Towards objective evidence on the quality of the ODF and Open Specifications

by Rick Jelliffe

The subject of the quality and completeness of standards is very much on people's minds. I suggest that the best objective evidence does not come from armchair review, necessary as that is in any case, but from implementation experience.

The document Known Issues from the OpenXML/ODF Translator Add-ins for Office project at SourceForge is currently the best source of objective evidence.

Now first a caveat. The translator project is concerned with issues for doing a direct conversion between formats. It is not interested in issues of information preservation as such: so we should not expect to see comments related to round-tripping, where going from one format to another and back again result in the same document structures. So the list misses out on features that can be faked or worked around.

So what do we find in the sections Undocument Features in the ODF Specification and Undocument Features in the Open XML Specification which relate to Word Processing documents?

In the case of ODF, they identify three places where Open Office does things differently to the ODF spec. Now of course here they are saying that the standard should support the product, and I don't know where these issues fit in with ODF development. But nevertheless people do take Open Office to be some kind of primary source or even reference implementation for ODF But three little issues sounds something to be proud of.

In the case of Open XML, we find ....(drum roll)...four issues. Two of these are editorial problems in the draft specification and two are where the draft does not quite capture all the constraints of Office 2007.

Now I don't know that the issues list for the converters for Spreadsheets and Presentations are complete enough to be usable in the same way.(There are five documentation issues so far for spreadsheets for Open XML and one for presentations in open XML.) Nor do I know that these lists were systematically prepared and so include all the issues they found. Nor the backgrounds and skillsets of the developers, nor external help they had in reading the various specs.

That not withstanding, I think that the converter project is mature enough to provide some good objective evidence of the quality and completeness of the specs for the particular important issue of conversion. And my interpretation of that evidence is that the ODF and Open XML specifications are both surprisingly good: I would have expected dozens of issues.

Strictly, what this issues list means is that Open Office and Office 2007 do generate documents that follow their relative standards, and that is evidence of the completeness of the standards for working with their major existing implementations.


Peter Sefton
2007-07-29 23:37:23
The small number of issues reported by the converter projects may indeed give a sense of the maturity of the specifications but for interoperability (and your great idea in a recent post of packaging multiple formats together) mapping between the specs becomes important.

I work on a project where we care a lot about real interop, with the ability to round-trip not just one way conversion and there are a lot more that three or four issues in that area. (This comment is really more relevant to your other post).

Rick Jelliffe
2007-07-29 23:52:21
Peter: Yes. Thanks! Write more!!

(Actually, Peter's blog is a really great place for information of the practical issues, for example issues with lists. Essential reading though not focussed on standards.)