Re: [oiic-formation-discuss] Draft Interoperability and Conformance TC formation proposal (0.2)

["Dave Pawson" <dave.pawson@gmail.com>]

2008/7/27  <robert_weir@us.ibm.com>:

>> 2008/7/26  <robert_weir@us.ibm.com>:
>>
>>
>> > Probably worth being explicit about what we mean by "doing tests". One
>> > way in which it could be done would comprise at least the following
> steps:


> You need a better dictionary.  Even MW Online says that "provision" also
> means a proviso or stipulation:

So 'doing tests' is all about a proviso or stipulation?



> More to the point, ISO/IEC Directives, Part 2 "Rules for the structure and
> drafting of International Standards" gives the following definitions:

Yet you won't reference
http://isotc.iso.org/livelink/livelink.exe/fetch/2000/2489/186491/186605/AnnexI.html

> I leave it as an exercise for the reader to figure out "what the heck has
> that to do with testing".

I'm still wondering.



>> Which is likely to cover around 55% of the standard.
>> The rest?
>>
>
> That should cover nearly 100% of the provisions of the standard.  The
> informative parts of the standard do not require testing.

Not according to my estimates. Currently stands closer to one half
of odf 1.1 (although I'm not yet half way through the spec)



>
>> >
>> > 5) The creation of a document which enumerates all of the test
> assertions,  categorized functionally, along with a statement of what preconditions
> are necessary to perform each test, what the expected results are, which
> test  files are needed for each test, and how the results are to be
> reported.
>>
>> Is this the program you said we shouldn't write? The 'glue'?
>>
>
> I said explicitly "document".  Where do you get "program" from that?
>
> The "glue" would be code that is tied to a specific ODF implementation,
> such as something that would add a "conformance test" menu item to
> OpenOffice and then perform an automated conformance test, or would
> semi-automate a test, by automatically loading a document, and giving a UI
> for the tester to enter their results.  This was #6.  Glue == Test
> Harness.

Simply because a 'document' won't do what you appear to claim?
When you inspect the current spec, it is clear that compliance
testing  will need 'glue' (programs) to tie
it all together. Validation, xml source level testing and manual
validation of an applications response to stimuli.
Or perhaps you have other ideas how a few hundred tests might
be tied together? I really can't see xyz macros doing this, all to
be re-written for the next application? Impractical

I'll leave you to your one man show Rob.
Enjoy it.

regards


-- 
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk