It is certainly worth thinking about, but I don't know if it will gain us
very much. If all the devices were 1284 compliant, then it would be
possible to design a 1284 API and implement drivers for the different
kinds of physical ports, but PLIP isn't a 1284 device, and from my
work implementing the driver for the ZIP I don't think it is either.
I tried to compare the ZIP protocol with 1284 - hoping there would
be a correspondence that would validate my reverse engineering effort.
1284 specifies how one strobes the nybbles in 4-bit mode, for example,
but it doesn't resemble the way the ZIP does it.
I've also been studying the protocol used by the ShuttleTech chips (as
found in the parallel EZ135, for instance). They use a 16-bit protocol
(in fact, they seem to be doing ISA bus replication), while the ZIP
and PLIP are 8-bit, as are printers.
There is, however, one thing that *is* desperately needed, and that is
a sharing protocol. The ZIP drive and the EZ135 both support chained
printers. And, if you want to risk your data, you can actually run both
the ppa driver and the lp driver on the same port. (Bill Riemers has been
playing with this and reports that it works - I discourage it because
of the danger that unexpected printing activity would corrupt a disk
data transfer.)
I recall that someone raised this question a few months ago - Ulrich
Windl, perhaps - but the thread died without any interesting discussion.
(And I was too busy to push it.)
--------------------------------------------------------------------------
Grant R. Guenther grant@torque.net
--------------------------------------------------------------------------