- From: Karl Dubost <karl+w3c@la-grange.net>
- Date: Sat, 26 Jun 2010 19:20:36 -0400
- To: tantek@cs.stanford.edu
- Cc: "Robin Berjon" <robin@robineko.com>, public-vision-newstd-request@w3.org, public-vision-newstd@w3.org
A few things that Tantek's good mail triggered. Le 24 juin 2010 à 20:25, Tantek Celik a écrit : > In fact I think W3C would do itself a great intellectual honesty service if it took all such RECs and reset them to CRs (to be fair most such RECs were published before the CR phase was introduced). Which makes me think that would be interesting to visualize that. For the benefits of everyone else, W3C's work has happened on 15 years and visualizing if there was any improvements (or not) would help understand. For a bit of background, when W3C created QA Activity, it has been very challenging to set things in place. Some of the difficulties I have encountered (when I was working Conformance Manager and still working for W3C): 1. Having companies committing resources to the QA WG 2. Having people committing time to create basic test cases in their own WG. 3. Having companies contributing their own internal test suites. 4. Having companies using test suites for marketing reasons 5. Having companies to give the name and the version of the products which was being implemented. It has improved a lot, and not all Working Groups are scoring the same. For example, Semantic Web groups have often adopted a test driven approach [1] to their works (RDF calendar, OWL[2], …) I would be inclined of having Working Groups doing that more often. You can NOT put a feature in the specification before it has a written test for it. It would create plenty of other issues and frustrations, but it might keep the specifications smaller and more reasonable (avoiding some features creeps). Interoperability reports are often suboptimal. [3] For example, you can have a double implementation of each features on 3 different products but not necessary consistently. Also there is the notion of Class of Products. A renderer and an authoring tool are not the same beast. The more we add constraints, the more we slow down the pace. * Slowing down gives w3c a bad image * Immature specifications give w3c a bad image. It is one of the issues to solve. Oh and if it was not clear. It has improved… a lot… but quality is always the aim to do better. :) [1]: http://www.w3.org/2001/sw/DataAccess/ftf2#specDev [2]: http://www.w3.org/2002/03owlt/ [3]: http://esw.w3.org/ImplementationReport -- Karl Dubost Montréal, QC, Canada http://www.la-grange.net/karl/
Received on Saturday, 26 June 2010 23:21:02 UTC