The main problem, as I understood it, is that "object[data]" expects
a URI, even if it doesn't know how to handle it, so the first
suggestion is actually requesting the relative path "./20050125"
which causes extra junk 404s (Ex. 1; not necessarily a bug). Some UAs
even requested relative paths for anchored resources in the page as
with the object include-pattern (Ex. 2; probably a bug and definitely
a reason to ditch it).
1. <object class="dtstart" data="20050125">January 25</object>
2. <object class="include" data="#foo"></object>
Problem noted here: http://microformats.org/wiki/include-pattern-
feedback
The next problem was browser display inconsistency with the human-
readable text being the innerHTML of the object.
The object example listed in the WaSP post circumvented both of these
problems, but wasn't very elegant markup and even looked sloppy
without the accompanying CSS. The solution was basically to ditch
"object[data]" and use "object param[value]" instead. The inelegant–
but working–object version was:
<span class="dtstart">
January 25
<span><object><param name="value" value="20050125" /></object></span>
</span>
There may be an additional problem of performance–what happens when
you load up 300 empty objects on a page even if they aren't trying to
reload the page 300 times–but it's as yet undocumented. I would have
spent more time finding out if the solution had been more elegant. As
is, I wasn't seriously suggesting it, but I wanted to leave the
possibility in there for consideration. This was as much homework as
I deemed necessary to commit.
James
On Apr 30, 2007, at 10:27 AM, Ryan Cannon wrote:
Perhaps I'm getting into this a bit late and this has already been
brought up, but I've skimmed through the conversation and haven't
seen it. Tantek's original proposal[1] was scrapped because it
didn't work in Safari 1.2.1 (WebKit v125). Hasn't that particular
browser version been obsoleted to that point that we can reconsider
using it? The latest Safari version for OS X.3 is 1.3.9, which is
soon to be two OS versions back. Any idea precisely when this bug
was fixed?
While few browser stats break Safari versions down to the WebKit
version, my site has received 1 hit from from WebKit v125, and that
tiny marketshare is reflected in other stats I've found[2]. If we
are going to talk about < 1% browsers, why are we holding back an
otherwise ideal design pattern for an obsoleted version of a minor
browser?
<object> is ideal, as Tantek described it, and it is both simple to
write and backwards-compatible.
[1]: http://tantek.com/log/2005/01.html#d26t0100
[2]: http://www.webreference.com/stats/browser.html
--
Ryan Cannon
Interactive Developer
http://RyanCannon.com
_______________________________________________
microformats-discuss mailing list
microformats-discuss@microformats.org
http://microformats.org/mailman/listinfo/microformats-discuss
_______________________________________________
microformats-discuss mailing list
microformats-discuss@microformats.org
http://microformats.org/mailman/listinfo/microformats-discuss
