Martin McEvoy wrote:
'-' + vendor identifier + '-' + meaningful name
http://www.w3.org/TR/CSS21/syndata.html#vendor-keywords
which leads me to believe that publishers can do something like this...
<span class="bday" style="-uf-bday:1968-01-04;">4th Jan, 1968</span>
but I believe that
this is a lesser evil than stuffing the values into @title, a machine
wont choke on the data it will just ignore it, where as at the moment
people do choke on the data because it makes little sense to them, its
in the wrong place.
1. Sticking a hyphen at the beginning of the property name is not all
there is to it. The CSS 2.1 specification only defines error-handling
for (unrecognized) vendor-specific properties only work when their names
and values can be parsed with the CSS core grammar. That might well be
possible with "1968-01-04"; I haven't tried to evaluate it. Alternately,
you might need to put it in a quoted string (like with the value of the
"content" property).
2. Just because the parsing of vendor-specific properties that do match
the grammar is defined does not make them conforming. I don't think
requiring microformat users to produce non-conforming, invalid CSS is
compatible with microformats' selling point of building "upon existing
and widely adopted standards". I don't think there's _any_ fundamental
difference in terms of conformance between sticking this data into a
vendor-specific CSS property and into a HTML custom attribute, and the
later would be a _lot_ less hacky.
3. The hidden data problem is very hard to solve in HTML 4.01 but I
think we already have more conforming "hacks" than this (data in class
name, empty span with title, object data, and hidden input among them).
--
Benjamin Hawkes-Lewis
_______________________________________________
microformats-discuss mailing list
microformats-discuss@microformats.org
http://microformats.org/mailman/listinfo/microformats-discuss
