[microformats-discuss] hSpec

[Tantek Ç elik]

On 8/27/05 2:10 AM, "Danny Ayers" <danny.ayers at gmail.com> wrote:

> On 8/26/05, Mark Pilgrim <pilgrim at gmail.com> wrote:
> 
>> I've been thinking about creating hSpec, a microformat based on RFC
>> 2629 http://www.faqs.org/rfcs/rfc2629.html
> 
> Good idea. 
> 
> An immediate benefit would be rationalising the creation of RFCs, and
> this is exactly the kind of material that could be usefully
> semantically indexed.
> 
> Hmm, but how much of rfc2629 is actually spec-specific? There's a fair
> bit of overlap with existing XHTML constructs (like lists) and things
> for which there are already microformat representations like hCard.

That's a good thing, that means that such an "hSpec" could reuse lists and
hCard and other things for which there are already microformats
specifications.
 
> There does also seem to be overlap with biblio vocabs like DC and
> PRISM, maybe some of this could be factored out into more
> general-purpose profiles..?

There are folks working on a microformat for publication style citations.

 http://microformats.org/wiki/cite


> I don't know, through RDF lenses it wouldn't matter too much if hSpec
> did redundantly duplicate terms found in other microformats,
> equivalents could be resolved later. What does the microformat style
> guide say on this?

It would matter through microformat lenses.  New microformats should reuse
existing microformats as much as possible, and avoid reinvention.

E.g. hResume is reusing hCard, hCalendar, and rel-tag.

Sometimes a larger, more compound microformat will surface the need for a
smaller, more elemental microformat.   hResume helped surface the need for a
"citation" microformat.  Such "simpler", "smaller" problems ought to be
solved first when possible.

Once the examples on the web of implied specification schemas, and explicit
specification formats have been documented, the next logical step would be
start brainstorming about what a minimal spec schema based on those would
be, and then, which microformats could be reused for portions of that
schema.

Thanks,

Tantek