[odf-discuss] Alex Brown responds to Rob Weir

[Daniel Carrera]

I see a disconnect between what Alex Brown is complaining about, and the 
way he is testing it. Perhaps someone who knows the technicalities 
better than I can tell me if I got something wrong, but let me try:

1. Alex Brown says that ODF requires every IDREF to point to an ID, and 
requires IDs to be unique inside a document. So far so good.

2. Relax NG DTD Compatibility is an extension of Relax NG. It adds 
additional optional features. One of them checks whether IDs and IDREFs 
are as described above.

3. The ODF schema does not use that extension feature.

4. Alex Brown concludes that this means that the ODF spec is broken.


I don't see that. I don't see that it is necessarily the schema's job to 
check that IDs and IDREFs are sane. Sure, it is great if it does, but I 
don't think that it is broken if it doesn't. A complete ODF validator 
should also validate aspects of ODF that are not reflected in the Relax 
NG schema. Right? So the validator can check that IDs are unique and 
that every IDREF points to an ID because the ODF spec says that this is 
a constraint.

Another problem with Alex Brown's line of reasoning is that there may 
also be features which are simply not possible to express in a Relax NG 
schema. The Relax NG is not the entire spec. It is part of the spec. The 
ODF spec says "for a document to be valid ODF it must satisfy the 
following Relax NG schema as well as the following additional 
conditions". For example, you can't use Relax NG to check that the .odt 
file is a valid Zip file. Does that mean that the ODF spec is broken?

Daniel.

marbux wrote:
> <http://news.zdnet.co.uk/software/0,1000000121,39411118,00.htm>;
> <http://www.griffinbrown.co.uk/blog/PermaLink.aspx?guid=ace3b1c6-7ce8-49c7-8485-1ff8c34b7038>.
> _______________________________________________
> odf-discuss mailing list
> odf-discuss at opendocumentfellowship.com
> http://lists.opendocumentfellowship.com/mailman/listinfo/odf-discuss
> 
>