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

Hi Larry,

I agree with you, at least in this message  :-) ...

From:  "Larry Masinter" <masinter@Adobe.COM>
To:  <connolly@w3.org>
Cc:  "Aaron Swartz" <aswartz@swartzfam.com>, <uri@w3.org>,
            <Donald.Eastlake@motorola.com>,
            "Donald Eastlake 3rd" <dee3@torque.pothole.com>,
            "Graham Klyne" <GK@ninebynine.org>,
            "Michael Mealling" <michaelm@netsol.com>,
            "Ted Hardie" <hardie@equinix.com>
Date:  Thu, 18 Jan 2001 19:37:21 -0800
Message-ID:  <NDBBKEBDLFENBJCGFOIJMEEAEEAA.masinter@adobe.com>

>> OK, so... let's fix that; let's set up
>> 	http://www.iana.org/assignments/media-types/text/plain
>> and so on, similar to the existing materials, and let's get
>> IANA to guarantee not to change it without due process.
>
>IANA may not be willing either to set up such a web site or to
>make a guarantee that they would maintain a "web site" for
>an indefinite period of time.
>
>In any case, the only mechanism for the IETF to give "instructions"
>to IANA to ask them to do such a thing is for someone to write
>an Internet Draft describing what it is they're supposed to do,
>and then to have that Internet Draft accepted for publication as
>a BCP.

Well, at least in principle it used to be possible for the IAB to
instruct IANA but it essentially never happened and these days IANA is
maybe sorta part of ICANN or who knows what...  But as a practical
matter you are right.

>In either case, you need to describe what it is that you want
>done. Certainly having IANA maintain an orderly web site is
>a good idea, but, since the data at the web site isn't necessary,
>it doesn't seem like it should be a requirement that there be a
>"web site" merely to have a mapping from IANA-registered tokens
>into URI syntax.

Agreed.

(If the idea is for a human to browse info about protocol parameters,
they can find their way around whatever currently is the site which
should certainly change rarely.  If the idea is that some info needs
to be automatically and mechanically retrieved, then you need to
engineer an appropriate solution, like "in-addr.arpa" or whatever,
which is almost certainly dependent on the particular protocol
parameter.)

>> Let's not pretend that a different URI scheme
>> will somehow magically provide more stability
>> than http/dns; stability depends on
>> social practices, not just technology.

The societal practice of DNS is that it's a giant political football
subject daily to litigation, arbitration, and loud arguments that it
has terrible data quality, provides terrible service, and should be
junked and replaced by something completely different and much better.

>I don't think there is any "magic", just simple engineering:
>systems with fewer points of failure or more stable than those
>with more points of failure.
>
>"IANA loses their domain name"

Or some other organization replaces IANA for some parameter(s).

>"domain name system changes dramatically"

Witness current suggestions to change to a new CLASS of DNS for
internationalization, etc.

>"http goes out of common use"
>"no one is willing to maintain a gopher server for IANA"
>  (ooops, I meant 'http', I know those old gopher URLs aren't
>   good any more, but those http URLs will be good for 20 years,
>   by golly).
>
>Those are all points of failure. Why introduce them? They're
>unnecessary bits of social practice.
>
>Larry

Thanks,
Donald
===================================================================
 Donald E. Eastlake 3rd                    dee3@torque.pothole.com
 155 Beaver Streeet                         lde008@dma.isg.mot.com
 Milford, MA 01757 USA     +1 508-634-2066(h)   +1 508-261-5434(w)

Received on Friday, 19 January 2001 01:11:07 UTC