need for advocacy groups: e.g. http://www.webstandards.org/

Monoculture: creates the vicious circle; much standards talk, little follow through; XML, platform independence, standards url now gone, no innovation

Conclusions

Most standards implementation due to grass-roots pressure.

Browser builders tend to be conservative.

New browsers tend to push the envelope; use standards as a method of selling value

Once browsers get established, they then slip back