/ Graham <dtcd@xxxxxxx> was heard to say:
|> With URIs, not neccessarily, but that's a different matter.
|
| Well OK. What's wrong with it being the same address made out of
| different characters?
There's no easy, reliable way for software to tell that it's the same.
Consider "http://norman.walsh.name/atom/whatsnew.xml"; on the one hand and
(1) "http://norman.walsh.name/atom/whatsnew.xml"; on the other. They're the same.
There isn't going to be any disagreement about that.
(2) "http://Norman.Walsh.name/atom/whatsnew.xml"; on the other. In order to know
that they're the same, you have to understand the spec for the 'http' scheme,
parse the domain component, normalize it, etc. Some folks are going to do
strncmp() and some aren't so there will be disagreement.
(3) "http://norman.walsh.name/atom/wh%61tsnew.xml"; on the other. This is probably
the same, but that depends on which character(s) are percent escaped. And
on the scheme. We don't want to go there.
(4) "http://norman.walsh.name/atom%2fwhatsnew.xml"; on the other. I believe that
this is really different, but I might be wrong.
And I haven't even opened the I18N can of worms yet.
If we're going to have semantics that depend on comparing URIs, I do
want the spec to say explicitly that comparison is by comparing the
characters exactly as they are writ.
Be seeing you,
norm
--
Norman Walsh <ndw@xxxxxxxxxx> | The trip doesn't exist that can set you
http://nwalsh.com/ | beyond the reach of cravings, fits of
| temper, or fears. If it did, the human
| race would be off there in a body.--
| Seneca
Attachment:
pgp00153.pgp
Description: PGP signature