RE: [oic] OOoCon and ODF tests

[robert_weir@us.ibm.com]

Andreas J Guelzow <aguelzow@math.concordia.ab.ca> wrote on 11/10/2008 
02:16:55 PM:
> 
> On Mon, 2008-11-10 at 05:48 +0100, Hanssens Bart wrote:
> > Hi Andreas,
> > 
> > 
> > Andreas Guelzow wrote:
> > 
> > >> Overall, formula interop was very good (at least for the
> > >> configurations that were tested that day)
> > >
> > >Where in the ODF specs are spreadsheet formulas specified?
> > 
> > They aren't (at least not in 1.1). ODF 1.1 implementations
> > tend to use an openoffice.org namespace or use a draft of the
> > upcoming ODF 1.2 OpenFormula.
> 
> Well, there aren't any definitions available for the ooo= namespace.
> Moreover while they might be using the OpenFormula Draft, I would more
> likely suspect that they typically implemented functions that emulate
> what MS Excel is doing. 
> 
> Since the OpenFormula draft is full of "implementation defined"
> sections, I would love to know where those sample formulae came from. 
> 

What we used were test cases from OpenFormula specification, extracted 
into a test file.  So the semantics were those defined by ODF 1.2's 
OpenFormula, but used in an ODF 1.0 document, using the ooo namespace.  So 
it was not a conformance test, but a test of how close current ODF 
implementations are to OpenFormula.

> > So unfortunately, for the time being, interop in this area
> > can only be achieved in a non-standard way.
> 
> I don't think that, with respect to spreadsheets, ODF and/or OpenFormula
> brings us any closer to interoperability. So I was quite surprised to
> read that spreadsheet formulas were used for this little exercise.
> 

I'm surprised to hear you say that OpenFormula does not bring us any 
closer to interoperability.  Can you explain what you mean a bit more? 

Regards,

-Rob