[uf-discuss] seeking clarification w/r/to hCard and RFCs 2425/2426
Brian Suda
brian.suda at gmail.com
Tue Mar 28 20:18:18 PST 2006
I think i am begining to understand what you are looking for, do you
have a solid example? that might clarify any issues myself (and
others) might be having.
Also, you mentioned that some of the services are "arbitrary" this
would make it vary difficult to standardise. One option MIGHT be to
look into the Rel-Directory[1] microformat. Similarly to tags it makes
the inference that this data should be in THIS category. This might
help to take very specific services and lump them into a directory
(which is dynamic) and let the community sort of drive that
classification.
I'm not sure if this is of interest, or relevance, but back in 2000
IBM wrote-up an idea called "The Advertisement and Discovery of
Services (ADS) protocol for Web services, Simplifying the announcement
of available Web services to inquiring software agents"[2]. It never
really went any where (to my knowledge). This was sort of the opposite
of UDDI, instead of centalising things, you advertised them
individually - like a robots.txt for services (early auto-discovery).
It was ment to be a directory of web services, but i imaging you might
be able to adapt/learn from some of it
-brian
[1] - http://microformats.org/wiki/rel-directory
[2] - http://www-128.ibm.com/developerworks/web/library/ws-ads.html
On 3/28/06, C. Hudley <chudley at gmail.com> wrote:
> On 3/27/06, brian suda <brian.suda at gmail.com> wrote:
> >
> > What are you trying to accomplish with directories that would be out of
> > the scope of hCard?
>
> Like I said: things like protocols and data formats. Version and
> revision numbers. Publication dates. Specification names and
> standards organization references. Arbitrary names of
> services-people-deliver that aren't standard in any way. Some of it
> could fit into hCard but other parts would be an awkward fit at best.
>
> This would be in the context of service discovery: for some resource,
> or for some organizational context, show me a directory of available
> or relevant services. If you're familiar with OpenURL resolvers, I'm
> thinking of something like their result screens, but bigger, and with
> a usable design and the option to machine-parse reliably. Which is to
> say, not really like current OpenURL resolver result screens.
>
> I don't really want to drop down to UDDI or whatnot... something
> user-centered is better. Though, I guess stuff like the UIs based on
> UDDI-driven services in apps like contemporary GIS tools with
> directories of remote data layers and remote feature processors might
> be the best examples I know of a current setting where something like
> what I'm after actually works and has a decent UI.
>
> I don't want to try to convince anyone here that there needs to be a
> microformat for this stuff... I'd just like to know if anybody's tried
> to do something like this already and whether they chose to "just use
> hCard" or something else or not at all. And, if it would definitely
> be wrong to "just use hCard".
>
> My hunch is that it would be wrong because most hCard processors
> likely expect to see information about people and organizations, not
> information about services and data and stuff. But my instinct is to
> try it anyway.
>
> Thanks.
> _______________________________________________
> microformats-discuss mailing list
> microformats-discuss at microformats.org
> http://microformats.org/mailman/listinfo/microformats-discuss
>
--
brian suda
http://suda.co.uk
More information about the microformats-discuss
mailing list