Sorry, I had forgotten that the original Q was phrased in the context of MIME,
my definition is a working one for the real-world as a whole, and may not be
applicable to the MIME definition list as it stands.
Charset - an algorithm ...
somehow the word algorithm is uncorfortable to me here---to me, it implys a
formulaic calculation of some sort. I *know* table lookup is just as valid an
interpretation, but it doesnt seem as right as "mapping" to me, rather subtle
differences, but still my perception, for what ever that is worth.
> "octet streams" is an overspecification.
No, because the stuff that is in a particular charset is
passed to a CTE encoder (if any), i.e. quoted-printable
or base64, and these take octet streams as input *by
definition*.
Ok, probably stirring up old graves, *maybe* we need Charset-8, Charset-16 and
Charset-32 then, with appropriate 16 and 32 bit CTE mechanisms/front-ends as
well, but I dont think that our definition should try to fold 32 bit streams in
with 8 bit streams, that is just beging for trouble. Of course, that leaves us
hanging for a way to deal with 8/16 streams such as apple uses nativly for S-JIS
(bit 8-off, next byte is a leading byte. Bit 8-on, look up value, some values
eat next byte), and other such nasty to scan streams (adobe has been
particularly inventive).
If you need that many words to describe your intended
meaning of the word "conventionalized", then it is
probably wrong to choose to use that word.
I choose to beleive that it shows off the power of the word, and therefore
indicates that it is well chosen :-)
Yup, and I showed how your definition didn't mesh well
with the consensus of this working group that
iso-2022-jp is a "charset".
Did you? I missed that somehow, in the volume of talk that has been going on
lately, but I will accede that you may well have, and since I dont really care,
I will stipulate it.
Wouldnt be the first time I bucked a consensus, occaisionally some truth emerges
from the struggle, dont really care whoose it is, mine, yours, the ET in the
corner.
If my definition is shared by the average reader, while our usage is going to be
significantly at odds with it, then we had better have a prominantly displayed
and truly memorable glossary entry to explain the differences.
Maybe we should seek a different wording?
(he said, ducking for cover :-)