ietf-822
[Top] [All Lists]

Re: Another approach to Encodings

1991-04-30 03:28:35
Nathaniel -- I split your paragraph up and reformatted it.  Just want
to warn you so all won't get vertigo or something equally bad from
from the mind warp this might produce out there.

The only advantage I see to putting the encoding into the
content-type field is that it makes it more clear that they're
related, i.e. that it is reasonable to say things like "multipart
content-types should use NO encodings."  

This part I can go along with, and I also want to say that I think
Greg somehow misread what you wrote to think that you were saying the
opposite of this.  

I can see how easily this can happen in this squirrel cage.

Clearly a disadvantage is that an MTA that needs to muck with
encoding will then also need to rewrite a content-type header.  

I thought we had agreed that all mucking with encodings in the
neighborhood of an MTA was to be done in a WELL SPECIFIED GATEWAY!  I
don't like to see talk like this that opens that door again, to
encourage sendmail to muck with whatever it feels like mucking with!
Gives me the heebie jeebies, it does.

If people think that its reasonable to say that the content of the
content-type field can determine what, if anything, is permitted in
the content-encoding field, then I'm happy to keep the two fields
separate.  -- Nathaniel

Yeah, and at this point we find ourselves trying to figure out how to
keep these two independent things in sync.  Seems to me that if we are
going to link things, then we should link them and be done with it.

Best...\Stef

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