Re: Draft: Universal Document Identifiers (Peter Deutsch)
Message-id: <>
In-Reply-To: Larry Masinter's message as of Mar 11, 10:15
From: (Peter Deutsch)
Date: Wed, 11 Mar 92 18:28:36 GMT-0:02
In-Reply-To: Larry Masinter's message as of Mar 11, 10:15
X-Mailer: Mail User's Shell (6.5.6 6/30/89)
To: Larry Masinter <>,
Subject: Re: Draft: Universal Document Identifiers
Cc: cni-arch@uccvma.bitnet,,,
> From Wed Mar 11 13:18:12 1992
.  . .
> Um, I think when I tell you about a document, I can tell you:
> a)Some attributes about it that you can remember and use for
>   finding it again.
> b)its signature/fingerprint/checksum whatever
>   This helps you know whether you already have exactly what I'm
>   referring to or can get it more locally.
> c)some information about where I think you can get it and how
> d)a set of instructions you can use for getting it.
.  .  .
[* analogy deleted *]
> The last two are a little strained in the analogy; don't jump on the
> analogy, please, I just want to point out that it is reasonable and
> customary to supply *MORE THAN ONE* of unique identifier, serial
> number, access path, common attributes, etc.

I have no trouble with this. Obviously, I'm not sure I
need all of these for each query, since I'm trying to
accomplish different things at different times so what I
want back will change but that's just detail.

I suspect that only a) and b) are really permanently
associated with the information and the other two may or
may not be provided by external services. Of course, if
someone's favorite information delivery service wanted to
provide all four, (either directly or through gateways)
that's fine by me.

				- peterd