Re: Proposal for ISSUE-11: Default prefix declarations

I am fully and absolutely in favour of having such default xmlns sets.
But, I think, the main challenge we have is to decide what these will
be. To be more exact, a mechanism that gives us the possibility to have
that list updated ever after an RDFa 1.1 rec has been published. Ie,
having this list being part of an RDFa 1.1 is probably not a good idea.

First of all, at least in my view, this option is related to ISSUE-1
a.k.a. ACTION-4, ie, the flexibility of defining vocabularies, maybe via
an indirect mechanism. _If_ we have such mechanism, then the RDFa1.1
text can say that a default @vocab (or @profile) file, maintained at
this-and-this-W3C-URI, is used by all RDFa files by default. Because the
this-and-this-W3C-URI file would not be part of the REC text, it would
then be possible to have an update of that content time to time, via
some properly managed procedure.

What this procedure would be? I think some sort of a social procedure
would be necessary: people may come in and suggest a new
prefix/vocabulary via a public mailing list; there may be some sort of a
minimum deployment requirement or some other quality check; after a
certain time, if there are no objections, the namespace could be added
to the 'central' vocab file. (Some sort of an RSS feed for changes, or a
mailing list, can be uses so that implementers may get an automatic
notification when the file is updated and they can update their own
local cache.)

There is a precedence for something like that at W3C, called the
XPointer Registry[1]. The policy is described in[2]. Maybe something
like that could work for us, too

Cheers

Ivan

[1] http://www.w3.org/2005/04/xpointer-schemes/
[2] http://www.w3.org/2005/04/xpointer-policy

On 2010-3-1 05:23 , Manu Sporny wrote:
> On 02/28/2010 10:58 PM, Shane McCarron wrote:
>> +1.  The only risk I see associated with this proposal is that people
>> might assume that some prefix is pre-declared when in fact it is not
>> (e.g., my favorite vocab is skiing: - that MUST be in there).
> 
> True.
> 
> Also, I don't know if we want to put a proviso in the spec that states
> that default URI mappings will never change. Placing such a proviso in
> there could be an issue with vocabularies like Dublin Core, that have
> changed their URL from:
> 
> http://purl.org/dc/elements/1.1/
> 
> to
> 
> http://purl.org/dc/terms/
> 
> but ensured the vocabulary was backwards-compatible.
> 
> Maybe we could say that default URI mappings will always map to the most
> recent version of a vocabulary, or a backwards-compatible implementation
> of a vocabulary? This could protect future versions of RDFa from
> vocabulary rot or implosions of large companies (like Google or Yahoo or
> Lehman Brothers).
> 
> -- manu
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF   : http://www.ivan-herman.net/foaf.rdf
vCard  : http://www.ivan-herman.net/HermanIvan.vcf