[Date Prev][Date Next][Subject Prev][Subject Next][ Date Index][ Subject Index]

Re: OS/2 And Procomm



Carl:

 Everybody has little troubles with OS/2 at the start. But soon, they pass.
You solved your 5.25" installation snag? For others who might have this
problem, CD-ROM is the best way to go, hands down, for direct-from-CD
installation of OS/2. It's pure pleasure. There is OS/2 driver support for
lots of CD-ROMs now, even for the Mitsumi (resting on the rocky bottom of the
CD-ROM barrel). In fact, the only remaining absentees that I know of are the
cheapo Scuzzony proprietary-SCSI drives (e.g. model 7205), where Sony and IBM
continue to bicker about who should write the drivers. Failing that, one can
beg/borrow a CD-ROM drive of any type, and make image disk copies in any
desired disksize, using a DOS|OS/2 compliant utility called LOADDSKF.EXE (found
on the CD-ROM). Failing that, I'd swap cables inside the machine and put 3.5
on the A: cable and 5.25 on B: -- something I did on one machine. It only took
5 mins, although you gotta change the machine's SETUP (CMOS) info at next
bootup. Nowadays it makes more sense to have the 3.5 drive on A: anyway.

 _Re_ ProComm: I should preface that I have no experience with this program.
First I checked various IBM documents for mention of ProComm as a misbehaving
program, but find nothing. Au contraire, many users report success with
ProComm. Still, your experience sounds much like mine with Hayes SmartCom,
which I tried to run in a DOS VDM under OS/2 (after buying a Hayes
14400+14400FAX Ultima). Sometimes I froze right on entry; but occasionally I'd
get to the dial stage, and then freeze.

 There are a slew of problems with DOS communications programs operating under
OS/2. Many such apps manipulate hardware ports directly, which is something
OS/2 would like to do in an orderly fashion. Others generate more than 1000
hardware interrupts per second, generally when they are communicating at speeds
higher than 9600 baud; 1000 is the upper limit for OS/2's default DOS
communications drivers (9600 baud=960 chars per sec, each of which=1
interrupt). DOS fax programs cause the most complaints of all.

 The solutions are several. First, not all DOS comm/fax programs are bad
actors. PIBterm, my DOS mainstay, runs fine. Second, comm/fax programs
designed from the ground up for OS/2 work very well, and allow asynchronous
communication at the highest rates. TE/2 (TE2_123.ZIP, Terminal Emulator/2
from Oberon Software) is a pretty good shareware comm program that you might
try, with all the standard features -- and there are others. Third -- highly
advised -- dump C:\OS2\COM.SYS and C:\OS2\MDOS\VCOM.SYS in favor of two
shareware drivers written by Ray Gwinn called SIO.SYS and VSIO.SYS (latest is
SIO110.ZIP). These create simulated 16550A chips, essentially a buffer for
data moving through the COM ports. The SIO drivers enable me to run PIBterm at
19200 baud, no coughing no wheezing. Be careful to specify correct port
numbers, port addresses, and IRQs for everything above COM2, if you use them
(COM1 and COM2 are auto-detected); depending on your bus -- ISA, EISA,
Microchannel -- you may have to specify unique IRQs for each COM port; the SIO
doc file has lots of info on these matters. (A real important point that IBM
staffers impressed upon me is that OS/2 is particular about its hardware
environment. Whereas DOS would work on any piece of junk, and was tolerant of
faulty configuration, OS/2 demands a certain standard: thus, a very cheap
TwinComm serial card of mine simply wouldn't work; and OS/2 sneered at a whole
(Windows 3.1!) computer belonging to my son with a cheap Taiwan 386-SX33
motherboard -- OS/2 just wouldn't install. Many DOS comm programs opened up
*all* your COM ports, and if devices were attached to one or more, the program
manipulated them. No discernment at all. Whereas OS/2 dispenses hardware
resources precisely.) You can assist ProComm by closing down other apps, by
keeping it in the foreground where it gets a bigger slice of processor time, &
etc. I don't know how much memory you have, but four megs total would slow
things down badly; another 4-8 meg makes an amazing difference in overall
responsiveness. And fourth, you can tune your DOS settings: IDLE SECONDS=0,
IDLE SENSITIVITY=100, HW ROM_TO_RAM=ON, INT DURING IO=ON (this may help, but
could also hurt), turn Sio_Allow_Access OFF for every COM port except the one
you're using, set DPMI and XMS memory to 0... and so forth.

 If none of this works, call Atlanta and activate your 60 days of free phone
assistance.

 What you really need, Carl, is more info than you're getting from the Command
Reference and other packaged docs. There's plenty of info out there, god
knows; you just gotta find it. Look for frequently-updated BBS files called
OS2TNT.ZIP, OS21PERF.DOC, OS21TUNE.DOC, OS2CNFGn.TXT -- and others. I can
bundle the above-named in a ZIP if you or anyone is interested; but at 2400
baud it's lengthy. There are very good BBSs with OS/2 support: Fernwood in
CT, OS/2 Shareware in VA, IBM NSC in Atlanta, CIS (if you can stand it), etc.
Lots of local gigafile BBSs echo every file found on the flagship BBSs.

 Like you, I don't relish the upcoming 32-bit OS war. Moving up vertically
from DOS was thrilling, so I didn't feel the exertion of learning a new OS; but
moving horizontally over to NT or Chicago or Cairo and a new set of quirks is
odious. I wish these guys would settle it amongst themselves first! Quality
may not be what matters (witness the UNIX ghetto); Microsoft's marketing clout
may be the all & only. The GENUINELY DISTURBING aspect of Microsoft's
dominance is that they refuse to embrace industry-wide standards in many areas,
intending instead to develop their own. They've chosen incompatibility very
deliberately. And the $price$ attached to WindowsNT is just a hint of what
will happen if Microsoft has no competition. All of this uncertainty makes me
cling to my DOS apps even more fiercely. I'm reluctant to buy or, more
importantly, invest the time to _learn_, software that exploits the new
technologies: big OS/2 apps like DeScribe (despite my great curiosity, and
Herb's enthusiasm), lest they quickly peter out; or Windows/DOS v3.1 apps, on
the ground that Win3.x is a temporary fix by everyone's admission, even
Microsoft's.

 Still, I think there's a solid chance that Microsoft won't meet ebullient
expectations, at least not right away. They don't have much experience writing
Operating Systems (after all, MS-DOS truly stinks, and the real wonder is, that
MS got away with such an atrocious hunk of dung for so long). Most of their
products are lousy in the initial implementations, witness OS/2 and MSWord (IBM
had to rewrite most of OS/2 from scratch, Microsoft bungled it so badly). NT's
hardware demands are steep and expensive -- OS/2 vs. NT is _very similar_ to
the situation ten years ago of MS-DOS vs. UNIX. Much of NT's cost represents
powerful capabilities that very few of us need -- that NOBODY running a
standalone machine needs (multiple processors, multiple keyboards, high levels
of security, platform portability, etc) but that you can buy (as add-ons) for
OS/2 if you do need 'em. In short, I don't think MS dominance is a foregone
conclusion. Lemme ask you: which MS product do you use and like? Think hard!