At 12:59 PM 8/15/99 -0700, Keith Moore wrote:
>At least when the labelling is accurate, the recipient has some chance
>of being able to do an appropriate conversion.
Accurate. Not unique. Simply sufficient to make it possible to attempt
conversion(s). This meets the technical requirement, and is appropriately
sanguine about outcomes.
There is an access issue (who has standing, procedural requirements) to/for
adding character sets to registries, which is distinct from the IETF/IANA
mechanism and policy for access. However, this isn't an issue unless one is
convinced of the technical necessity for registry ordering (as distinct from
accuracy of labels, etc.), and thence of uniqueness of the registry/registed
object relationship. The latter two points have been argued, one via "full
stop" reasoning, without reference to the access issue, or it is argued off
this list as a non issue. I don't find this at all compelling.
P.S. Least my early partial agreement with Frank be taken too broadly, we
do (IETF) standardize things not simply "on the wire". Examples of this are
mibs (and the machine state they represent in some canonical ordering), and
the socket API, though the latter will superficially appear to be an act of
The Open Group, and I trust evenually, a QoS API. I'm sure there are more,
I simply want here to limit the scope of Frank's original (and dashed off in
haste) note, or at least my concurence to it.
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:51 EDT