ietf-822
[Top] [All Lists]

Re: Thoughts about characters transmission

1993-07-10 10:24:51
        [please excuse this cross-post;  I am following a thread]

The _most_important_point_ is that a single common representation code
be defined _for_the_line_ (suiting the purpose, namely to cover all national
languages in one single way) and that people be instructed that every bit
of text should travel in that code on the wire, whatever_the_protocol_is.

I agree to most of what Andre'' is saying and I have an additional
point here: that the single common representation code should be something
that can be handled by existing software and hardware,   ...

        I agree with most of what Andr) said,  and agree with you on
this one point.   But ...

will take a long time before the conversion software is installed
on all machines, or even a large share of the installed base.
Also I would like to emphasis the need for world-wide solutions.
This would mean that ISO 8859-1 would not be a good candidate,
we need something ASCII based (or even with a smaller repertoire
than ASCII to cover the problems with EBCDIC and national ISO 646
variants).

        I don't understand the warrant here,  Keld.   You're right that
we need world-wide solutions and you're right that we should have some-
thing ASCII based.   How does these make ISO 8859-1 a bad choice?

        I've spent a significant part of *my* life working with others
toward a true solution to the  ASCII <---> EBCDIC  problem.   Some form
of concensus was reached a long time ago and folks have successfully
"beat IBM over the head"  with it,  and IBM has finally acknowledged a
"de facto network EBCDIC"  [my term]  which they call CodePage 1047.
CP 1047 maps one-for-one with ISO 8859-1.   The mapping of 1047/8859-1
is the most palatable mapping to the most sites on the InterNet.

        I see the common code Andr) mentions.   I see ISO 8859-1
"on the wire".   I see some  greater-than-8-bit  code in the future
that is a superset of  8859-1.   (and whether TCP has been super-
ceeded or wether we "tag" things,  I am NOT addressing here)
What's the problem?

        [I think it was Nathaniel who said,  "memory is cheap and
bandwidth is cheaper".   In agreement,  I say we scrap the 16-bit
stop-gap solution and go directly to 32-bit and then start looking
toward bit-unconstrained (bit-free?) representations.   Just my opinion]

Keld

--
Rick Troth <troth(_at_)rice(_dot_)edu>,  Rice University,  Information Systems