[core] Re: I-D Action: draft-ietf-core-uri-path-abbrev-02.txt

[Jon Shallow <[e-mail]>]

Hi Christian,

I don't have any concrete examples where discovering UPA helps - apart from
perhaps a multicast scenario where there may be legacy servers responding.
Also "in which case discovery can indicate a short non-well-known URI" which
is similar to also reporting the upa=nnn as an alternative for clients that
can use it.

"concerns of lifetime" leads me to ask - should we have a
Location-Path-Abbrev in a response where the server after (POST/PUT)
dynamically adds in a new Uri-Path-Abbrev mapping?

Appreciate the client is likely to remove that dynamic resource with a
DELETE (and hence should know better), but a 4.02 response following a later
use of the Uri-Path-Abbrev mapping for the now deleted resource is a bit
confusing compared to a (not allowed) 4.04.

This then leads me to ask - if the server understands Uri-Path-Abbrev, but
not the particular value, should the response be, say, 4.07 (Specific
Uri-Path-Abbrev value not supported)?

Regards

Jon

-----Original Message-----
From: 'Christian Amsüss' <[e-mail]> 
Sent: 23 October 2025 10:09
To: Jon Shallow <[e-mail]>
Cc: [e-mail]
Subject: Re: [core] I-D Action: draft-ietf-core-uri-path-abbrev-02.txt

> Should a response to .well-known/core include something like a 'upa=nnn'
> attribute for any known Uri-Path-Abbrev option values to indicate that 
> this resource can be accessed using Uri-Path-Abbrev?

I don't think so. .well-known URIs are there because something is used
without prior discovery. Either support for something is discovered (in
which case discovery can indicate a short non-well-known URI), or it is
available without discovery (in which case Uri-Path-Abbrev makes sense).

Also, discovery carries with it concerns of lifetime (the time that any
discovered EST resources would be available for was part of the early
discussion), and leaving this out simplifies things.

Do you have any concrete example where you think discovery of an UPA would
be useful? (Best that comes to mind for me is .well-known/edhoc, but for
that there's the updated ACE EDHOC profile document that just makes support
for the compressed form mandatory, and thus discovery
moot.)

BR
c

--
To use raw power is to make yourself infinitely vulnerable to greater
powers.
  -- Bene Gesserit axiom