[uf-discuss] uF Discovery? (was: FYI: Jeff Jarvis on microformats and Google Base)

[Brian Suda]

With hCards there has been a discussion about "Auto-Discovery"[1].

If the CURRENT page has an hCard in it, then rel="alternative" would
be correct, because the vCard is an alternative representation of the
hCard.

But i think you are talking about "the data over there contains ..." a
page containing a PROFILE link does not always mean it will have a
microformat within it. So detecting profile (that page over there has
these profiles) is not sufficient to determine if it has any of the
microformats.

There are a few ideas listed in the rel="" to auto-detect an hCard,
the same train of thought could be expanded for all microformats.

Just remember that what i point to and see is not always the same as
what you see. (that has always been the debate over adding the
HREFLANG[2] to hrefs, if you say this is in ENGLISH, but because i
send different HTTP headers i get the SPANISH page instead, then your
assertion that it is english is incorrect.[3])

If you look at any other "distributed" web services, for instance
SOAP, it has WSDL for description and UDDI (and other competing
services) for discovery. ADS (Auto Discovery Service)[4] was an idea
at IBM to help populate UDDI services. The user would create a file
(my old notes say svcadvt.xml althought google knows nothing about
this) that would act similar to Robots.txt, a spider would come along
looking for that file, read the XML, parse it and file it into the
UDDI server, therefore keeping your record "fresh". This does suffer
from Data Drift between services offered and services advertised! (i
could easily see microformats data drifting). The advantage to
crawling rather than submission was that other data could be extracted
from the HTML page, like scanning the page for TAGS, thus using Human
visible data to help categorize the Service correctly.

Those are just a few thoughts on Discovery of microformats.

-brian

[1] - http://microformats.org/wiki/hcard-brainstorming#Auto-Discovery_for_XFN
[2] - http://www.w3.org/TR/REC-html40/struct/links.html#adef-hreflang
[3] - http://annevankesteren.nl/2004/06/hreflang-and-type
[4] - http://www-128.ibm.com/developerworks/web/library/ws-ads.html

On 11/23/05, Luke Arno <luke.arno at gmail.com> wrote:
> On 11/23/05, Ryan King <ryan at technorati.com> wrote:
> > On Nov 22, 2005, at 9:11 PM, Luke Arno wrote:
> >
> > > I am sure this has come up before, so sorry in
> > > advance, but could we use rel values to say
> > > "there are x and y uFs over there".
> >
> > In most cases, this would be semantic abuse. Remember, rel and rev
> > describe the *relationship* between the current page and the referred
> > URL. Type could get you a big closer, but still wouldn't be very
> > reliable.
> >
>
> Right, of course. There is no particular relationship
> here. Maybe just rev="knowsabout" :-)
>
> So class names then?
>
> <a
>    href="http://lukearno.com/#hCard"
>    title="Luke's hCard"
>    class="disco-hCard">Luke</a>
>
> class ideas:
>
> disco-disco
> disco-uF
> disco-hCard
> disco-XFN
> ...
>
> Examples would be feed autodiscovery and
> just about any use of links really.
>
> - Luke
>
> > > In other words, should there be a microformats
> > > discovery microformat? uF-links or some such?
> >
> > -ryan
> > --
> > Ryan King
> > ryan at technorati.com
> >
> >
> >
> > _______________________________________________
> > microformats-discuss mailing list
> > microformats-discuss at microformats.org
> > http://microformats.org/mailman/listinfo/microformats-discuss
> >
> _______________________________________________
> microformats-discuss mailing list
> microformats-discuss at microformats.org
> http://microformats.org/mailman/listinfo/microformats-discuss
>


--
brian suda
http://suda.co.uk