There is at least one serious problem with the "mnemonics" document:
It refers to the "charsets" document, which you say may become an
informational document (as opposed to a standard).
Erik,
Keld has split the nmemonic document into two at my request.
The previous document joined two types of information, the algorithm
for translating a stream of nmemonic characters into a stream of nmemonic
characters, and the informational listing of existing character sets
and characters. The "algorithm" is not expected to change, and should
be published separately. The nmemonics are similar to SNMP MIB's
in that they can be defined separately and used in the framework of the
nmemonic proposal. An initial "Cib" Character INformation base is
defined in the charsets document. Also included in the charsets document
is a tabling of character sets using the terminology (nmemonics) of
the charset document. This is in no way standardizable material in that
the IAB cannot claim any change control over say, 8859-1 or any other
character set.
Now, it could be argued that the two documents should be
split into three, 1) Algorithm, 2) NMemonic "CIB", and 3) informational
republication of character set tables. A nmemonic is really just a simple
naming of a well defined character. With a loose interpretation of
standard, the CIB could be considered inforatmational and be
joined with the list of character set tables.
The important thing is this is just editorial and does not really
affect the intentions for nmemonic. If this is the only issue, I'm
sure Keld will make yet-another-editorial pass.
The real issues is: Is the current "profile" adequate for the intended
uses? Other profiles are possible to define later. The main objection
I see so far are that it does not deal well with oriental character sets.
This working group has yet to come to a full understanding of the
requirements and the mechanisms needed for those languages. Attempts
to solve these problems have been incomplete at best. The current
hope is that some set of experts will take the MIME format and define
a character set which will work. If NMEMONIC is not it, then work
to find one that will work! NEMONIC is not likely to be the
end-all of character sets, and if this groups gets bogged down trying
to make it so, we will remain forever in painful flamming deadlock.
Greg Vaudreuil