Re: Errors in the test suite

On Thu, 22 Apr 2004, David Carlisle wrote:
>>
>> But were those implementations then used to exit CR?
>
> yes but for cr we needed to show implementations of mathml, not
> necessarily a renderer that can render mathml-in-xhtml without any
> assistance in the way of a processing instruction.

I thought to exit CR you had to show interoperability. Requiring per-UA
hacks to get them to render MathML doesn't really count IMHO... (The
MathML 2.0 Implementation and Interoperability Report doesn't really
explain how the interoperability claims were verified -- for example there
are no test results as far as I can see.)


>> Yes; my point was they also have a lot more. For QA purposes, the ideal
>> test is one that has nothing but the test and the pass criteria; the
>> current XHTML tests have lots of other stuff that is distracting.
>
> A conforming mathml system needn't be able to render xhtml at all.

Sure, but the implementations I would be testing do, so... :-)


> To test such a system you need to render the .mml file and then look at
> the sibling image file, you don't get the convenience of the xhtml
> wrapper that shows them both together but that's the way it has to be if
> you can't rely on having xhtml support.

Well, you already have combined XHTML+MathML files, I'm just saying it
would be very useful to one day have combmined XHTML+MathML files that
only have the test and the pass criteria.

(The MML files alone aren't tests as they don't have pass criteria.)

-- 
Ian Hickson                                      )\._.,--....,'``.    fL
U+1047E                                         /,   _.. \   _\  ;`._ ,.
http://index.hixie.ch/                         `._.-(,_..'--(,_..'`-.;.'

Received on Thursday, 22 April 2004 09:44:14 UTC