Internet Info 1997 December/Internet_Info_CD-ROM_Walnut_Creek_December_1997.iso/ietf/urn/urn-archives/urn-ietf.archive.9611/000074_owner-urn-ietf _Wed Nov 6 19:37:15 1996.msg

Internet Message Format | 1997-02-19 | 7KB

Received: (from daemon@localhost) by services.bunyip.com (8.6.10/8.6.9) id TAA20472 for urn-ietf-out; Wed, 6 Nov 1996 19:37:15 -0500
Received: from mocha.bunyip.com (mocha.Bunyip.Com [192.197.208.1]) by services.bunyip.com (8.6.10/8.6.9) with SMTP id TAA20465 for <[e-mail]>; Wed, 6 Nov 1996 19:37:13 -0500
Received: from loki.fsc.fujitsu.com by mocha.bunyip.com with SMTP (5.65a/IDA-1.4.2b/CC-Guru-2b)
        id AA15698  (mail destined for [e-mail]); Wed, 6 Nov 96 19:37:02 -0500
Received: from ishtar.fsc.fujitsu.com (ishtar [192.240.3.45]) by loki.fsc.fujitsu.com (8.7.5/8.6.11) with ESMTP id QAA29690 for <[e-mail]>; Wed, 6 Nov 1996 16:35:19 -0800 (PST)
Received: (from tallen@localhost) by ishtar.fsc.fujitsu.com (8.7.5/8.6.11) id QAA22323 for [e-mail]; Wed, 6 Nov 1996 16:35:41 -0800 (PST)
Date: Wed, 6 Nov 1996 16:35:41 -0800 (PST)
From: Terry Allen <[e-mail]>
Message-Id: <[e-mail]>
To: [e-mail]
Subject: [URN] Re meaningful names
Sender: [e-mail]
Precedence: bulk
Reply-To: Terry Allen <[e-mail]>
Errors-To: [e-mail]

[subject line changed to decouple Unicode from meaningful naming]

I wrote:
|    Our longest-lived names actually are meaningful to humans; to support
|    your argument you have to establish that semantic drift affects
|    names specifically (Moby Dick, Julius Caesar, Assurnasirpal) and
|    that using them in the context of a defined name space does not
     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
|    insulate them from semantic drift.

Lewis Girod replied:
| The names you are using as examples are all long lived because they
| refer to conceptually immutable objects.  This is the case in which
| names are afforded some protection from semantic drift, although it
| does not necessarily protect them completely.  Out of context, I can't
| tell if you mean Julius Caesar the guy who got stabbed or Julius
| Caesar the play (and that was not an issue when JC was around!,) but
| in the context of a namespace of people or plays that problem would be
| resolved.  Either way the name refers to a piece of history that,
| whether or not it is well defined, people have an idea of what you
| mean that doesn't change too much.

Exactly.  urn:roman.emperors:julius.caesar is not ambiguous.

| However, I think what Keith was getting at was a problem that arises
| when semantics is embedded in names in an attept to structure
| hierarchical namespaces.  The sort of semantic mechanism involved with
| categorizing things is much more likely to drift over time; a
| collection of companies originally categorized under ``Mainframes''
| might want to recategorize under ``Enterprise systems''.  (Over longer
| spans of time spellings and meanings of words change -- Julius Caesar
| was originally filed under playes perhaps.)  In the context of URNs,
| every move we make we are stuck with forever... which eventually leads
| to silly results and overall cruftiness.  There is little point in
| having semantically significant names unless you retain semantics that
| makes sense as time goes on.

Well, no.  The name is the name; if you come to dislike it or think
it doesn't mean the right thing you can map something else onto it
in a new name space.  If an existing name space uses "Mainframes",
so be it; it will still work just fine as part of an unambiguous
identifier (even if users have to be told "what name space FOO
calls Mainframes we now call Enterprise Systems").

| Of course, as you point out there is no way to stop people from
| putting semantics into names; the most effective way to deal with the
| issue is to implement a system of user friendly names over the top of
| URNs; 99% of the time the user would see and type only a UFN, and
| these UFNs would be resolved to the URNs that could then be used as
| long-lived references.

Twice the work or more, doesn't deal with grandfathering, prone
to error; the most effective way to deal with the issue is to
drop your objection to meaningful names.

Keith:
| > Few people on this planet can accurately transcribe anything.
| That's not true.  People regularly transcribe telephone numbers,
| email addresses, ISBNs, URLs, credit card numbers, etc. with
| sufficient accuracy for their purposes.  But ask people to

Actually, they regularly mistranscribe them.  In the past two
weeks I have mistranscribed my VISA number and the VIN of
my newly acquired used car, precisely because they are inherently
meaningless.

Martin, I think it was, pointed out the value of meaning for
correct transcriptions.  I think he's right.  But I also think
it's an issue this group should not legislate on:  tell inventors
of new name spaces what you like in the way of good practice,
but don't prescribe.

| > Your
| > last point would have merit were it not that we absolutely must
| > be able to grandfather in existing name spaces.
|
| Yes, we must be able to grandfather existing name spaces *that serve
| similar purposes to URNs*.  ISBNs qualify.  Handles qualify.
| Usenet message-ids qualify.  But URNs don't have to handle existing
| name spaces that don't have the characteristics that URNs have.

This is circular argument unless it's grounding in the group's
charter.  What I find in the charter as Leslie posted it is:

"Although the framework will allow URNs to be defined
that vary in terms of degree of scalability and persistance, ensuring
"user friendliness" of all resultant identifiers is beyond the scope of
this group."

That's far from saying that URNs shall not be meaningful, legible,
or user-friendly.  And RFC 1737 (section 5) explicitly takes no
position on use of meaningful semantics.

I don't find grounding for your argument.

| I don't think this is a problem, because I don't think existing URN-like
| name spaces are likely to be human friendly in the first place.  (If someone
| knows of a counterexample, please tell us.)

Circular reasoning again.  "Human friendliness" is not a defined
property of URNs at the moment.

Lewis:
| One idea we were experimenting with here was to use a subset of ASCII
| that included only numerals, punctuation and consonants, which would
| make most words hard to express (at least in some languages).

I cannot imagine that such an approach would find favor in the real
world.



Regards,
    Terry Allen    Fujitsu Software Corp.    [e-mail]
"In going on with these experiments, how many pretty systems do we build,
 which we soon find outselves obliged to destroy?" - Benjamin Franklin
  A Davenport Group Sponsor:  http://www.ora.com/davenport/index.html