Re: Another viewpoint on validation

On Sat, April 14, 2012 1:03 am, Arved Sandstrom wrote:
> Probably a good start, let's get an idea of what people would like to do.

That was the purpose of the Wish List page [1].  It already has a few
entries on it. (And, IMO, adding something to the Wish List does not mean
that you've just volunteered to do the work on it.)

...
> In the context of this group I'd like to:
>
> 1. Provide elbow-grease help with tests;

We've talked about validation and about tests (not least thanks to you)
and they're not really the same thing.  Note that doing tests isn't on the
current Wish List.  Should it be?

Producing tests can be a gruelling amount of work (as Ken Holman can
attest from the OASIS XSLT Conformance TC days) but at least we are the
people (ppl?) best placed for deciding things when there's a question
about interpreting the spec (which relates to your point below).

Of course the easy answer is to ask the vendors for their tests, but we
don't even have all the vendors here.  The vendors that we do have may be
able to comment on the ease or otherwise of just handing over test files,
and Liam may be able to comment about how the test cases were accumulated
and the test suite opened up for XSLT/XQuery, and maybe David may have
some experience from MathML.

I don't even know about the test cases from the open source formatters
since they are under different licenses but may need/want to be under a
W3C license if they're to be promulgated by the W3C.

So the XSL 1.0 and XSL 1.1 CR tests (and the NIST tests if we can find
them) are the low-hanging fruit since they're already submitted to the W3C
and/or are public domain.

> 2. Participate in an effort to identify "fuzzy" or ambiguous spec
> pieces, and attempt to deliver group interpretations of how
> implementations should behave. These certainly existed when I was active
> with FOP and I'd have to assume they exist now;

They do.  One came in in the last month or so, and I was hoping the
submitter would also raise it here.  And there's some doozys about tables
from Vincent that may still be partially unanswered.

That would fit with the 'water cooler' aspect from the proposal for this
CG, and it would reduce any resemblance to the "Cascade of
Attention-Deficit Teenagers" model [2] if we only focused on XSL-FO 2.0.

> 3. Continue participating in the discussions with respect to validation:
> how best to validate, what can we do to help, etc.

That is another topic, yes.  Between polling and being away, I've lost
track of where we're up to on that.

> That's a start for me.

Thanks.

Regards,


Tony.

[1] http://www.w3.org/community/ppl/wiki/Wish_List
[2] http://www.jwz.org/doc/cadt.html

Received on Monday, 16 April 2012 14:33:14 UTC