At 07:16 PM 11/22/2001 -0500, Donald E. Eastlake 3rd wrote:
I don't see it as standardizing a practice but just allocating an
identifier for an already standard protocol. Like it or not, people
want to be able to express everything that they reasonably (or
sometimes unreasonably) can as a URI.

I concur with Patrik and Donald.

Trying to discourage use of a protocol by making it a bit more difficult for a user to invoke it feels like the same logic that tries to achieve "security through obscurity".  It even harkens back to the argument that van jacobsen's header prediction, and related work, should not be allowed on an ethernet because it permits a user to take too large a percentage of the LAN capacity.

We DO need to find ways to explain why a practise is bad and to discourage that practise.  However a parental tone is certainly not the way to achieve more competent developers and users.

Offhand, my guess is that some sort of formal specification effort, to develop explanatory details in Worst Current Practises documents, would be more productive. 

d/

----------
Dave Crocker  <mailto:dcrocker@brandenburg.com>
Brandenburg InternetWorking  <http://www.brandenburg.com>
tel +1.408.246.8253;  fax +1.408.273.6464