ietf-822
[Top] [All Lists]

Re: charset philosophy

1991-07-11 08:35:34
      Content-Type: text/us-ascii

Today, when ASCII messages cross a border into an EBCDIC world, they
are converted to EBCDIC. Assuming that such a gateway is fully
upgraded to conform to the new RFC-XXXX, what should be done about the
content type header? Should it be changed to the following?

      Content-Type: text/us-ebcdic

With the understanding that Erik asked this question very carefully and 
did not refer to the BITNET-instance of this problem, but the more 
general "EBCDIC world" case...
I have just put on my "member of the committee that gives technical 
advice to the CREN Board" hat, and....
  Once RFC-XXXX and the 8bit transport issues settle in a bit, and a few
of us are able to come up for air, a BITNET document will get written 
that attempts to offer guidance for the handling of these mail 
extensions things at the gateways.  Since, in practice, two MHS systems 
account for in excess of 95% of the gateways, the authors of one are 
watching for that document and one of the authors of the other one is 
clearly one of the perpetrators, I'd expect rather speedy agreement and 
implementation.
  Now, part of that is intended to say, in a backwards way, what I've 
been saying in a forward way for some time:  other than making sure that 
there is enough information present at the gateway for it to make decent 
decisions, what happens on the other side of the boundary is not an 
internet issue and I'd like to take this discussion off-list in a hurry.
  While I can't predict what will come out of the BITNET-side drafting 
process, I would assume (as likely co-author) that at least the first 
draft will specify:
 o  Some trace ("Received") field action that specifies *exactly* what 
has been done and by whom.
 o  Transformation of the header field to reflect the actual character 
set in the message as it is to be delivered, e.g., yes, "something like 
that".
  I would suspect that, when we ask Ed Hart's group about this, they 
will tell us to specify a code page number, rather than "us-ebcdic", so 
you would be more likely to see something like "text/ebcdic-nnn", but
that is still clearly in the "something like that" domain. 
   --john

<Prev in Thread] Current Thread [Next in Thread>