ietf-822
[Top] [All Lists]

Re: MIME's 'Content-Disposition' Header

1995-01-13 12:45:11

"Terry" == Terry Crowley <tcrowley(_at_)kilsythe(_dot_)banyan(_dot_)com> 
writes:

  Terry> This is sort of contemplating your navel here, but a
  Terry> content-disposition of attachment on the main body does make
  Terry> logical sense, it's just that most MUA's don't let you
  Terry> compose such a beast.  If the Content-Type of the message as
  Terry> a whole is Image/gif, then it certainly makes sense to
  Terry> indicate that the disposition of that gif object might be
  Terry> attachment rather than inline.

  Terry> So why should this be considered "undefined"?  Granted, MUA's

Here's my original thinking on the subject. 

cdisp was to (and at some point probably will, I envision) feature a
"hidden" disposition type. This is to support compound documents with
links to neighboring bodyparts. What would it mean for a whole message
to be Hidden? Other disposition types might be defined that did not
make sense for whole messages, but only for body parts. Can of
worms. Best to ignore it and declare it illegal at the message level.

If one says a top-level message is an attachment...what's it attached
to???  

Your point about the GIF is good, though. Filenames for toplevel
messages, too. maybe we should loosen the spec.

-Rens