On Feb 12, 2006, at 4:08 PM, Scott Reynen wrote:
On Feb 12, 2006, at 5:21 PM, Ryan King wrote:
...
This method is something that Tantek and I came up with while
brainstorming about hResume with James Levine of Simply Hired. We
haven't gotten any feedback on it from others (except Brian Suda,
who says it will be workable for X2V to support this), so, please
examine it and give your feedback.
Here are a few potential problems I see:
1) it allows for a complete disconnect between what a human reads
and what a machine reads, as humans read by proximity, not pointers.
Indeed, but in this particular case the humans don't have a problem
figuring out the connection. So maybe, just maybe, its ok to be a bit
lax here. I'm not convinced that this is an ideal solution, but it's
probably the best one we have so far.
2) the resultant hCards contain out-of-date information, with a
person identified under a job title they no longer hold (except
possibly the most recent experience).
Right, except this data is within the context of an event, which has
dates associated with it.
3) it makes parsing more complicated. Currently, a parser needs to
consider only the fragment of the documents within the root of the
microformat, but allowing data to be scattered about the document
means a parser needs to have the entire document available during
parsing.
Indeed, it does complicate things a bit for parsing. However, we
already have this level of complexity with hCalendar in tables.
Backing up a step, what is the advantage of using the hCard class
on education and experience data? Just reusing components from
hCard without labeling them hCards would remove the need to follow
hCard's fn requirement, but I'm not clear on what would be lost in
the process.
What you're essentially asking for is a variant of hCard that doesn't
have the required-fn constraint. This is certainly an option for
hResume, we could reuse hCard semantics to describe job titles/
positions. However it would not be a generalizable solution.
I know we had talked about this, but we wanted to try and see if we
could get it to work this way first. I'm not totally sold on the
<object> reference method, but I think its the simplest thing we have
that could work.
-ryan
--
Ryan King
[EMAIL PROTECTED]
_______________________________________________
microformats-discuss mailing list
[email protected]
http://microformats.org/mailman/listinfo/microformats-discuss
