My suggestion would be to try and promote some sort of "flag day"
(say N months after RFC-MIME becomes a proposed standard) after which
people are strongly encouraged to use base64 instead of uuencode.
Since I'm the one who has been making these picture posts, I feel I should
reply here to all group readers. First, the defence would like to present
"exhibit A": all my posts have had the following type of disclamer at the top:
| (Please note: this posting has been made in a experimental MIME-like format;
| It should be compatible with all the old posting formats, but of course
| x-uue is not a standard MIME content encoding -- it has been used here for
| old posting compatiblity but it should be simple to add to your MIME system)
Now, getting people to accept and use MIME will be a pain, no matter how
good a standard it is. There will always be sites where their hardware or
software limitations prevent it from being used (remember, "the net" is not
a bunch of Unix machines with virtually unlimited disk space -- even a
humble MeSsy-DOS machine could nowadays receive a full newsfeed using the
right software and hardware). If tools for manual decoding of the images
(sounds, whatever) are available, it will be easier to convince people that
base64 and all the other MIME features Are Not A Bad Thing. There will be
people who will complain bitterly whatever one does to insure ease of use
("I have no C compiler on my machine, I can't compile the converter program
you sent me !"), but any change has it's opponents. A "flag day" would seem
to be the most workable solution, giving most users time to get ready for
the change.
I'll also criticize the MIME setup slightly: there is no provision for the
simplest error detection, which seems to crop up randomly all over the net.
At times, some parts of the multipart images I've been posting end up
scrambled at some sites. This seems to have no pattern, I've rarely
mailed these damaged parts several times to the same machines. This is
why it is customary to include some kind of checksum information with the
images posted, to assure people attempting to decode them that failure is
not their fault, but the image has been damaged somewhere along the route
to them. It would seem that making this a part of the standard would not
be a bad thing, probably the easiest way would be to make it some kind of
a extension to base64. Some kind of a simple CRC calculation (preferrably
incremental, so errors could be pinpointed in large multipart messages),
which the simplest implementations of decoders could ignore would nicely
enhance base64. How about a simple "ignore rest of line" type of char,
which could be used as a preamble for more sophisticated decoders ?
PS. Group replies please, I am not a member of the MIME group. Should I be ?
--
/* * * Otto J. Makela <otto(_at_)jyu(_dot_)fi> * * * * * * * * * * * * * * *
* * * */
/* Phone: +358 41 613 847, BBS: +358 41 211 562 (V.32bis/USR-HST,24h/d) */
/* Mail: Kauppakatu 1B18/SF-40100 Jyvaskyla/Finland, ICBM: 62.14N25.44E */
/* * * Computers Rule 01001111 01001011 * * * * * * * * * * * * * * * * */