[microformats-discuss] microformats and semantic HTML best practices

[Danny Ayers]

On 9/4/05, Bryan J Busch <bryanjbusch at gmail.com> wrote:

> Or, when the time came to build the application that would index such
> pages, would it necessarily be upgraded to the status of microformat,
> or would it not even matter?

IMHO, that depends on one single thing - the existence of a profile
URI in the documents.

I agree 100% with the practice of reusing existing formats wherever
possible, but feel the "don't do this alone" attitude is unnecessary
and potentially counter-productive in a global environment.

Interoperability should and can be enabled by working from the same
base specifications, not through any potentially exclusive community
process. The exclusivity may not be intentional -  for example,
someone who doesn't speak English may be able to make sense of the
specs, but may have trouble with the nuances of discussion on this
list.

The Web works because it supports decentralised development and
document publication. More specifically, one of the reasons HTTP+HTML
work is because the use of URIs as names ensures clears partitioning
of the Web space.

Microformats support effectively the same kind of partitioning through
the use of profile URIs. Even if I were to ignore existing formats and
developed something that fulfilled the same purpose and reused the
same class names as  e.g. hCard, it needn't break anything. If I
included a profile URI that made clear it wasn't hCard, and didn't
include hCard, in practice that should be enough. To be 100% certain
it'd probably be advisable to also avoid any of the other microformats
that have interdependencies defined with hCard. Ok, a blatant
reinvention approach would be far from desirable, but I believe it's
important for it to be possible.

My own interests are mostly around Semantic Web tech, and those
technologies work because they support decentralised development and
data publication. That brings in the creation of the vocabularies, the
data schemas used to express the data. People can develop these
independently because the languages (URIs, XML+namespaces, RDF) are
designed to allow interoperability. The base level for the interop is
that the vocabularies are partitioned off from each other through
namespaces. Using GRDDL, microformats can be mapped directly to RDF
vocabularies and their independence assured by namespace partitioning
(and careful XSLT writing ;-)

I'm certainly not suggesting that community development is a bad
thing, rather that it shouldn't be (and isn't) necessary for
interoperable data.

Cheers,
Danny.

-- 

http://dannyayers.com