RE: [oic] The Hague expense sheet scenario, continued

[Hanssens Bart <Bart.Hanssens@fedict.be>]

Hi Dennis,


> "Test point" in the sense of a specific measurement point in an
> electronic circuit is something I think of like that.  In our case, the test
> point definition would include the threshhold condition.

"Test point" sounds good


> It appears that we are asking that an ODF document of a particular
> nature (more or less) be created, but the tester has to know how to
> interact with the application in a way that achieves that result.
> I guess we are going to learn how to specify what the outcome should
> be in a way where a test developer would have to identify how that
> outcome is achievable, or not, with their product.

Correct. I'm focusing on office suites, and I'm trying to find a balance
between being too implementation-specific ("press F11") and too
"clueless" ("here's the raw XML, now you figure out how to recreate it")


In fact (I think I mentioned this before), a developer could create scripts
for their favorite office suite. IIRC, the KOffice guys already have some
scripts to create test documents for conformance / regression testing. 


> For example, how is it to be determined that the document conforms to
> ODF 1.1 (and that it is the intention to create an ODF 1.1 document) ?  


> Also, how are satisfaction of layout measures to be determined ?
> I assume the document must be rendered (on a printer?) in this case.

Rendering is challenging indeed :-) So the PDF I've added is merely a
suggestion, not "the" correct answer: how "bold" is "Bold" ? I have no
idea. Maybe the margin is 1.999cm instead of 2.0, that's probably
good enough. But if I see a red border while the XML says it's black,
that's obviously wrong...

> For example, what exactly does "Apply 'oic-page' to current page"
> mean? What is the intended outcome.

Well, maybe it is possible in implementation X to define a page style
(by clicking around or using a wizzard, whatever), but X might fail in
actually using / applying it.

> For example, test scenario step 1.1 assumes there is a UI provision
> for independently defining and naming page styles.  That's a big item
> with tons of dependencies (A4, portrait, metric dimensions, and UI
> language and locality-specific behavior).

Hmmz yes, once again I'm trying to find the right balance between a
scenario with truck loads of explicitly defined post- en pre-conditions
(cumbersome to write and not very encouraging for potential testers)
or being too vague...


Best regards,

Bart