Re: CSS discussion

I agree with you about valid markup. Validation must be performed as
part of preprocessing no matter what. The validation results are of
general interest to consumers of the moki document, not just the
tests.

But for the case of the MEASURES test, this test does not *have* to be
performed in preprocessing. Its results are specific to the MEASURES
test. Therefore I think it should be run as a test, and not part of
preprocessing. Otherwise we move towards merely running all tests
within preprocessing, and having elements like
"<PageSizeTestPassed>true</PageSizeTestPassed>" true in moki, and
trivially selecting these elements to implement the test -- in which
case we have merely pushed the whole implementation into
preprocessing.

I don't think there is a requirement that all tests be implemented
solely with XSLT, because I can think of several tests that can't. We
should view XSLT as a convenience and a quick way to implement some
tests. Where it is harder to implement in XSLT, don't use it.

Given that, yeah, I agree, we do not strictly need an XML
representation anyway. Laura's implementation does not use it, so
let's take a look at that next.

Sean

On 8/8/07, Miguel Garcia <miguel.garcia@fundacionctic.org> wrote:
>
>
> Isn't supposed that moki document should contain all information needed
> in order to be able to pass mobile OK tests with any XSLT parser?.
>
> I think it was discussed that the result of the tests will be achieved
> from moki document. Tests like Valid_Markup are done during
> preprocessing but they don't provide the outcome of the test they just
> include some information in moki and that information is checked later
> during XSLT processing.
>
> If capture all information in moki isn't a requirement then neither CSS
> representation in XML is needed for SAC parsing approach.

Received on Wednesday, 8 August 2007 12:50:02 UTC