[Geowanking] geo microformat BOF session at Where 2.0

[Mattias Konradsson]

> Man, I just want to say for the record that microformats look 
> to me like a big mistake waiting to happen. Everyone probably 
> thinks that RSS is the shiznit on toast, but I gotta say, 
> anyone who's had to write an RSS aggregator and deal with the 
> 8 or 9 different flavors of RSS out there will tell you that 
> "microformats" are a big mistake waiting to happen. 
> Application-specific XML (micro)formats are, like,
> *so* 2001.
> 
> Come on, kids, it's the 21st Century already, RDF isn't 
> rocket science, it already offers bottom-up style pluggable 
> ontologies (i.e.
> namespaces) and you'll be glad you went there when you need 
> to add new semantics to your application without having to 
> *touch* a (micro)format parser ever again.  Gee whiz.
> 
> SDE

Well personally I disagree because I find rdf to be a very "messy" (not that
basic rss is much better) format, while a good implementation-specifik xml
format can be simplistic and easy to read and parse. However I don't believe
"rdf vs microformats" is the issue, both are xml-based formats and it's
fairly trivial to take them and translate to each other or some new format
using xslt.

 No, the issue is the design of the actual data, staying true to a defined
schema and keeping things wellformed

1) design of the actual data
You need to think really hard about what information your format should
describe,in what detail you should describe it and what implementations it
need to support. Take rss for instance. Structure doesnt matter much but
neglecting things like guids in the original spec made it unecessarily
difficult for aggregators to keep track of items, not defining a utc format
for date/time  was another big mistake and they didnt address the issue of
embedded html. You really need to think ahead of all possible uses when
designing a format, of course you can't predict *everything*. Stuff like
podcasting was probably very hard to design for. On the other hand if you
design for everything you're left with a bloated format so guess the rule is
to atleast think hard about the basics.

2) Using schemas
Half of the rss providers don't even follow the basic schema of the rss
version they're using. This really is the story of html all over again, if
you don't force authors to validate their content you'll end up with lots of
ugly hacks to fix broken formats. Unfortunately it's a bit hard to enforce
schemas since adoption of technology tends to favor loose standard but
still.

3) Keeping things wellformed
Well, part of 2) really but still. You'd think people authoring xml-content
atleast could make sure it's wellformed so it can be parsed without a lot of
second-guessing and trying to fix stupid errors, but nooooo :)

So I hope people designing formats in the future will think about these
issues a little harder so we don't have to keep repeating the same mistakes
over and over again.

mvh
--
Mattias Konradsson

"GlobalVillage - world domination or public humiliation"
http://www.globalvillage.nu/mattias