Re: [whatwg] Alternative method of declaring prefixes in RDFa (was Re: RDFa is to structured data, like canvas is to bitmap and SVG is to vector)

Hi Manu,

just a quick answer on this specific issue, noting that these are
personal views of the matter and not necessarily the group's.

Manu Sporny wrote:
> I have forwarded this to the RDFa developer mailing list, Eduard... just
> so that it's noted in a record somewhere public.
> 
> Eduard Pascual wrote:
>> Just some opinions:
>> On Mon, Jan 19, 2009 at 12:18 AM, Manu Sporny <msporny@digitalbazaar.com> wrote:
>>> Speaking as an RDFa Task Force member - we're currently looking at an
>>> alternative prefix binding mechanism, so that this:
>>>
>>> xmlns:foaf="http://xmlns.com/foaf/0.1/"
>>>
>>> could also be declared like this in non-XML family languages:
>>>
>>> prefix="foaf=http://xmlns.com/foaf/0.1/"
>>>
>> Could <meta name="prefix" content="foaf=http://xmlns.com/foaf/0.1/">
>> do the job? I think adding a new @name value for meta would involve
>> less work than adding an entire new attribute.
> 

Personally, I always regarded the @prefix solution as being conceptually
equivalent to @xmlns which also means that I would have the possibility
to add a @prefix on any element (thereby defining the prefix for that
part of the subtree). (This may also be a help in the cut-and-paste
issue that you referred to in your discussion with Henri and Ian.) Using
the <meta> element for the same purpose would either (a) restrict its
usage to where that element can be used in the current HTML spec, ie,
the header or (b) extending the content model of HTML to include <meta>
anywhere. In my view, (a) would seriously reduce the usability of prefix
definition; (b) would lead to much more work for HTML parsers, that is
why the RDFa group (in spite of its initial designs...) decided to keep
strictly to attributes.

Cheers

Ivan


> This was discussed as a possibility at one point, I don't quite remember
> what the argument was against it. Perhaps somebody else on the list can
> remember?
> 
>> In either case, may I suggest an alternative syntax, like
>> "foaf:http://xmlns.com/foaf/0.1/"? That's a quite subjective opinion,
>> but the "=" inside attribute values seem too wrong to me, and they
>> quite ruin the readability.
> 
> We had been contemplating separators other than '=', ':' did come up in
> the conversation as a possibility.
> 
> -- 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