- From: Patrick Curran <Patrick.Curran@sun.com>
- Date: Sun, 09 Feb 2003 18:31:06 -0800
- To: Lofton Henderson <lofton@rockynet.com>
- CC: www-qa-wg@w3.org
Good summary - good ideas. Talk to you tomorrow... Lofton Henderson wrote: > Others have expressed virtually every point that I would make. See my > comments after theirs... > > At 02:10 PM 2/4/03 -0800, Patrick Curran wrote: > >> [...]key messages are. For example: >> >> * A conformance test suite is as important as the specification >> * Incorporate QA requirements and practices into your work >> * It's easy to get started (but expensive to do a good job) >> * We have documents and guidelines to help you >> * Here they are (pointers) > > > At 08:36 AM 2/5/03 +0100, Dominique Hazaël-Massieux wrote: > >> Le mer 05/02/2003 à 08:02, Olivier Thereaux a écrit : >> [...] >> > I'm looking forward to seeing the other's points, but my apprach today >> > would be: >> > >> > * you can't afford to ignore QA >> > * analyse what conformance to your spec(s) means >> > * turn this into tests >> > * Do it. Now. Don't wait for LC, or CR, or whatever >> > * we can help [pointers...] >> >> Mine would be pretty similar: >> * cost/benefits of QA >> * integrate QA in your WG (OpsGl) >> * integrate QA in your spec (specGL) >> * coordinate with us, exchange your QA experiences with others > > > At 03:30 PM 2/5/03 -0500, David Marston/Cambridge/IBM wrote: > >> Patrick proposed: >> [...] >> >From Karl's earlier draft [1], I would add: >> * Appoint a QA person to track what you do in this area >> >> Also, something like: >> * Three levels (A-AAA) of goodness exist >> * Test harness or automation may not be your burden > > > All good points. I don't really have new ones (except see #4 below). > But I have some thoughts about emphasis. Viewed from 30,000 feet, one > can summarize the points above: > > 1.) points to sell the case for QA > 2.) what we are doing in QA Activity to help them > 3.) what we (QA) think they (WGs) should be doing > > no one has mentioned, > > 4.) what do you (the WGs) want from us? (reviews, TTF, tools, > consultancy, hard labor, to be left alone, etc) > > My priorities/emphasis about others' points (above). I think that we > should be expecting a range of audiences, from just starting to deep > into test suites. For the former, spending some time on #1 is > appropriate. For the latter, it is "preaching to the choir", and > emphasis might better shift to #3 and #4. > > The points and perspective that I would favor for #3 are: > > * Easy to start (appoint QA Moderator now; read Intro; OpsGL > self-assess; SpecGL self-assess) > * 80/20 (or 90/10, or 70/30) rule: lots of benefit from initial easy > steps (GL self-assess, basic test suite) > * OpsGL is simpler than it might look (Quick tips summary) > * SpecGL is simpler than it might look (Quick tips summary) > > -Lofton. > > > > > >
Received on Sunday, 9 February 2003 22:07:26 UTC