See also: IRC log
05.11.2008 Participants: Shadi Abou-Zahra, Carlos Velasco, Johannes Koch, Yehya Mohamad
Test samples task force:
Starting point:
BenToWeb test suits: word suite changed to samples because suite does main the software is correct and sufficient. Samples does not have that meaning and does not imply the tool running the samples is accessible
Audience
Task force, review notes
People, who want to test their software
Goal for today's meeting:
Basic concept and a mock up for today morning
discuss it in a wider view
Create a stable concept as a start of implementation work.
Extensibility and maintenance are core issues of the architecture:-
CVS repository as a component
for reasons of extensibility and reusability, the architecture should stick to w3c architecture e.g. Using technologies like PHP, MYSQL etc.
Carlos presented the BenToWeb repository:
A cron job was synchronising with CVS every 6 hours
A static copy of the test (TCDL) was copied to the website
table automatically generated
The server (Cocoon) reads from the meta data and recognise the relevant data and generate them dynamically
SAZ: Are there filtering (sorting) technologies in the current system ? Answer No
There exists a human readable test viewable in a browser
There are different type of tests
plain HTML files
Change of the current BenToWeb JSP test files to the target language like PHP, python etc.
The Editor Parsifal would not be considered as it would mean a deep level involvement in Software development
Structure of CVS
metdata
test files
resources
CV: For tools FIT is working on development of the next version of their tool imergo, which will be able to run groups of tests
Accessing directly meta data vers. Accessing directly the HTML test files
Accessing the test files would be the start point, having a search page selecting criteria e.g. Technologies as a result would be
a tool friendly page: set of links
EARL output with expected results
Human friendly format
Accessing the meta data would be the next and ultimate step towards tool development implementing EARL
Search functionality
1.Technology � any, html, css, scripting...
2.Requirement - any, principle guidelines, success criteria
1.Guidelines, success criteria
1.Technics should show up
3.Level � A, AA, AAA
4.Select output parameters
Display defualt human readable output
Display just links to test files
Display expected results as EARL
SAZ:
we need to develop more test cases
AND a web interface to the test cases
3 Step search:
1.Select global parameters
2.Depending on the first selection select a level (optional)
3.Select the techniques
Output of the form is a URI pointing to a list of test files. This URI would be fed into the tools, which will then read all test files and process them sequentially and generate their own output.
File format:
Id � title � Techniques � Expected outcome - Test file - XML Meta data
Note: there is the problem of blocking IP's because of too much traffic! In this case the tool developer will create during the tests huge amount of requests to the w3C servers.
hi mike!
just back from lunch, setting up
http://www.w3.org/WAI/ER/2008/11/tsd-agenda
<CarlosV> xslt file for TCDL 2.0 files: http://bentoweb.org/refs/TCDL2.0/tcdl2.0tsdtf_to_xhtml.xslt
http://www.bentoweb.org/ts/XHTML1_TestSuite3/metadata/sc3.2.4_l2_002
http://www.w3.org/WAI/ER/tests/usingTCDL
<MikeS> we're in discussion - done with lunch
<JohannesK> mc: approach seems to be quite WCAG-specific
http://www.w3.org/WAI/ER/tests/usingTCDL#structure
<scribe> scribe: JohannesK
<shadi> http://www.bentoweb.org/ts/XHTML1_TestSuite3/metadata/sc3.2.4_l2_002
<Christophe> HTML view (sortable and with expected result): http://tinyurl.com/ynu7q4
<Christophe> http://canada.esat.kuleuven.be/bentoweb/tsdtf/xhtml/index.html
works in FF 2
<MikeS> will redial in 20
<MikeS> IBM's auto software installer just did an install and I need to reboot. will be back on in a few
<MichaelC> Tim http://www.w3.org/WAI/ER/tests/WebInterface/Mockups/design.html
<MikeS> trying to dial back in - not getting anything. anybody else try yet?
<shadi> yes, all here :)
<MikeS> +496940359803
<MikeS> using our internal VOIP SameTime suite
<shadi> yup, correct #
<MikeS> worked fine this AM
<MikeS> pls proceed without me; I'll listen in on IRC
<shadi> http://www.w3.org/WAI/ER/tests/WebInterface/Mockups/search.html
<shadi> http://www.w3.org/WAI/ER/tests/usingTCDL
<shadi> MC: sorty by pass/fail -- expected outcome
<shadi> MC: sorty by technology element -- such as "img", "table" etc
<MichaelC> Note I mean more "search by" than "sort by" those features, though sorting on them useful too
<shadi> http://www.w3.org/WAI/ER/tools/
<shadi> http://www.w3.org/2006/tsdtf/TestSampleStatusList
saz explains status list in WIKI
saz: can JIRA be a replacement?
CV: JIRA is an issue tracker
http://www.atlassian.com/software/jira/
CV: requires MySQL and Tomcat
<shadi> tool should be self-contained
<shadi> the basic interface should be a list as currently implemented in the wiki
<shadi> templates for the structure review, content review, and issues are available
<shadi> automated aspects such as schematron checking etc should be automatically built into the tool, rather than need to be installed separately
<shadi> + changing status (taking it over from the submitted reviews)
<shadi> http://www.w3.org/WAI/ER/tests/process
saz: we should think about how to get
information about the submitter into the database
... either remove status from metadata or have the script update the field
CV: wish: give public read-only access to CVS repository
saz: then the status has to be in the metadata