Crete F2F minutes - June 16, A.M.

Minutes: F2F Crete IG Meeting
June 16, 2003 A.M.
--
Scribe: Mark Skall

Attendees:

(KD) Karl Dubost (W3C, WG co-chair)
(PC) Patrick Curran (Sun)
(DH) Dominique Hazaël-Massieux (W3C)
(LH) Lofton Henderson (CGMO - WG co-chair)
(LR) Lynne Rosenthal (NIST - IG co-chair)
(MS) Mark Skall (NIST)
(OT) Olivier Thereaux (W3C)
(DD) Daniel Dardellier (W3C)

Regrets:

(PF) Peter Fawcett (RealNetworks)
(SM) Sandra Martinez (NIST)
(dd) Dimitris Dimitriadis (Ontologicon)
(KG) Kirill Gavrylyuk (Microsoft)
(AT) Andrew Thackrah (Open Group)
(DM) David Marston (IBM/Lotus)


Summary of New Action Items:

AI-20031016-1  DD to answer question in last call about relationship of QA 
specs and guideline to ISO 9000 by July 16, 2003.
Agenda:  http://www.w3.org/QA/2003/06/f2f
The agenda for the morning is as follows:
o       Introduction, Logistics, Agenda review/approval.
o       Specific: Report from AC Meeting (DD + OT/DH)
o       Specific: new EO project(s) -- cost-now vs. cost-later for QA
o       List policies / logistics review
o       Other IG and outreach topics TBD


Minutes:

IG Meeting

DD - QA activity summarized and future of activity discussed.
Feedback from internal activity.
First thing discussed was that quality has a price  does W3C still want to 
pay for it? Recognize there is a cost to QA process  more process and more 
time spent on spec; need recognition early on.  Benefits are obvious.
Feedback positive.  Lower cost if done early on.  If early, cost is on 
QA.  Later on it’s spread among the community.  Director is supportive.

Process used to deliver guideline was discussed next.  Should QA guidelines 
be on Rec track?  Advantage is better opportunity for member/WG 
review.  Consensus is that QA should do the process they think is the best. 
Rec track is ok.

Certification was the third question.  Strong opposition.  However, some 
people in favor like big user companies.

Are liability issues important to W3C.  They may be but W3C will try to 
have certification IF the membership wants it.  One certifies that a test 
has been passed; not that the implementation is correct.  We should think 
about a candidate for the first certification if we proceed with 
certification.  Self certification may be an option.  Legal liability would 
probably be no different but costs would be lower.  Certification is for 
advertising and procurement.

Next step is to identify 2 or 3 W3C technologies that have enough quality 
test materials to be considered for certification.  SVG may be a candidate.

LR  Id a company passes the test do we have netter interoperability because 
there are so many options?

KD  If we take SVG, do we really improve implementations that much since 
implementations are pretty good now.

DD  First product is just an experiment to see problems, liabilities.

LH  2 problems.  SVG full is way too big and many don’t implement all.  SVG 
tiny does not have conformance.  SVG tiny is for cell phones and they might 
want conformance.

KD  If everyone doesn’t implement everything how has it helped?

MS  If certification results in more companies running the tests then we’ve 
improved quality of the web.

DD  Can we comprehend what‘s going on if there are too many components to 
certify?

MS  Will certification improve quality of web?  It won’t improve vendor 
implementations since they are already using the tests since they’re 
free.  It will improve user applications since they will know the best ones 
to procure.  But what are the costs?  Let’s have an rationale for why we 
want certification then do a cost-benefit analysis.

DD  Disagree that everyone is running the tests.  Some companies don’t 
know  tests are available.  Certification would be a good way to advertise.

DD  Perhaps first step is to look at logo validation and see if people are 
using them legally.

LR  Is there a benefit for people that have been using branding icons?

LR  Maybe certification will make tests known to companies outside of W3C.

DD  Goal is that we want more people to pass the tests.

MS  Goal should also be to allow users to make better informed decisions.

DH  Key point is timeline of spec implementation.  Implementations that 
already exist are already instantiated in the market.

MS  Certification may lead to tests being developed later on in the process.

Last question was taxonomy of certification.  for web content, for 
software, for service providers (certifying people).

Conclusions

4 tiers of testing/certification:

No organized W3C effort
         People using tests
         People not using tests
Self-Assessment organized
         SVG
         WEbCGM
Self-certify with icon
         (lax) CSS, WAI, HTML
         (strict) none in  W3C
Formal certification
         None in W3C
         ATA CGM

KD  Can outreach and education have same effect?

Conclusion  One can take another path to achieve more usage and passage of 
the tests. The other path is better communication, more advertising, 
easier-to use, free tests.   However this path doesn’t preclude 
certification later.

PC  We’re jumping the gun in certification.  We (W3C) doesn’t even have the 
resources to develop and and/or ensure good test suites.  This is a 
necessary prerequisite to certification.

DD  How well is outreach working?

KD  Different parts.  Articles on web sites  have complete framework that’s 
complete and useful.  LR does excellent job for Month in QA.  Announcement 
on Chairs mailing list caused awareness of QA work.


A discussion ensued regarded re-chartering the IG.  Each aspect of the IG 
charter was reviewed.  The QAIG charter includes:
·       how the QA activity is moving forward (i.e. QA of QA)
o       This is ok  it’s what we did this morning.
·       sharing experience in validation of Web content, documents and 
protocols;
o       This is ok.
·       first to review QAWG deliverables
o       Is this going on?  Should we include it when we re-charter?  DD 
suggests removing “first”
·       discussing QA role in W3C standardization cycle (from early draft 
to Candidate Rec);
o       Mostly discussed in WG
·       issue with external funding, partnering with commercial sector, IPR;
o       Mostly WG matter.
·       coordination with external organizations
o       Mostly from WG
·       use of formal QA practices (ISO9000, etc)
o       We don’t do.  If we will, it will be WG.
·       how to address certification: branding, service, logo, metrics;
o       Yes, we did that.  Very important; also done within W3C.
·       education issues: QA tutorials, documentation, etc.;
o       Yes.  Done in IG.
·       evolution of the Activity into a more comprehensive domain to 
guarantee full compatibility between all W3C Recommendations and the 
unification of the various groups doing specification reviews (like WAI PF, 
I18N, TAG)
The meeting adjourned at 1730.