- From: Sander Tekelenburg <st@isoc.nl>
- Date: Mon, 6 Aug 2007 10:40:21 +0200
- To: <public-html@w3.org>
- Cc: <wri-talk@webrepair.org>
At 16:32 +0900 UTC, on 2007-08-06, Karl Dubost wrote: > Le 4 août 2007 à 05:56, Ian Hickson a écrit : [... dropping <http://www.whatwg.org/specs/web-apps/current-work/multipage/section-presentational.html#wysiwyg0>] >> We still need a better solution for handling the two tiers of >> document quality, [...] > <meta name="conform" content="html5-bp"> > > html5-bp = HTML 5 Best Practices. It would acknowledge a set of rules > defined by the Web community and considered as "good HTML". What is the problem that is being solved here? Obviously I recognise a use for easy identification of quality. It's why The Web Repair Initiative aims to certify authoring tools: <http://webrepair.org/strategy/certification>. But I don't see how a mere string claiming quality, or lack thereof, could ever be useful -- it would be easy to be spoofed. The only way for it to have meaning would be if either it somehow can only be added to a document after the quality of the document has been established, which I don't see how to achieve, or if it would be verifiable. The latter might be possible, if it were combined with a mechanism to easily verify the claim. For certified authoring tools, and for individually certified sites, that might be possible through something like: <link rel="quality" href="http://webrepair.org/certificates/id" title="WRI certified"> pointing to a document that confirms that the web page in question, and/or the authoring tool that generated its code, are certified. Or to apply it to your idea of a HTML WG "best practices" document:<link rel="quality" href="http://w3.org/certificates/bp/id" title="W3C Best Practices certified">. (In which case W3C would have to bother to do certification.) But without any such confirmability 'built-in' to the claiming mechanism, it seems it would be an empty claim. Any crap site could insert <meta name="conform" content="html5-bp">. Obviously, any site could as easily insert <link rel="quality" href="http://crap.example/file" title="ACME certified">. But the machanism allows inspection of the claim. If it links to a certifier that is unknown to you, and/or you can see that their certificate is meaningless/useless, you know how to interpret the claim. [Aside: possibly such quality certificates could actually be in a machine-redable format, similar to SSL certificates. Or even easier, have the <link> element point to a 'secure resource' using https. Once the certifier has been granted trust, a claim can automatically be verified through verification of the SSL certificate.] -- Sander Tekelenburg The Web Repair Initiative: <http://webrepair.org/>
Received on Monday, 6 August 2007 08:43:50 UTC