W3C home > Mailing lists > Public > uri@w3.org > June 2003

deprecating URNs and the contemporary view

From: Larry Masinter <LMM@acm.org>
Date: Sat, 7 Jun 2003 21:27:21 -0700
To: "'Hammond, Tony \(ELSLON\)'" <T.Hammond@elsevier.com>, "'Roy T. Fielding'" <fielding@apache.org>, <uri@w3.org>
Message-ID: <003801c32d76$472713c0$6ace8642@MASINTERPAD>

> Maybe Larry can tell us which one is closer to the contemporary view.

I have to confess that the idea of deprecating "URL" annoys me;
I'm not a big fan of making a 'old way/new way' shift.
However, I thought it was "rough consensus" to go ahead
with deprecating "URL", and it does seem to hold promise
to fend off some complaints about the terms being confusing.

Probably some _other_ people will complain _even more_
that we've confused them _worse_, though; I'm not sure this
is a battle we can win.

With that said, I'll try to justify deprecating "URL"
and not "URN":

I don't think 'balance in the glossary of terms' is a goal.
Keep terms which have or can be given clear definitions,
and deprecate those that can't. Defining "URN" to mean
"a URI whose scheme is 'urn'" is pretty clear.

> Given that a URI scheme may be classified as a 'locator', a 'name' or
both,
> how can the term 'URL' be deprecated while maintaining currency of the
> term 'URN'?

I'm not in favor of attempting this classification, either
individually or by scheme. One person's name is another
person's locator. I think that's the main intent of
the "contemporary" revisionism: the classification doesn't
work.

> The 'urn'
> scheme just marks out a certain class of URIs which have a particular
> semantics - i.e. 'persistence'. Nothing more.

Hardly. There are URIs with 'persistence' (whichever definition
you might choose) but which don't start with 'urn:', and there
are URIs that start with 'urn:' that don't have any guarantee
of 'persistence' (for most reasonable definitions of same.)
So I don't think persistence it's definitional. It's a hope
and a promise; we all hope that URNs are persistent, and we
ask those who hand them out to make the recipients promise
to make it so. We also provide no mechanism anyone can
count on for updating what is identified, thwarting
non-persistence.

Leaving the term "URN" bound to "URIs that start with 'urn:'"
means we don't have to ask people to call those things
'urn URIs'. It's just a little bit of linguistic niceness.

The complement ("all URIs that don't start with 'urn:'")
isn't as useful. Personally, I would have no trouble with
redefining the term "URL" as "a URI used as a resource locator,
such as inside the 'href' attribute of an 'a' element in HTML",
but it would lead to some additional confusions if you ever
did see a 'urn:' being used as a locator, so we might as
well let it go.

We can then all be proud that we're modern... or contemporary...

Larry
-- 
http://larry.masinter.net
Received on Sunday, 8 June 2003 00:27:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 13 January 2011 12:15:31 GMT