ietf-822
[Top] [All Lists]

Re: NULL

1994-10-17 22:59:32
I don't believe that the 8bit content-transfer-encoding was intended
to be used with contents that can contain the entire range of octet
values from 0-255.

If I understand Ohta-san correctly, he believes that 8bit is an
appropriate label for such cases.

OK.

I think we agree that the current MIME spec does not sufficiently 
specify when the 8bit label is or is not appropriate.

No, I don't. I think some people are accustomed to ignore clear
specifications on allowable code points.

I believe the current spec does not say what was intended for 8bit,
and also that the spec is different from the way 8bit is currently used.
The solution in this case is to change the spec to say what it should
have said in the first place.
 
And, then, some people even say 8bit is CR, LF, 32-126, and 160-255
and impose further non-standard restrictions on what can be transmitted.

This is only my attempt to state clearly what I believe *was* intended
by 8bit.  

For example, though Keith think that line wrapping, which is not so
harmful for Euuropean languages, should be allowed, it is harmful
for messages with ISO-2022-JP or ISO-2022-KR encoding.

I think that the content-transfer-encoding of 7BIT or 8BIT or BINARY
is NOT meant to prohibit certain operations in transport (0x80 bit 
stripping, line wrapping, or otherwise).

Instead these content-tranfer-encodings are only meant to describe 
what range of octets is used in the body part when the message is 
submitted.

So it is not really a matter of "allowing" line wrapping on 8bit 
encoded body parts.  Informally (and imprecisely) the labels were
intended to mean:

7bit:   this body part will survive being transmitted through 
        traditional Internet MTAs and gateways, without losing too
        much information.

8bit:   this body part will survive being transmitted through
        traditional Internet MTAs and gateways, that have been 
        modified to NOT strip the 0x80 bit during transmission, 
        (or on receipt), without losing too much information.

binary: this body part requires more transparency than can be
        expected from traditional internet MTAs and gateways,
        it might even require true end-to-end 0-255 transparency.

Keith

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