- From: Sandro Hawke <sandro@w3.org>
- Date: Fri, 27 Apr 2001 20:44:13 -0400
- To: "Sean B. Palmer" <sean@mysterylights.com>
- cc: michaelm@netsol.com, "Tim Kindberg" <timothy@hpl.hp.com>, uri@w3.org
> tag://hpl.hp.com;date=1/tst.1234567890
> tag://exploratorium.edu;date=1/pi.99
> tag://sandro@w3c.org;date=1/my-dog
> tag://myIDs.com;date=1/TimKindberg/doc.101
> tag://champignon.net;date=1/
> tag://champignon.net;date=1-3-22/99
> tag://champignon.net;date=2-4/100
I don't see what the "//" or "date=" gets you. The "//" in URIs lets
you do relative links within the same scheme, but that's meaningless
with tags, since there are no defined semantics (across all
authorities) for linking. (I've never actually seen the // be used, but I
guess one could conceivably use it to stay with http vs. https in some
set of host-crossing documents.) The "date=" lets people who know nothing
about the scheme know that the following bit is some kind of a date,
but it doesn't tell them nearly enough to make sense of "date=1".
I also think its important that the date be visually grouped with the
name more than the following text, which is a little tricky. Maybe
there's not much difference among the following:
tag:sandro@w3.org/1:foo
tag:sandro@w3.org,1/foo
tag:sandro@w3.org,1:foo
tag:sandro@w3.org;1/foo
I'll need to think more about the argument that "/" is misused in the
first version.
-- sandro
Received on Friday, 27 April 2001 20:46:40 UTC