[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

MIME acceptance test--where's the break-even point?



OK, Tim, I'll see your 80 column ASCII and raise you a proportional font :).

Most of the objections you (and others) have raised about MIME have centered around graphics. However, MIME isn't fundamentally about graphics, QuickTime� movies, and so on. It's a mechanism for tagging parts of messages, and associating meta-information (such as a digital signature) with one or more such pieces.

This message, for example, is formatted as text/enriched instead of text/plain. It's still readable on 80-column ASCII terminals. It's got some extra stuff, but so does every PGP-signed message, or worse yet a PEM-signed message, and I would argue that the MIME formatting itself is no more objectionable than these, just as I would argue that base64 encoding (the format that graphics generally appear in when MIME encapsulated) is no more objectionable than other mechanisms that serve the same purpose, such as PGP "ASCII armor", uuencoding, etc.

However, for people that do have a MIME mailer, this message will appear in a proportional font, with all of the little touches that we expect out of even the most basic office memo these days. More importantly, I find it literally more readable. If people start using HTML for mail, messages will look like this (only with yet more stuff in angle brackets, since HTML was not designed with non-aware readers in mind the way text/enriched was.

Do you find this message to be "out of bounds" the way you found my (intentionally excessive) GIF signature from a while back to be? If so, where's the boundary between this message and the innumerable PGP formatted messages we see come by on this list? This isn't a rhetorical question, and I'm interested in feedback from other folks on the list as well (which is why I'm sending this message to everyone and not just Tim).

How far are we willing to inconvenience the least common denominator in order to provide the services we want (whether those services be authentication & encryption, multimedia content, or anything else)? Where to the cost and benefit curves cross?


Amanda Walker
InterCon Systems Corporation