Re: 2.3 URI Ambiguity

Good comments, Walden.

> The example of natural language ambiguity could also
> be tightened up by removing the URI altogether.  It would
> be better to just say "Moby Dick" is ambiguous because
> it could mean: A, B, C...  Also, I think too many examples
> are included.  Three would be more than enough.

Agreed.

> 2.3.1 URI in Other Roles is fallacious.  It says URI are
> sometimes used other than to identify, but then it goes on
> to give examples of URI identifying companies, websites,
> people, mailboxes.  No examples are given in which a URI
> is not used to identify something.

Agreed.

> Bottom line:  you're saying "avoid using a URI to identify
> more than one resource", and also "it's okay when a URI
> sometimes identifies more than one resource" (e.g., you and
> your mailbox, both).  Is there a critical distinction that
> can be made so I can see why this is bad in one case and
> good in another?

I personally don't see a distinction.

> Is there a counterexample of the "avoid URI ambiguity"
> practice that would help make that clear?
> 
> Is there enough critical mass in the concept of "indirect
> identification" to warrant a definition of same?

The previous text went further into that, but I think it's good that
this text did not.  IMO, indirect identification is ambiguous.
Moreover, I don't know any other way to be ambiguous than to use
indirect identification.  Maybe that's just me though; I'm happy to be
shown to be incorrect, but 2.3 doesn't do that, AFAICT.

I also wanted to point out that my previous concern[1], which Dan
suggested[2] that the revised text may address, still applies to this
text.  Specifically, it seems to take a position on httpRange-14 (the
wrong one in fact, otherwise I wouldn't have mentioned it 8-).

As for how to move forward, I'd personally just like to see the existing
good practice note combined with a single example describing what
ambiguity is; the database merging one seems fine (minus the preamble).

Thanks.

 [1] http://lists.w3.org/Archives/Public/www-tag/2003Nov/0046.html
 [2] http://lists.w3.org/Archives/Public/www-tag/2003Nov/0047.html

Mark.
-- 
Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca

Received on Monday, 1 December 2003 09:54:44 UTC