RE: comments on draft-eastlake-cturi-01.txt

Persistent is fundamentally due to people, and technology can assist but not
guarantee. 

I very much agree with the point made here that there are legitimate,
desirable, and unavoidable reasons for changing organisation names, domains
etc.; so let's deal with it.  One aim of naming entities or resources is to
avoid tying an entity  name to a domain name (or any other piece of variable
metadata); compare the domain names/trademarks issue.  The entity can be
persistently named as a first class object irrespective of its location,
owner, licensee, etc.  Distinguishing names from locations is essential for
e-commerce:  it is trivially true that "all names are locations" (in a
namespace) - but practically, most people worry about spaces like URLs and
that's the wrong level.  Naming entities as first class objects, rather than
locations, enables better management of e.g. multiple instances of an
object.

In the DOI (Digital Object Identifier) initiative (http://www.doi.org) we
have concluded that technology can help with persistence: e.g. a redirection
resolution step enables management in the redirection directory, rather than
in bookmarks and the like, thereby ensuring that one change can be picked up
by many users even if they are unaware of the change.  But to manage the
data in the directory takes effort, time, incentive, etc. - either you do
that locally (purl like) or as a global service (DOI).  In our case we see
that as ultimately a role for registration agencies - we can learn from
other proven activities like bar codes, ISBNs, and other data systems.  And
people aren't free, so there's a cost to this (just like the physical bar
code system , the DOI aims to be a self-funding operation).  DOIs won't be
appropriate for many things, and some people won't feel this  people cost
merits the reward.  But we do think DOIs are a viable solution for content
management of intellectual property on a large scale.


-----Original Message-----
From: Donald E. Eastlake 3rd [mailto:dee3@torque.pothole.com]
Sent: 29 January 2001 03:15
To: Weibel,Stu
Cc: 'michaelm@netsol.com'; Aaron Swartz; Larry Masinter; uri@w3.org;
ietf-type@iana.org; dee3@torque.pothole.com; Graham Klyne; Ted Hardie
Subject: Re: comments on draft-eastlake-cturi-01.txt 



From:  "Weibel,Stu" <weibel@oclc.org>
Message-ID:  <A4DCD9B43F237A41BE31C98D434C22880B5EE9@OA4-SERVER.oa.oclc.org>
To:  "'michaelm@netsol.com'" <michaelm@netsol.com>,
            Aaron Swartz
    	 <aswartz@swartzfam.com>
Cc:  Larry Masinter <masinter@adobe.com>,
            "Donald E. Eastlake 3rd"
    	 <dee3@torque.pothole.com>, uri@w3.org,
            Graham Klyne <GK@ninebynine.org>, Ted Hardie
<hardie@equinix.com>
Date:  Sun, 21 Jan 2001 09:26:57 -0500

>URLs that are responsibly constructed and well managed will be as
persistent
>as the committment of the organizations that manage them.

That isn't true, as an orgnaization's DNS names can get taken away
from them due to conflicts of which they were not originally aware.
But even if it were true, many organizations change name and/or domain
name for marketing reasons, because new management wants to make a
change, because old management wants to appear to be making a change,
etc.

It makes no sense to me to have the stability of MIME Type /
Content-Type <-> URI mappings to depend on the name stability and duty
assignment stability of IANA or ICANN or the IETF or W3C.  This
mapping will get embedded into widely deployed code, essentially
impossible to get globally updated.

Donald

>The vulnerabilities discussed in these messages simply don't pertain to
well
>managed URLs (that is, managed according to publically stated policies and
>with reasonable attention to IPR issues).
>
>stu
>
>-----Original Message-----
>From: Michael Mealling [mailto:michael@bailey.dscga.com]
>Sent: Sunday, January 21, 2001 3:26 AM
>To: Aaron Swartz
>Cc: Larry Masinter; Donald E. Eastlake 3rd; uri@w3.org; Graham Klyne;
>Michael Mealling; Ted Hardie
>Subject: Re: comments on draft-eastlake-cturi-01.txt
>
>On Sun, Jan 21, 2001 at 12:21:26AM -0600, Aaron Swartz wrote:
>> Larry Masinter <masinter@Adobe.COM> wrote:
>> > That you, Aaron Swartz, do not see the need to use anything
>> > other than "http://www.iana.org", which has sufficient
>> > stability for your own purposes, doesn't mean that it will
>> > meet the needs of everyone else.
>> > 
>> > I suppose this argument will persist until we resolve
>> > the W3C/IETF split over the utility of URNs and their
>> > role in protocol element identification.
>> 
>> I hate to see this argument pointlessly persist, so I will stop arguing
>> after this question:
>> 
>> Why aren't my URLs safe? That is, why do I have to worry about an address
>at
>> iana.org suddenly disappearing one day? What needs are not met by this
>> system?
>
>>From my standpoint there are two reasons:
>
>One of the main reasons is that due to existing case law you 
>don't own your domain-name (the same way you don't own your telephone
>number). If a court says so a registry is required to remove that
>domain-name from service and either not give it back out to anyone or
>sell it to someone (probably a competitor).
>
>Reason two: since there is nothing inherent to domain-names or
>http or anything else, the only way I know I can use your
>URLs anytime beyond tomorrow afternoon is that you have told me so.
>That may be fine if I interact with you on a daily basis but
>if I come across some URI 'in the wild' I have no idea how
>persistent it may be and if I do act as though it were useable
>beyond tomrrow afternoon then _I_ am the one making an error in
>assumptions. Now, if the URI scheme requires that it be persistent
>then I can start doing some pretty powerful things since I can
>now make that assumption safely. If the URI I find 'in the wild'
>is part of that scheme and it doesn't follow those rules then
>I know that it is an error in the network, not in my making
>erroneous assumptions....
>
>
>> I use URLs for a lot of the work I do and I'm curious whether I'm making
a
>> mistake.
>
>Probably not for most cases. But if you plan on using a URI 10 or 20
>years from now it just might be a problem. Heck, there was a large amount
>of discussion at the last IETF about creating a new DNS class
>and changing the rules _completely_ (in a new class you don't
>have to follow any of the old rules, including delegation models).
>
>-MM
>
>---------------------------------------------------------------------------
-
>Michael Mealling	|      Vote Libertarian!       |
>www.rwhois.net/michael
>Sr. Research Engineer   |   www.ga.lp.org/gwinnett     | ICQ#:
>14198821
>Network Solutions	|          www.lp.org          |
>michaelm@netsol.com

Received on Thursday, 1 February 2001 07:03:57 UTC