[odf-discuss] OOo OOXML filters

[Alex Hudson]

marbux wrote:
> I raise the question because as I understand it the lack of XPath
> capability in EOOXML means probably 60 per cent fidelity in ODF <>
> EOOXML XSLTs at best, as compared to probably 85 per cent or so
> fidelity using OOo's traditional file import processes.

I don't think that's true - at least, OXML's inclusion of XPath is
neither here nor there.

What is true is that the MS Converter is using post-processing: e.g.,
OpenDocument uses Unicode to handle unbreakable spaces; OXML uses XML
tags. You could convert those with XSLTs, but it would be horrendous and
slow.

Ditto, the style models of the two formats are sufficiently different
that converting in XSLT would be painful.

>  For example, if Microsoft implemented the following
> requirement from ODF section 1.5, its Office customers would have both
> forward and backward compatibility from that version of Office
> forward. I.e., no more upgrade treadmill.
>
> "Conforming applications shall read documents containing processing
> instructions and should preserve them [the processing instructions]."

I think you're reading that a bit too widely.

Processing instructions are things embedded into XML <?like this ?>.
OpenDocument doesn't define any, and I've never seen them used in an
OpenDocument, aside from the <?xml version="1.0"?> standard PI at the
start of a document.

What it doesn't mean is that if the OpenDocument app comes across markup
it doesn't understand, it preserves but ignores it - I don't think there
is any such requirement, and it would be quite hard to do in many ways
(e.g., how can that application tell if the tag is just standalone, or a
range marker? How can it tell if it can/should alter information within
that range?)

Cheers,

-- 
Alex Hudson
IT Director, Stratagia Ltd.
P: 0845 226 17 13     W: http://www.stratagia.co.uk/
F: 0845 226 17 14     E: alex (at) stratagia.co.uk