R> Maybe because the 3270 is so complex, weird and little-used nowadays?
it is still a major i/o style for many legacy mainframes. that is why
there are many software emulators around. not much in the way of
physical terminals may exist anymore but i don't know for sure.
R> The original 3270 was connected by Coax cable, not serial cable, not
R> by Ethernet (nor even Token Ring originally).
well it was used way before those were invented. hell, the simple ascii
terminal wasn't in use then too. and they did use serial lines when used
over a modem (at least the modem part did).
R> The original 3270 used a proprietary block-mode communications
R> protocol. I think it was synchronous rather than asynchronous.
bisync. it used a 2 byte sync char and was half duplex/polled.
R> The 3270 uses the EBCDIC character set.
the least of its craziness.
R> Perhaps Uri is busy scrutinizing RFC1576
nope. i actually used a 3270 emulator package from DEC on RT-11 way
before most of you were born. i even found a bug in it which proved that
we were the first real world users of it (it started an ack0/ack1
pattern on the wrong foot). so i know about the protocol (bisync) and
polling and such. this last fall i even helped someone i know with using
some 3270 emulation package under perl. but i won't be writing a module
for this as no itch could possibly make me care about the 3270.
a cluster of 3270s would share a single controller/modem which would
connect over a phone line to the mainframe. so it used some form of
serial line at that level. there might have been some form of coax
between the terminals and the controller. i didn't ever get into
physical use of 3270's so i may be wrong on the wiring but the protocol
i did know some way back when.
uri