- From: pfawcett <pfawcett@speakeasy.org>
- Date: Mon, 10 Mar 2003 09:02:03 -0800
- To: www-qa-wg@w3.org
Begin forwarded message: > From: pfawcett <pfawcett@speakeasy.org> > Date: Sun Mar 9, 2003 8:21:13 AM US/Pacific > To: Olivier Thereaux <ot@w3.org> > Subject: Draft Minutes from Friday PM > > QA Working Group Face To Face > Friday, 7-March-2003 Afternoon session. > -- > Scribe: Peter Fawcett > > Attendees: > (KD) Karl Dubost (W3C, WG co-chair) > (PF) Peter Fawcett (RealNetworks) > (DH) Dominique Hazael-Massieux (W3C - Webmaster) > (LH) Lofton Henderson (CGMO - WG co-chair) > (SM) Sandra Martinez (NIST) > (PC) Patrick Curran (Sun) > (LR) Lynne Rosenthal (NIST - IG co-chair) > (MS) Mark Skall (NIST) > > Regrets: > (dd) Dimitris Dimitriadis (Ontologicon) > (KG) Kirill Gavrylyuk (Microsoft) > > Guest: > Susan Lesch. > > Summary of New Action Items: > > > Agenda: http://www.w3.org/QA/2003/03/f2f > > Previous Telcon Minutes: > http://lists.w3.org/Archives/Public/www-qa-wg/2003Mar/0011.html > > > Minutes: > > Framework Completion Plan: > > TestGL > - 3/10 mon telcon will be for TestGL. > - Publish 2nd working draft around 20th. Likely wont make that date. > Will we need another telcon other than monday. likely yes... > - Second TestGL telcon scheduled for 24th march. > > - The 17th telcon will focus on OpsGL and SpecGL. > > - If we are not done with issues for Test on the 24th we will need > to schedule > another telcon at a time other than the regular time. > > - Publish another Working Draft of TestGL two weeks after the Second > Telcon (around april 7). > - The go to last call with TestGL by 6/4 > > - TestGl is running about 5 months behind. > > Issues processing on Ops and Spec. > - We should try and be done with issue processing by around the > 26th-27th. > Then we prepare for next cycle. > - what is next cycle? 2nd Last call or CR. > - one concern is that we may be rejected if we make too many > changes... > > - some comments about the organization of ops and spec. Loften > mentioned a desire to do some re-org but likely for V2. > Dominique, ops has to take into account time line of process and may > need > some reorganization as well... > > - If we are going to CR we need to be very exact and careful about > our > issue processing. Send responses to issue submitters, document all > issues > and resolutions. Document issues that we have decided are not issues > and > so on. > - Also need to prepare documents and do actual editing. > - Take about 3 weeks for the two documents. Taking us to Mid may. > > - Is there any other issues for going to CR. > Need CR Call. Meeting to cover CR. > - Once we go to CR, it will be 3 to 6 months (give or take) till we > can > finish CR. Part of the exit criteria is to have two implementations. > - Our current criteria is two implementations that conform to all p1 > and p2 > checkpoints. > > - That means that it will be late 2003 give or take to get out of CR > for OpsGL and SpecGL. TestGL is running behind and likely wont be > getting out of CR till early mid 2004. > > - Still not positive that the mid may date will be CR or 2nd Last > Call. > > - What does that put on our agenda for Crete. > - Tests for last call. > - What ever is going on with OpsGL/SpecGL > - Rechartering. > > - We still have not determined if our final goal is Rec or Note. > it seems that people expect us to go to Rec. > > Since we are dealing with process, the advisory committee may have > something to > say about our work. At what point do we start thinking about > integrating some > of these Guidelines in to the bigger process. For example pub rules. > > - Karl pointed out that we should comply to our own guidelines. We > need to at > least address the guidelines and determine if they are applicable or > not. > This should happen before CR. > > - There are issues for what our 'test materials' are. From the break > out group > on Thursday, we had determined that our checklists are our 'test > materials' but > it is some what different than conventional 'test materials'. > > - Karl would like to try to apply our own criteria to our selves. We > should > make an attempt to move our own process/orginization in the > direction we > hold others too. > > - Loften - of we get our process doc and an addendum to the charter > we can > start passing a number of the checkpoints that we currently fail. > > - This should be done before we are done with issues processing. > > - If we have the process document and the charter are updated we > should > re-evaluate our selves. > Put an update on this on one of the April Telcon Agendas. > > - Given all this the earliest that we can get to Rec is late this > year. > Test later, probably early next year. Perhaps around next plenary. > > TTF/QAWG future and re-charter. > > - Loften had heard a comment that TTF may violate the w3c process > document. > however Loften claims that section 4.1 does allow us to do this with > an > informal charter. > and as an aside to the above comments about conforming to our own > checkpoints, our current draft charter does not comply with our own > OpsGL in that we do not mention qa resources.... > > - QAWG2 will likely be mostly TTF and outreach mostly. > > - we have draft charter for TTF. We have some deliverables, scope and > success > criteria. > > - what do we need to do to the draft TTF charter in order to say that > we > are done with it? > > - Karl suggested to see if there is interest in the next charter. Let > people > know that it's almost done but we don't currently have staffing > resources > and is any one interested in giving us some help? > > - Deliverables can be included in task force charter as it's not as > formal > as a wg charter. Some of these may migrate to the WG re-charter. > > - Milestones is one part that needs to be filled in with uncertain > dates > as we don't have staffing yet. > > - List of possible deliverables for TTF. (from lofton with some > additions from > others, largely from out reach sessions ) > - TS Management framework. (Tool) > say take VoiceML testing tool that was demoed to us yesterday, > and generalize it for reuse. > - Test Assertion prototype/specification or markup. (Spec) > what ever that becomes should be moved toward incorporation in > style guide. We do not want to end up with multiple markup > guides in multiple places. > - Binary serialized infoset and comparator. (tool) > - Test Authoring Principles - (document). (from this morning > discussion). > have samples from css already. > - Test suite and Test Materials assessment and description. > (document/research) > an extension to the matrix perhaps. > describe how it is done in the w3c currently. > stay away from rating. Only description and assessment. > - Reporting/Earl/tools framework. (Tool) > - Other new templates. (document) > - Test case description language /metadata for testcases (spec) > > - who will build these tools. > > - Patrick - Sun is working on Tools to do some test assertion analysis > of specifications. If possible he may be able to leverage it and > move it into back into our group/w3c. > > - Likely two incarnations of such a tool, one for xhtml markup with > classes, the other for xml like xmlspec. > > - Susan Lesch, don't worry too much about manual of style as it can > be changed, > still a flexible document. > > - Once TTF Charter document is done and include above deliverables > and then > send out invitations for participation to help develop these tools. > > - If call for participation is not met favorably, we don't really > loose anything. > We will just be in the same boat that we are already in. > > Also even if we are over reaching and can't really build some of > these tools, > the templates have proven useful and people like them, so more of them > could be easy fast way to get some support out there and give us a > good affect > with out a lot of effort. > > - the XSLT group had done a very good job creating all aspects of a > test > framework including contribution and review of test cases. This is > one thing > that every group has to do so if we came up with a review process > template > it could well help many groups. > > - A test challenge process template is another one that every WG > needs. > > - There is a lot out there that is reusable, we can likely steal, > generalize and > reuse what is good. > > Patrick, Concepts/catigories from Sun in process and tools to help > process. > - tools for Program management collaboration, > - tools for test case generation > - test execution framework > - documentation for test harness. > - deployment, maintenance, update > > - Another tool could be XSLT tools that have a selection mechanism > that can > determine what tests are valid for what profile or implementation. > > - Now that we have a notion of some of the possible deliverables: we > can update > the scope and deliverables sections of the draft charter. > - Milestones can then also be started, at least with some of the > easier tools. > - One milestone would be a solicitation from other WG's about what > the priority > of tools to build should be, a determination of what tools are more > important. > what do groups need first. > - Loften - we can also grab the easier ones that are more obvious > while waiting > for comments on priorities. > - People may contribute more or other ideas that we have not thought > of. > > - Patrick agreed to be co-editor of document. > > - Summary: > Patrick will be co-editor. > Came up with list of first possible deliverables. > Seem to have enough resources to at least start on some of the easier > tools > Send out call for participation. and call for > comments/priorities/ideas. > > - We have been talking about this for a long time, lets see if we can > start > on a few of the small things. > > Back to TestGL: > - Loften's comments from his last review of the document: > - Intro needs to be updated with new intro that matches the headers > from > SpecGL/OpsGL and rework it as appropriate for TestGL. > Editor should do this likely. It also needs a lot of work. See the > next > few comments on specific issues. > > - Chapter 1 (intro) 1 second comment in paragraph 3 is likely out of > scope - WG agreed. > > - Chapter 1 (intro) 1 paragraph 4 / bullet list/5 the is really hard > to read. and understand. > Lofton suggested that first sentence of 5 could be first sentence > with > bullet list and nothing else. > > - Chapter 1 Section 1.1 - Really confusing and hard to read. > > - In this document comply, compliance is used. Should use > conform/conformance > instead. We do not test compliance, we test conformance. > > - 2nd bullet in 1.1 list, lofton disagrees with this point, it's > false. > WG agrees. > > - 1.5 "test area" what is it, better definition? or remove it. > also glossary should be moved to end as the rest of the documents. > > - Many checkpoints have multiple sentence long prose. Needs to be > shorter > and terser. Many of them will become rational, further descriptions > or > fulfillment criteria. > > - others are just long and not concise like 5.3 and 7.2. > > - Need to reformat as SpecGL/OpsGL. > Title, Conformance Requirements, This will be more than a formatting > excise. Lead editor will call on co-editors for help as needed. > Need to first do basic restructure. > > - Checkpoint 1.3 - The wording is bad. What we believe that we mean > is that > dimensions of variability should be taken into account. > By imposing grouping, it forces it to be a done in a single way. > It currently focuses on asserts. It should likely focus on test cases > so that tests can be selected based on dimensions of variability. > > - Much of the issues with this current document is that it focus on > the > "How to write a test suite" not "what a test suite should be" > > Move to adjourn. > Agreed. >
Received on Monday, 10 March 2003 15:17:07 UTC