Minutes 2003-10-22 / QA F2F

* QA Panel at WWW2004

We called Dave Marston (QA IG) on phone. Dave Marston presents <a href="http://lists.w3.org/Archives/Public/www-qa-wg/2003Oct/0008.html">his proposal for a Panel on Quality</a> at the conference <a href="http://www2004.org/">WWW 2004</a>. After a poll, all people in the room agree that it will be a good idea.

Discussion:

Interopability might be the central question of the panel. Is interoperability the main challenge, or should we explore other domains. We need a catchy title if we want to attract people to this panel, maybe something provocative. We will have to find panel participants who are willing to take the roles of pragmatists, idealists, etc. 

How do we start the discussion ? Should we talk about the W3C QA Guidelines. It might not be a good idea because it's very specific to W3C or at least we should have something more general among others organizations. 

The audience of WWW Conferences is very diverse. You have journalists, webmasters, researchers, developers, managers, etc. So we have to define the target, we want to attract people and maybe find more or less controversive topics. For example, people are afraid about the future. Web standards are supposed to be stable but they evolve. Is W3C relevant or not? Many kind of topics could raise discussion. We can have a progressive method by telling a story and the birth of Web standards, the issues, the problems of implementability.

Is the purpose of the Panel to advertise our work or is it to explain the work of Web standards and their importance? We have to define the target and the goals. We can do that in the next ten days, which will be deadline for proposing a panel. 

David Marston, do you have ideas of panelists? It really depends on the topics. Jeffrey Zeldman for example could play the role of the person who has to use the Web standards but he's suffering about bad support. Who could be a user victim? For example, people from Boeing or have large users base. Difficulties for developers like Thierry Kormann and implementation of SVG for Batik.

David Marston will look out for details about the minimal requirements for proposing a Panel and we will decide about the next action and precise the topics, titles, etc.


* Tools

Olivier reports on qa-dev mailing-list and <a href="http://www.w3.org/QA/Tools/qa-dev/">QA-DEV Structure</a>. It has been developped aside the IG list because we needed to find a new moderator for the markup validator and we decided to create a forum where developers could join and participate to the work of developping tools like checklink, validator, etc. The people in QA-dev have a status which is a bit like invited experts in the sense that some developpers have some rights on machines. 

	It's a kind of incubator for software developers. It's a structure that could be reused for developing Frameworks tools, test tools, etc. A few groups inside W3C have created their own tools. Maybe there's an opportunity for gathering the effort.

	Olivier explains what has been started to be developped for testing the Markup Validator. How we started to use n3 for developing a test framework for the tool. EARL has been used to develop as a reporting languages, but the vocabulary might be reused for test decriptions and in this sense EARL needs to be extended for it.

	One of the main issue raised by Alex Rousskov is that declarative testing doesn't take into account the dynamic aspect of it. Do we need to have a framework which will permit testing dynamic features? For example, Protocols and descriptive markup languages have very different conditions of applications. Patrick Curran explained that the programmatic part is not necessary an XML language. There's a need for a language where the Test Harness can be reused easily. It's more about the definition of the tests, the way the results are given back by the tests harness.

We have clarified the issue. In fact, we start to identify 4 parts

	- the test itself (whatever it is piece of code, text documents, xml, etc)
	- the metadata about the tests	(TCDL)
	- The test harness
	- The test result				(EARL)

We agreed that should work on a language to express description and results, and if the same language can do both, it might be benefitial.


* NIST

Lynne is reporting on contracted works.
NIST has contracted with Lofton Henderson and Dimitris for tools methodologies which are related to the W3C QA work as well. 

Dimitris Dimitriadis: The purpose of the first work is to document the Test methodology which has been used for the DOM Test framework and makes it generic to be used for other technologies. The first deliverable is a review of DOM Test Framework. The second one is about reporting best practices to achieve testing, conformance, licensing, etc. The third one is to document the DOM test suite methodologies. The last one is to develop a general test framework. This work will be in one form or the other submitted to the QA WG.

Lofton Henderson: Test reporting tools: Assess the state of art in this activity and generate the requirements for this tool. The most significant and the largest task is about Test Assertion Markup language, define the requirements, and create a beta version of the language and after comments the final version.
It will be good to define templates, tools, etc to help WG to use our documents. [TAML applies to markup of test assertions, conformance requirements, or whatever, in the specification itself.]

Schedule: Everything has to be done by the end of June/July 2004. The templates are very important. 


* Bigger scheme of things for QA WG/IG

- reminder
How should we organize the work of the WG? The activity is becoming more operationnal so we may need to organize the work. Assignments.
Until now, we had sometimes not been very good at respecting our deadlines. 
Do we have a management issue or do we need reminder, or is it because all people are too busy? Dominique Hazaël-Massieux can develop a system which is a reminder. The WG thinks it's a good idea. We just have to decide when AI must be reminded to the person. 

- teleconferences
Should we rearrange the way we organize our teleconferences. One of the possibilities is to assign a rotating chair at each teleconference who's in charge to present a topic. Another possibility is to remove the teleconference to have more time to work. But the email is sometimes not enough, teleconferences generate often better discussions. 

Lofton correction:
"a rotating chair at each teleconference who's in charge to present a topic" -- as I proposed it, I didn't mean *each* teleconference.  I meant one teleconference each month.  E.g., the 1st Monday of the month, or the 2nd Monday of the month.  That is when we would do a "guest topic, guest chair".  Each QAWG member would take a turn.  Any QA-related topic is fair.


AI: DHM, Setup a system for reminding people about their AI, 2003-11-01