Excerpts from TODO: 30-Mar-92 Re: mime formats and versio.. Larry
Masinter(_at_)parc(_dot_)xero (1497)
If I want to send you a Postscript-2 document (say, with a JPEG
compressed image), can I use the "ps" content-type? Is that
legitimate, as long as the enclosed postscript file is clearly marked
internally?
You certainly can, but I don't recommend it -- using separate multipart
pieces increases the probability that a recipient will at least
understand SOME of it, but there's nothing illegal in what you describe.
Say, in a couple of years, I want to send you a GIF94b image which
includes a little animation and a audio description of the image --
can I use the "gif" content-type? Presuming of course that GIF94b
format is somehow 'upward compatible' with GIF89a?
Given that presumption, I don't see why not.
Recentmessages from you and Ned Freed seem to indicate that you prefer
not to use versions at all, and that we should rely on markers within
the content of the body to detect different versions. That is, the
"ps" body-type will be used for postscript now, and for postscript
later; the "gif" body-type might well be redefined to include "GIF94b"
at some later date, and that will be OK because readers will be able
to detect that it isn't GIF89a and act accordingly.
I think John Klensin summarized my position quite well on this: I don't
mind this approach so long as the new version of the format is largely
backward-compatible, which is a fuzzy concept that will have to be
evaluated on a case-by-case basis. But if a format already has
mechanisms built-in for version evolution, I suspect we're better off
using them than layering another version on top.
Yet, in another discussion, I got the impression that when you say
"ps", what you really mean is "postscript version 1 as defined only by
the 1985 book", and that if I wanted to use Postscript version 2, I'd
have to use a different content-type marker, e.g., "ps2" or "gif94b".
I don't think you can have it both ways. Could you please clarify what
you intend?
I defer to Ned on matters of PostScript, for reasons that his latest
missive should make obvious... -- Nathaniel