[uf-discuss] Disambiguation Conventions? (was Comments
fromIBM/Lotus rep about Microformats)
joe at andrieu.net
Tue Dec 12 11:28:37 PST 2006
> Mike Schinkel wrote:
> Ryan Cannon wrote:
> > If the community is slow to develop a format that makes
> > sense, we often encourage authors to develop their own
> > systems, which then can inform how a format will function in
> > the wild. This is where documentation and the oft-belabored
> > "process" becomes powerful. Although it can be annoying for
> > early-adopters and people who need solutions now, it creates
> > strong formats once the issues are solidified.
> Arrgghhh! (he says, frustrated that people address the
> tangent to his issue, but don't/won't address his actual issue!)
> So I repeat: How then can we achieve a disambiguation
> conventions to keep official Microformats from conflicting
> with "proto- Microformats?"
Microformats has no mechanism in place to disambiguate, either amongst
its own uFs or between its own and other "non-official" microformats.
> A "microformat" is such because it is a use of semantic class
> names, etc. that IN PARTICULAR:
> 1. Are designed according to microformat principles 
> 2. Follow the microformats process 
>  http://microformats.org/wiki/process
Although the process says to use the microformats wiki and mailing list,
there is no "official" blessing as both of these media are either
unmoderated or "community" moderated. Any user could make
recommendations and edits. A team of antagonistic users could hijack it
completely. There really is never an "official" version of any given
microformat, although there have been statements that eventually their
will be when shepharded through a standards body. The only real
authority is the autocratic role played by Ryan and Tantek.
For example, since it was initially stabilized hCard has been changed to
include "place" in its semantics, yet we have no way to let parsers know
that the "new" hCards may not be people, companies, or organizations,
but instead may also be places. vCards were for /contactable/ entities.
hCards changed the semantics to include "locatable" things because the
address capability of existing hCards made that convenient. Yet, there
is no versioning to disambiguate between these two. That means that
once a standards body blesses any particular microformat, it will be
locked in stone. I for one think hPlace (or something) should be a
separate uF, based on the location-related fields in hCard.
Similarly, there is no mechanism to distinguish class names in one hcard
from names in another, so we are forced to make sure that the entire
namespace is flat and hopefully unique enough from random semantic HTML
in the wild.
If it looks like a duck and quacks like a duck, you really can't tell if
it is a duck, but the parsers will try to treat it like a duck anyhow,
even when it is simply semantic HTML that happens to have the same class
names as a microformat.
The initial concept, as I understand it, was that divergent,
namespace-enabled variants are bad. That instead, a community-forged
single namespace based on consensus is better. I agree with the intent
of the latter. It is good to have a shared tongue. However, I think
that goal is compatible with the existence of namespaces (or other
disambiguation). Just as having an international language of commerce
and science is good (English serves this purpose today), it is also
entirely compatible with a world where people are allowed to speak other
languages, especially when knowing which language they are using helps
you clarify their meaning.
I, for one, think we need a mechanism to disambiguate, but I didn't have
much success with my early arguments. Apparently I hit a lot of
hot-buttons and got shouted down (some of my suggestions were somewhat
tangential to this particular point and probably deserved getting
shouted down). The idea of disambiguatable namespaces stirs up
exhortations against a "Babel"-like profusion of uFs, which apparently
would be the end of the world and somehow inherently prevent the forging
of the consensus uF namespace. I don't buy that argument, but it is the
one that has held sway here since the creation of microformats.
That's ok. Microformats are still cool. I just don't think they scale
well, which is apparently by design.
joe at andrieu.net
+1 (805) 705-8651
More information about the microformats-discuss