[microformats-discuss] Late to the party!

brian suda brian.suda at gmail.com
Thu Jul 14 15:29:49 PDT 2005

I was just added to the discuss list (i thought i signed-up, but that
must have been the dev list?). I read through the archive as to not
repeat anything and to try to bring myself up to speed with everyones
questions, comments, and thoughts. My X2V came-up several times, so i'll
try to explain the state of it and other Microformat related projects in
the hopper.

X2V is a gigantic XSLT file (actually two, one for hCard, one for hCal -
in hind-sight this was a pretty bad name for the software. It originally
was X for XHTML and XSLT, the 2 was to represent the two x's the two v's
and the work 'to' and the v's are for vCard and vCal. As we know now it
is actually iCal, and also works with bad HTML via Tidy to make it
XHTML. So all-in-all the name is less catchy, but i keep it so not to
confuse everyone using it). The reasons for an XSLT file are multiple:
- it is cross-platform and cross-language, anything that has a good XML
lib could use these files
- it tries to mimic GRDDL
- it can be used in other webservices
- it is RESTful, so it is like a UNIX pipe for the web
- it taught me alot more about XSLT

The XPath expressions have become more and more convoluted with each
iteration of the version. This is due to LOTS of things, one being my
limited use of XSLT (it probably could be optimised), and PHP's
implementation. I tested many XPath expressions offline in Firefox
locally, then ran the same code through the PHP XML lib and it failed,
so some expressions are used because they work in PHP. Others aspects
deal with the nesting of elements, our change over from
'first-letter-capital' to 'all-lower-case' class values. Then also,
class elements can be seperated by spaces, tabs, and/or returns, so
those had to be normalized to spaces then split on spaces. It is a long
process that is never finished. (hence the BETA)

I have also been working on a generic XMDP parser, so you could validate
any page. That page's profile would be extracted, the URIs fetched and
then compared against the XMDPs and some sort of output would be shown.
There were SEVERAL problems (if anyone is interested i can explain
further), mostly dealing with collisions of class values (e.g. url in
vCard and url in iCal) and deciphering if a class value was part of the
CSS style or for a microformat?

i'm certainly on this mailing list now, so i can respond directly to any
Questions about implementations, or is it Consuming?


More information about the microformats-discuss mailing list