RE: Ping

Thanks for your reply and your comments, Craig.  

First, I'd like to point out that the OID-based scheme is no longer in the
process draft.  After listening at the last WG meeting, talking with many of
the more active members of the WG, and putting the idea out on the list,
Rich and I rewrote the draft to (a) document the process for the IETF tree,
and (b) document further process for creating new trees, through the filing
of other RFCs.  The vnd doc is the first effort in this direction.  (I've
spoken with Yaron Goland, the author of the DNS scheme, and he stated an
interest in touching up that draft and resubmitting it as an alternate tree,
pursuant to the process doc.)  

You mention other properties that appear to have vanished -- can you be more
specific?  I'd really like to discuss them, both here on the list and at
least informally in Minneapolis; the process doc is not set in stone, but we
would like to get it finished soon.  

BTW, I agree that the DNS-based method has some benefits, but there were
also some drawbacks and concerns discussed at the last WG meeting and on the
list.  The vnd.* scheme syntax is already in use, at least informally, and
seemed to have the benefits of broad interest and significant utility.
That's why I chose it for formalization.  DNS may well be next... if there's
enough interest.  

Again, thanks for your input.  Please feel free to expand further on your
observations and opinions, and let's talk about it!  
------------------------------
Ian King <[email protected]>
Internet Bill Delivery & Payment
MICROSOFT CORPORATION

-----Original Message-----
From: Craig A. Finseth [mailto:[email protected]]
Sent: Thursday, March 11, 1999 1:19 PM
To: Ian King
Cc: [email protected]
Subject: Re: Ping


   As I'm sure you're aware, we do not have a meeting in Minneapolis;
however,
   we do have a draft out there, <draft-ietf-urlreg-procedures-05.txt>.
I've
   seen no public discussion of it -- is everyone satisfied with this?  (The

I'm speaking for myself and not the group as a whole, but I'm not sure
that _anyone_ is satisfied with it.  My recollection is that after it
was last revised, I asked a number of questions (so, how does one get
an OID?  is the process cheap and easy?  etc.) and that there were no
good answers to these questions.  Also, the properties that some
people had thought best about it appeared to vanish upon closer
examination.

Again, speaking for myself, I don't see myself using any number-based
scheme.

[snip]  
   While you're thinking about this, please consider the draft I sent out
   yesterday on the vnd.* URL scheme tree.  
[snip]  

I like it much better.  However, I'd personally prefer the re-use of
an existing identifier (e.g., a DNS string) rather than invent a whole
new one.  Each such identifier incurs a substantial overhead to track
and keep current.  Where's the advantage to not using it (and existing
identifier) that offsets the extra overhead?

Craig