Re: Feedback on RDFa Core 1.1

Hi Dan,

I'm not convinced the case is as clear-cut as you say. For example,
when it comes to caching documents such as OWL files and RDFa
profiles, then having a version number or date in the URL will greatly
help.

Similarly, when it comes to 'hard-coding' the processing of particular
profiles in a parser, then if you know that you process 'version x' of
a profile, you can code for that, and not worry that your code will
ever be out of sync with a parser that loads the profile dynamically.

Anyway, the particular URI referred to is a made-up one, and as
Stephane says, there is now a real bib vocabulary so we'll use that
instead. (Those examples are *very* old, from early 1.0 drafts, and
could definitely do with a refresh.)

And on using capital letters on class names...sure -- no problem,
although in the case of the single example that you are referring to,
it will fall out as a consequence of using a real vocabulary.

Regards,

Mark


On Sun, May 2, 2010 at 8:13 PM, Dan Brickley <danbri@danbri.org> wrote:
> On Sun, May 2, 2010 at 9:01 PM, Stéphane Corlosquet
> <scorlosquet@gmail.com> wrote:
>> Hi,
>>
>> This is feedback on the RDFa 1.1 core W3C Working Draft 22 April 2010
>> http://www.w3.org/TR/2010/WD-rdfa-core-20100422/
>>
>> [[[
>> biblio:         http://example.org/biblio/0.1
>> ]]]
>> any reason for not using an existing biblio ontology such a bibo:
>> @prefix bibo: <http://purl.org/ontology/bibo/> .
>> and further bibo:Book which is an existing class?
>
> Oh, I just noticed the "0.1" thing there. Dear RDFa WG, please do not
> encourage any more poor souls to put version info into their namespace
> URIs. Really! It's probably the single biggest irreversible mistake in
> FOAF. It started out as a silly little prototype and just kept
> growing, and there was never a right time to switch to a new ns URI
> that didn't contain "0.1'.
>
> Please don't encourage this practice. Instead the simplest rule for
> namespace URIs is the best: "if in doubt, leave it out". Looking back,
> many kick themselves for including stuff in a namespace URI, creating
> maintainance baggage (and I count here things like using your personal
> domain, since it puts your stuff and the ns in the same bucket
> forever). I think very few have kicked themselves with regret for
> leaving stuff out of a namespace URI. Particularly version numbers! So
> +1 on switching to real examples like bibo: and dcterms:, ... but
> regardless please drop the version number.
>
>
>> [[[
>>    '<span about="urn:ISBN:0091808189" typeof="biblio:book"
>> ]]]
>> how about using some common practice and capitalize RDF classes, e.g.
>> biblio:Book?
>
> Seconded. This might seem picky, but being able to do a first-cut
> parse of RDF simply by skimming for capital letters keeps a lot of
> people sane. Do please stick to those conventions...
>
>
>
>> [[[
>>  <span property="foaf:givenname">Albert</span>
>> foaf:givenName
>> same for _:a foaf:givenname "Albert" .
>> ]]]
>> s/foaf:givenname/foaf:givenName
>
> Yup, we fixed that one late last year in FOAF (keeping the old
> properties but marked as 'archaic' forms). If you can encourage the
> newer spelling that would be really great!
>
> cheers,
>
> Dan
>
>

Received on Sunday, 2 May 2010 22:07:53 UTC