[ODFPlugtest] Scenarios and What Conformance? (was Re: MSO does not read LO files)

[robert_weir at us.ibm.com]

None of this is really interesting from a specification or conformance 
perspective.  Microsoft Office supports ODF 1.1, not ODF 1.2. 

So, when Office loads an ODF 1.2 document, it is by definition processing 
a document that may have surprises for it.  The manifest version attribute 
is one.  But we also know from before, that they are not processing 
spreadsheet formulas in the OpenFormula namespace defined in ODF 1.2. 
There may be other such cases, where ODF 1.2 changes cause unexpected 
things to happen in Office. 

This has nothing to do with how late something was added to ODF 1.2. 
OpenFormula, for example, was added a long, long time ago. 

This has nothing to do with the attribute being mandatory.  Office 2010 
has no idea that it is mandatory, right?  And even if it were optional, 
the mere fact that it is in the document would cause this warning message. 
 Any new feature in ODF 1.2, mandatory or optional, has the potential to 
cause similar problems with an ODF 1.1 application that is implementing a 
strict policy about what it allows in documents.

It is simply a case that Microsoft Office supports ODF 1.1.  There is 
nothing more to this.   If you want a reliable way for MS Office to load 
your ODF, then save it in ODF 1.1 format.  Or wait for Microsoft to 
support ODF 1.2.

-Rob

plugtest-bounces at opendocsociety.org wrote on 06/29/2011 12:17:44 PM:

> From: "Dennis E. Hamilton" <dennis.hamilton at acm.org>
> To: "'ODF Plugfest mailinglist'" <plugtest at opendocsociety.org>
> Date: 06/29/2011 12:18 PM
> Subject: [ODFPlugtest] Scenarios and What Conformance? (was Re: MSO 
> does not read LO files)
> Sent by: plugtest-bounces at opendocsociety.org
> 
> I like your test.
> 
> Why not just provide the file as part of the test, since one 
> question is about consumers.
> 
> With regard to who produces it, I think that is a separate scenario 
check.
> 
> Even though what it is all about is interoperability, from those two
> cases we know what the interop cases are.
> 
> It is, of course, perfectly reasonable that OO.o 3.3.0 and 3.2.0 
> (the one I still run) wouldn't produce the attribute, its 
> requirement wasn't in the specification yet (or not long enough for 
> anyone to have noticed, etc.).  So a late changed breaks whatever 
> their implemented anticipation of ODF 1.2 conformance is.
> 
> So when the validators check conformance, do they provide an 
> assessment of the particular conformance that they encounter?  I.e.,
> a Conformant OpenDocument [1.2] Spreadsheet Document shall use 
> OpenFormula.  A Conformant Extended OpenDocument [1.2] Spreadsheet 
> Document could still use the OpenOffice.org formula (cute that it's 
> an extension), or whatever the formula namespace and syntax that 
> Microsoft Office uses in their ODF 1.1 documents.  (This distinction
> also applies to certain use of formulas in Conformant [Extended] 
> Text Documents.)
> 
> Are these far more material provisions being reported on by 
> validators?  Or do the validators assume there is just one 
> conformance as was the case for ODF 1.1?
> 
> That sort of thing is probably useful to kick around at a plugfest 
> too.  What conformance are we talking about?
> 
>  - Dennis
> 
> 
> 
> -----Original Message-----
> From: plugtest-bounces at opendocsociety.org [mailto:plugtest-
> bounces at opendocsociety.org] On Behalf Of Hanssens Bart
> Sent: Wednesday, June 29, 2011 00:23
> To: ODF Plugfest mailinglist
> Subject: Re: [ODFPlugtest] MSO does not read LO files
> 
> 
> Hi,
> 
> >Several people mentioned that, and I fail to see how it can have
> >anything to do with a plugfest.
> 
> It is a good test to see what ODF 1.1 consumers do when opening an
> ODF 1.2 document (not only MS-Office, but for instance also Google
> Docs - don't know if GD supports 1.2 - or SoftMaker Office etc)
> 
> 
> >This is not an issue of having applications agree on a certain
> >interpretation of ODF. ODF1.2 as currently written unequivocably
> >requires that attribute.
> 
> It is also a reminder for ODF producers to check if the files they
> generate are actually valid ODF files (there are some fine validators
> out there)
> 
> Even while it is clearly defined in the spec, it is still a good test 
because
> 
> a) we know that not all ODF 1.2 implementations follow this spec
> b) it would be nice to know how various ODF 1.1 implementations
> handle ODF 1.2 documents 
> 
> Anyway, I've created a scenario for those who want to perform a test:
> 
> http://plugtest.opendocsociety.org/doku.php?id=scenarios:20110715:odf12
> 
> 
> Best regards
> 
> Bart
> _______________________________________________
> Plugtest mailing list
> Plugtest at opendocsociety.org
> http://lists.opendocsociety.org/mailman/listinfo/plugtest
> 
> _______________________________________________
> Plugtest mailing list
> Plugtest at opendocsociety.org
> http://lists.opendocsociety.org/mailman/listinfo/plugtest