[odf-discuss] Basic Compatibility Issues in Office 2007

[marbux]

On Feb 19, 2008 6:03 AM, Pete Harlow <peter.harlow at gmail.com> wrote:

> On 19/02/2008, Sander Marechal <sander.marechal at tribal.nl> wrote:
>
> > I think it's pretty clear that the current MS-Office is not compatible
> > to DIS-29500. It can't possibly be with all the changes and comments
> > still going on.
>
>
> This was going round in my head. If MS state that MS-Office is compliant
> with DIS-29500, to whom does it fall to validate that compliance? Presumably
> if that compliance is not stated, (and, heaven forbid, DIS-29500 gains ISO
> status) MS-Office remains ineligible for those procurements where compliance
> with an ISO file format standard is required?
>

JTC 1 Directives say at page 145 that the interoperable implementation of a
draft standard has to be demonstrable. I read that as requiring two
interoperable implementations and an adequate interoperability conformance
assessment procedure. (There's some detail about development of the
conformance assessment procedures.)

On the other hand, lots of standards get through JTC 1 without any specified
conformance assessment procedure or demonstrable interoperability, even
though it's required.

There is no OOXML conformance assessment method specified. Italy requested
that it be done and a test suite and reference implmentation be developed,
but Microsoft shined them on until after the ballot in the Ecma response.
Carlos Piana, a member of the Italy NB, blogged about it here.<
http://www.piana.eu/?q=en/disposition_comments>. Scroll down to the
"Conformance" subheading. He is not a happy camper. :-)

There is a two step validation process using different methods. First for
the specified part of the schema, then another method for the custom
extensions. Grep OOXML Part 4 for exhLst Over 570 "future extension" points
with no functionality described and only the exhLst attribute to identify
them when processing the markup. As I recall, it was 573 extension points.

Then look at the compatibility framework in Part 5. Implementers get to
choose what compatibility markup they will apply to what markup. One of the
compatibility tags is kind of interesting, the MustUnderstand attribute.
It's the only one with mandatory processing requirements. A conformant
implementation that doesn't understand the custom markup so tagged is
required to abort processing of the document. It's a real 1-way interop
weapon for Microsoft. Create a custom extension, tag it with the
MustUnderstand attribute, and flash an error message on the screen telling
folks they've received a document that's incompatible with MS Office along
with a hint that the folks who sent you the document should buy a license
for the Real Thing to avoid such problems in the future.

I think all of the above says something about how enlightening the the
validation process is as an interop conformance assessment procedure. :-)

But back to the beginning, I think there's a stronger argument right now
than MS Office OOXML being non-conformant. That's the argument that there
are no implementations. Can't be, because Microsoft has offered to change
the spec in response to the comments but JTC 1 hasn't decided whether to
accept those changes, add more, or just reject the thing. So I think it's
entirely accurate to say that no one knows what the OOXML specification will
be at the moment, and therefore there are no implementations.

Interoperable implementations? Forgeddaboutit.


Best regards,

Paul
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.opendocumentfellowship.com/pipermail/odf-discuss/attachments/20080220/4b40ccde/attachment-0001.htm