Experts have defended the Open Document Format standard against suggestions that its schema is broken, but the critic who highlighted the alleged flaws has defended his position.
Alex Brown, a major contributor to the progress of the ODF's Microsoft-backed rival, OOXML, within the international standards process, carried out a "smoke test" last week, which he said showed the OpenOffice open-source application suite does not produce documents that conform to the ODF format — and also showed the schema defined in the OpenOffice standard is broken.
Other document experts have responded to Brown's post, criticising his methods and results, but Brown maintains there is a problem with ODF that should be fixed.
The standardised version of ODF, known as ISO/IEC 26300:2006, "has a defect which prevents any document claiming validity from being actually valid", said Brown last week. "Consequently, there are no XML documents in existence which are valid to ISO ODF." He had uncovered the problem while preparing to test whether the output of OpenOffice genuinely conforms to the ISO/IED 26300:2006 standard.
"Alex Brown has a problem," responded Rob Weir of IBM, who works on ODF at Oasis, the group that submitted ODF to ISO (the International Organization for Standardization). "He can't figure out how to validate ODF documents." Weir raises two apparently elementary flaws in Brown's tests, in a post pointedly titled "ODF validation for dummies".
Brown was validating OpenOffice's output against the wrong version of ODF, according to Weir — he tested the current version of OpenOffice, which claims to output the 1.1 version of ODF — a newer version that that in the ISO standard.
Brown says he knew what he was doing and the test was partly performed to see the divergence between the standard and the products which implement it — as he did in a similar test of OOXML and Microsoft Office 2007 — and suggests Weir's claims back up that part of his story. "The point of the test is to see how far adrift they are," Brown said to ZDNet.co.uk.
More fundamentally, the two disagree on whether the schema in the ISO version of the ODF standard is broken. Brown used jing, a validator for Relax NG, a schema language that is used in both ODF and OOXML. Written by the same person who wrote Relax NG, jing can validate schemas and then validate documents produced by them. Brown found that ODF had problems at the first stage, and claimed there is an ambiguity in the schema it uses.
Weir said in his post that this can be ignored by turning off an option in jing, but Brown argued in a response (titled "ODF validation for the cognoscenti") that this should not be done because it affects the way the schema handles ID and IDREF attributes . "IDs and IDREFs can't include ambiguity," he told ZDNet.co.uk. "As a result the schema validator can't understand what is meant."