W3C home > Mailing lists > Public > www-webont-wg@w3.org > April 2002

TEST: agenda for f2f breakout

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Thu, 4 Apr 2002 17:57:43 +0100
To: <www-webont-wg@w3.org>
Message-ID: <JAEBJCLMIFLKLOJGMELDGEJKCDAA.jjc@hplb.hpl.hp.com>

According to the draft f2f agenda we have a 90 minute break-out on Monday.

It is not clear to me who is drawing up an agenda for that. Since it might
be me, and I would be embarrassed if it were, and I didn't have one, here's
a first cut.

(If it wasn't meant to be me, sorry).


1: round table: test war stories (15 mins)
2: objectives and non-objectives (20 mins)
3: explore possible test types (15 mins)
4: identify first deliverables (20 mins)

I am excluding implementation discussion in this f2f.


1: round table.

We could each have the opportunity to spend a minute or two describing a
successful or unsuccessful 'test' experience (or any other experience!).
This is intended as an ice-breaker.

2: objectives and non-objectives

Test cases can be used for a number of different purposes

e.g.
- conformance testing
- test suite covering the functionality
- tests that document decisions
- tests that document critical questions and require decisions to be made
- performance tests

In RDF Core we have an issue driven test suite. If something isn't an issue,
then it doesn't get a test - so for RDF Core performance testing is a
non-objective.

(20 mins)

3: explore possible test types

I did an initial exploration
http://lists.w3.org/Archives/Public/www-webont-wg/2002Mar/0152.html

do we want to pursue this avenue at all?
(15 min)

4: identify first deliverables

Should we be jumping in at this stage and making a few tests to highlight
issues? Both Dan and I have been doing a bit of that.

Should we be more pushy about trying to get agreement for or against the
test cases we have?

We could jump into the LANG discussions by providing test cases very like
the RDF XML & N-triple pairs. These would be pairs of frame and non-frame
portions of the language that were in some sense equivalent.

Should we be arranging a test repository? manifest file? a working draft
that gives an overview of our non-existent tests?

(20 min)


and 20 mins for overrunning ....


Does this look appropriate?
Received on Thursday, 4 April 2002 11:59:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:49 GMT