OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

opendocument-users message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Re: [opendocument-users] RE: Foreign elements and attributes



> Now, I think dual conformance classes sounds like a good idea, but ODF
> TC must enhance the text describing it to a level that greatly
> exceedes the current detail level of ODF 1.2 CD01.

It depends on what the requirements actually are. "Banning extensions" is
a mechanism, not a requirement.

If the requirement is to prevent any document from having any information
that would not be known by another vanilla application, then yes, a
two-conformance class system with the strict class going a *lot* further
that the proposal would be necessary: banning the kinds of things that the
ODF Wiki has on the Extensions page, and including banning arbitrary
application settings as you point out. (And in ODF, lots of things are
dumped in this class.)

If, however, the requirement is merely to prevent miscegenated documents
where, for example, someone has taken a DOCBOOK file and wrapped it in the
minimal ODF elements (or put a real SVG graphics file where ODF is
required), then the idea of banning foreign elements makes complete sense.

But, as I commented, foreign attributes are different to foreign elements.
For all intents and purposes, elements provide basic semantics, while
attributes merely decorate. (I think the only attribute-only vocabulary I
know is XLink.)

If the requirement is "allow signatures" then all that is needed is to
nominate the particular information that will be used to derive the
signature. For example, a signature could be the whole document
canonicalized according to XML C14n. Or it could be the documents with the
foreign elements and attributes removed (i.e. to know whether the ODF
information had changed or not.) Banning extension elements and attributes
is not necessary for reliable signatures.

So what are the requirements? (I dare not posit any, because the last time
I tried I was greeted with accusations of conspiracy: a spade cannot be
called a spade in polite company it seems.)

Obviously having a single conformance level is simpler for all concerned
than having two, if such a compromise is indeed workable. Rob Weir will be
happy to know that my company has no tools that relate to complex
validation of ODF, nor plans for them, nor demand: so his claims that I am
raising this in order to suit my business are something that people need
not fear.

Cheers
Rick Jelliffe


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]