home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Power Programmierung
/
Power-Programmierung (Tewi)(1994).iso
/
magazine
/
c_news
/
06
/
unix1.txt
< prev
Wrap
Text File
|
1988-04-09
|
58KB
|
1,411 lines
#: 7966 S9/UNIX for the PC
17-Nov-86 14:53:53
Sb: #Special Topic Intro.
Fm: Levi Thomas (Sysop) 76703,4060
To: ALL
Hello folks and welcome to the ever-changing world of SubTopic 9. This
week's special topic is UNIX on the PC. Our guest is Dwight Leu from
Microport Systems, Inc. of Aptos CA. Microport has ported UNIX System V
to the PC-AT and compatibles and is selling the Runtime System for
about $159. Wow. Feel free to ask Dwight about the product, the
project, the legal hassles from AT&T (just guessing Dwight),
and...well...y'all can take it from there. Mr. Leu will be here until
next Monday, so pace yourselves. <grin>
--Levi (emcee) Thomas
*** There are replies:
7968, 7983
#: 7983 S9/UNIX for the PC
17-Nov-86 21:08:11
Sb: #7966-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Levi Thomas (Sysop) 76703,4060 (X)
Greetings! This is Dwight Leu. I led the System V UNIX port, and did
much of the technical work, from the boot blocks, the kernel hacking,
and AT specific user level utilities. Since I have no idea as to what
people are interested in knowing about Microport, I will leave this
open to you folks, and start off by answering any specific questions.
For those of you who aren't familiar with what Microport has
done, we are offering real true blue UNIX (and not a look alike) for
the IBM AT, and we sell the basic OS and utilities (including vi and
uucp) for $159. The complete package (with software and text
development) for $450. We were also the first people to port System V.3
to the 386. We've also just introduced DOS MERGE, which allows one to
run both DOS and UNIX at the same time. Our basic philosophy was
inspired by Borland: to sell UNIX at the lowest price possible.
Levi wasn't just guessing about some of the legal hassles we've
experienced, and not just from ATT. I believe I mentioned some of them
at the hacker's conference a few weeks ago. If there's enough demand,
I'll go into some of them. Others I won't discuss here, because we are
still under the active threat of lawsuits by various competitors.
One final note: we are no longer located in Aptos. We moved to
Scotts Valley a couple of months ago, into a much bigger building.
-dwight-
*** There are replies:
7985, 7993, 7998, 8019, 8072
*** Reading replies to 7966 ***
*** More ***
#: 7985 S9/UNIX for the PC
18-Nov-86 00:56:45
Sb: #7983-#Special Topic Intro.
Fm: Fred Buck 73327,3604
To: Dwight Leu [Microport] 76703,4266 (X)
Er, can I ask a general question? What's the real diff between System
V and its predecessors (i.e., if someone has, say, Sys 3, why switch)?
I'm not talking here about purely cosmetic or trivial changes like
better utilities X, Y and Z: in what way is the operating system
superior (if any)?
I realize this has little specifically to do with Microport, but on the
other hand, presumably, you guys expect to sell yer product, and also
presumably you expect to sell it to people who may already have a non-V
Unix or equivalent.
*** There is a reply:
8002
*** Reading replies to 7966 ***
*** More ***
#: 8002 S9/UNIX for the PC
18-Nov-86 20:34:51
Sb: #7985-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Fred Buck 73327,3604 (X)
System V is better than it's predecessors in several ways. First, the
OS is faster in how it deals with various lists; it uses hashing, where
previous versions used linear searching. Buffers, inodes and filesystem
freelists are the first areas that come to mind. This results in a very
noticable performance increase.
Second, the OS is much more powerful in its capabilities.
Interprocess Communication and Shared Memory are two obvious examples
in V.2; built in networking is one example in V.3. I don't share your
belief that better utilities are cosmetic or trivial changes; and there
are a number of changes in this area that are definite improvements.
The generic version for the 286 and 386 ports have an great
feature called "file system hardening"; whereby buffers are continually
aged and written out to disk, rather than being sync'ed every 30
seconds or so. The upshot of this is that you can pull out the plug on
your AT at any time, and not have to worry about losing files, or
having serious filesystem damage. With fs hardening, we can allocate
many more buffers without having to worry about trashed filesystems,
and this results in tremendous performance increases.
Finally, System V is "The Standard", according to ATT, but only
you can determine the importance of that to your work. As to why you
should switch to System V, you'll really have to weigh your own needs
in this area. I know of some sites that are still running V6. All I can
say is that I've found that my own work is made easier when I deal with
something that everyone else is trying to support.
-dwight-
*** There is a reply:
8010
*** Reading replies to 7966 ***
*** More ***
#: 8010 S9/UNIX for the PC
18-Nov-86 21:17:00
Sb: #8002-Special Topic Intro.
Fm: Fred Buck 73327,3604
To: Dwight Leu [Microport] 76703,4266 (X)
A fair answer. I didn't genuinely believe that AT&T would waste its
time coming up with System V if there weren't such significant changes
to be made, but I did genuinely have no idea what they might be.
Thanks for the details.
*** Reading replies to 7966 ***
*** More ***
#: 7993 S9/UNIX for the PC
18-Nov-86 14:28:48
Sb: #7983-#Special Topic Intro.
Fm: Bob Peterson 76703,532
To: Dwight Leu [Microport] 76703,4266 (X)
Dwight,
Which of your competition is not, as you put it, "...real true blue
UNIX (and not a look alike)..." As I understand the situation, Xenix
_is_ licensed from AT&T and is based on Unix sources just as your port
is. Is there another port I'm not aware of that you consider to be
your competition?
*** There is a reply:
8006
*** Reading replies to 7966 ***
*** More ***
#: 8006 S9/UNIX for the PC
18-Nov-86 21:00:51
Sb: #7993-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Bob Peterson 76703,532 (X)
In the micro UNIX world, there are only 3 competitors with a
significant share of the market: XENIX (by Microsoft), VENIX (by
Venturecom), and System V/AT. In my opinion, Venix has never been a
serious contender. Xenix, on the otherhand, holds the greatest market
share. Although Xenix is licensed from ATT, it is not real UNIX. While
it is compatible at the System call level (and thus has passed the
System V validation suite), it differs from real UNIX in many ways.
By and large, Microsoft is constrained to be backwards
compatible to its previous releases. For this reason, you see a number
of outdated incompatibilities at the user level, dating back to V7
(e.g. /etc/ttys and termcap to name but two). And it is these
incompatibilities which continually get in one's way when you try to
switch between the two.
While system call compatibility is nice, it is not sufficient
to guarentee that a program which compiles under UNIX will port to
XENIX. We saw a dramatic case of this recently. A customer had 20 Mb of
*source* code that they wanted ported to the AT. It took them 3
man-years (i.e. 3 engineers working a solid year) to port it to Xenix.
One of my engineers did it in two weeks.
The new release of UNIX (V.3) is the best example, though. The
kernel is much more complex that V.2. This means that Microsoft is
having to spend a tremendous amount of effort to come up with Xenix
V.3. In contrast, I and 1 other engineer brought V.3 up on the 386 (a
full port) in *two* weeks. To my knowledge, that is one of the fastest
UNIX ports in history (it usually takes 3-6 months to bring UNIX up on
a new architecture).
Ultimately, I expect to see DOS pass the System V test suite.
But in my opinion, it won't be UNIX either.
-dwight-
#: 8019 S9/UNIX for the PC
18-Nov-86 22:39:17
Sb: #7983-#Special Topic Intro.
Fm: Darryl Okahata 75206,3074
To: Dwight Leu [Microport] 76703,4266 (X)
Is there any chance you could upload an "electronic data sheet" on
Microport UNIX? Do you have any benchmarks -- I'd like to get a feel
for how fast/slow it is.
-- Darryl
*** There is a reply:
8023
*** Reading replies to 7966 ***
*** More ***
#: 8023 S9/UNIX for the PC
18-Nov-86 23:10:55
Sb: #8019-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Darryl Okahata 75206,3074
I'll see what I can do for an electronic data sheet; I suppose that
sales must have something along these lines. As far as benchmarks go,
drystones are in the bottom range of a VAX 750; between 1500-1800 as I
recall. In our standard lit pack is an article about the Nelson
benchmarks, which were performed independently, by Dmitri Rotow and Bob
Glossman, of Bell Technologies. It compared us, IBM XENIX and SCO
XENIX. We came out ahead of both XENIX's by factors of between 10% and
1000%. These benchmarks are weighted towards measuring multi-user
performance though.
-dwight-
*** Reading replies to 7966 ***
*** More ***
#: 8072 S9/UNIX for the PC
19-Nov-86 13:40:16
Sb: #7983-Special Topic Intro.
Fm: ERIC G. ECKBERG 74176,2731
To: Dwight Leu [Microport] 76703,4266
Hello Dwight, thanks for participating in this SIG. I have seen the
ads for your UNIX System V and am trying to get it stuck into my
budget. I have also used PC/IX from Interactive Systems (UNIX System
III) and was real excited about running UNIX on a PC. Now for the
questions: You talk about one partition in the ad. Does this mean that
UNIX and DOS reside in the same partition or does each require it's
own? I'm sure you've got utilities that allow users to transfer files
between partitions if that is required, but what about running say
Lotus 1-2-3 from UNIX? Is that possible? Final question is on a
different note: How could I get a DOS version of NROFF with the
Memorandum Macros without having to go out and buy some complete
commercial package? Is there any way to get ahold of some of the old
UNIX I source code or technical documentation? Thanks.
#: 8085 S9/UNIX for the PC
19-Nov-86 16:57:05
Sb: #Special Topic Intro.
Fm: Bob Peterson 76703,532
To: Dwight Leu [Microport] 76703,4266 (X)
Dwight,
So your definition of "real Unix" is one derived from AT&T sources
_and_ doesn't support any backward compatibility? Does this mean that
Berkley doesn't ship "real Unix"? After all, 4.3BSD still uses
/etc/ttys and termcap!
If you couldn't tell, I find your definition of "real Unix" extremely
narrow, almost to the point of "real Unix" being what AT&T sells
_exclusive_ of ports and/or passing the System V Validation Suite.
There is, after all, a _lot_ of software out there, both commercial and
free, that needs backward compatibility in areas such as termcap.
For example, I'm aware of problems porting to System V very similar
to your anecdote, ie, where moving from System III to System V required
_lots_ of time. In my opinion, the portability of Unix applications is
one of the biggest myths around. I say this having been there.
Please tell me what a "...full port..." of System V Release 3 is for
the 80386. Are you telling me you utilized the 32-bit nature, ie, flat
address space instead of segmented space, by coding a brand new code
generator and recompiling all of SVR3 in two weeks? And are shipping
this capability now? Does this mean that your 80386 version doesn't
have memory models?
*** There are replies:
8103, 8104
#: 8103 S9/UNIX for the PC
19-Nov-86 22:30:13
Sb: #8085-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Bob Peterson 76703,532 (X)
I agree completely with you, Bob; my definition of "real UNIX" is
extremely narrow. And no, I don't consider UCB UNIX to be "real UNIX".
Please note that this by itself doesn't place a value judgement on
whether UCB or ATT UNIX is "better". Such a judgement is based upon the
metric that you use. For speed high tech features, and just better
thinking, UCB wins hands down. Occasionally they screw up, but that's
part of what being a leader in technology is all about.
But I live in the world of business, and have had to deal with
industry software for many years. For this reason, I prefer to stick to
standards; where the strength is using something that people agree on.
One of the reasons that UNIX has done so miserably in the marketplace
is because there have been so many different versions of it around;
which is expensive when you want to port something from Funknix to
Dosnix (as you have apparently seen).
But this is not true when you are dealing with software that
runs on the certified ATT UNIX ports (or at worst, it is considerably
less expensive). ATT is spending literally millions of dollars in
making sure that their certified ports are consistant across all
architectures. And this pays off greatly when you have to port from SV
on a Vax to SV on an AT. It has saved us millions of dollars, and given
us instant application software, because our copy of the certified 286
port is *BINARY COMPATIBLE* with the copy that runs on ATT's 6300+.
Finally, as to the 386 port, you must realize that we didn't
have to recompile *any* 386 UNIX utilities. As with the 286 port, we
merely *copied* them from the generic release. And this includes the
compiler. For the port, we merely had to change the kernel, which
originally ran on Intel's 386 310 box. And to be honest, half of that
time was spent on developing the big boot code, which wasn't part of
the original release. Normally, it takes about 2 weeks just to compile
the utilities. As far as memory models go, yes, there is indeed just
one, with the tiny model thrown in for whoever needs it. The 386 SVR3
has been shipping in Beta form
*** There is a reply:
8115
*** Reading replies to 8085 ***
*** More ***
#: 8115 S9/UNIX for the PC
20-Nov-86 02:21:22
Sb: #8103-#Special Topic Intro.
Fm: Chuck Forsberg Omen Tech 70007,2304
To: Dwight Leu [Microport] 76703,4266 (X)
Will 386 Unix be able to run 286 Unix programs, or must they be
recompiled for the 386 system? Whch 286 models will be supported?
*** There is a reply:
8142
*** Reading replies to 8085 ***
*** More ***
#: 8142 S9/UNIX for the PC
20-Nov-86 17:21:14
Sb: #8115-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Chuck Forsberg Omen Tech 70007,2304 (X)
As things stand right now, 286 Unix programs must be recompiled for the
386! Yes, this is incredibly stupid, and I don't know why the folks who
are developing the generic 386 port have done this. I also don't know
if this is going to remain the same for the final release of the
generic 386 port. We, and many others, have been applying a lot of
pressure on Intel to support 286 binaries. But so far, they haven't
made a firm decision on this, to the best of my knowledge.
All I can say is, that if they don't support 286 binaries, we
most likely will, as all of our application base is on the 286.
As far as 286 models go, right now just the small model is
supported on the 386.
-dwight-
*** Reading replies to 8085 ***
*** More ***
#: 8104 S9/UNIX for the PC
19-Nov-86 22:37:26
Sb: #8085-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Bob Peterson 76703,532 (X)
<this is a continuation of message 8103> SVR3-386 has been shipping
from us since September, in beta release form, but only to select OEMs.
End users will have to wait until ATT certification, which is scheduled
for Feb 87.
I also apologize for the length of this and other messages. I
really have been trying to keep them all as short as possible; but some
questions need a thorough answer. And when many questions are asked at
once, it is difficult to keep a letter short.
-dwightp.s. Now I know about the !#$%^& 35 line limit! jeez,
this smacks of censorship <grin>
*** There are replies:
8124, 8137
*** Reading replies to 8085 ***
*** More ***
#: 8096 S9/UNIX for the PC
19-Nov-86 20:00:07
Sb: #Special Topic Intro.
Fm: Peter Retief 76224,71
To: Dwight Leu [Microport] 76703,4266 (X)
Dwight, It's great to have you here! In my opinion, the most important
question regarding which Unix to buy is the question of what additional
software is available. As far as I know, Xenix Sys V has a large
selection of spreadsheets, database, word processors, etc. How
compatible is Microport with Xenix in terms of these products. Could
you give us some idea of what is available under Microport.
A second question: Does DOS-Merge allow more than one user to run DOS
programs at the same time? What restrictio?s are there on which DOS
applications will run? Thamnks, Peter.
*** There are replies:
8109, 8112
#: 8109 S9/UNIX for the PC
19-Nov-86 23:05:13
Sb: #8096-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Peter Retief 76224,71 (X)
This brings up an excellent topic: applications! System V/AT is in no
way capable of running Xenix binaries. A year ago we had announced that
we would be supporting this in 86. However, we got a very nasty letter
from Microsoft, saying that they considered Xenix technology
proprietary, even if we didn't need source code to implement it. I
believe they threatened to sue us if we so much as even demo'ed this
capability to anyone.
In my opinion, they had no grounds whatsoever for a lawsuit,
just a very large legal department.
Around the beginning of this year, we realized that we didn't
even need their applications anyway. ATT was kind enough to provide
more applications than AT Xenix has. ATT spend millions of dollars in
porting application software to the 6300+. And all of them (in the
hundreds, at least) are binary compatible with our System V/AT. The
*only* thing that we've had to do is to de-engineer the installation
format on the floppies. Every single one has come up without a hitch.
So virtually overnight we had instant applications. And, with
our new DOS-MERGE technology, I believe we now have more applications
than anyone else in the world, including Microsoft.
-dwightDisclaimer: All of the above statements are my own
personal
opinion, and not those of my employer. And they are
certainly not official Microport statements, or policy.
*** There is a reply:
8224
*** Reading replies to 8085 ***
*** More ***
#: 8224 S9/UNIX for the PC
22-Nov-86 10:42:51
Sb: #8109-#Special Topic Intro.
Fm: Bill Hersh 73117,3320
To: Dwight Leu [Microport] 76703,4266 (X)
Dwight:
Some questions. My programming experience is almost exclusively in
the MS-DOS world. I have used other computers, including my
university's CMS/VM mainframe (yuk!), but have never used anything
beyond high-level languages with them. I would be interested in
reading up on Unix, just to get a flavor for the machanics of it. What
book(s) would you recommend? Most of the UNIX books available around
here seem to be very introductory and not tell you anything about the
nuts and bolts of the Unix OS. I would also be interested in some info
about the various releases by AT&T, which you have briefly alluded to.
Essentially, I am an MS-DOS Turbo Pascal and assembly language
programmer. I know all about DOS and BIOS calls. I would be
interested in how programmers do the same under Unix. I would also be
interested in how graphics are coded and implemented.
Thanks,
Bill
*** There is a reply:
8294
*** Reading replies to 8085 ***
*** More ***
#: 8294 S9/UNIX for the PC
24-Nov-86 22:39:09
Sb: #8224-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Bill Hersh 73117,3320 (X)
This best book that I know of about the internals of UNIX is by Maurice
J. Bach. I believe the title is "The Design of the UNIX Operating
System". It has just been published, by Prentice-Hall. While it does
has it's limitations, it is a must for anyone who wants to learn about
the internal mechanics.
As far as User-level books go, I'm less well-informed about the
current offering. S.R. Bourne's book is a classic, as well as Kernigham
and Plaugher(??). And of course, the C "bible" by Kernigham and
Ritchie. All of these are for the more sophisticated user, though. And
they were written by people who were instrumental in UNIX's early
development.
-dwight-
*** Reading replies to 8085 ***
*** More ***
#: 8112 S9/UNIX for the PC
20-Nov-86 00:00:25
Sb: #8096-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Peter Retief 76224,71 (X)
I forgot to answer the second part of your question, so here it is. 286
DOS-MERGE only runs DOS on the main console. 386 DOS-MERGE however runs
on each remote (serial) terminal, in addition to the main console.
Thus, we now have multiuser DOS currently running. Where's Microsoft
with this, you might ask? About a year away.
-dwight-
*** Reading replies to 8085 ***
#: 8099 S9/UNIX for the PC
19-Nov-86 21:47:36
Sb: Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: ERIC G. ECKBERG 74176,2731 (X)
Currently, System V/AT (like the other UNIX ports) requires that DOS
and UNIX be kept on separate partitions. Under DOS MERGE, they can live
on the same partition; and this will allow one to run DOS programs
(like Lotus and Flight Simulator) from UNIX. As for DOS versions of
Nroff, I don't know of anyone who sells such a package; the closest
thing I know of is some company in the midwest (?) who sells UNIX-like
programs under DOS, but I don't know their name.
As far as old UNIX source code, I suspect that you may actually
still need an ATT license (they are incredibly fussy about these sorts
of things). I don't know where old documentation can be had; but some
of the current doc is availible in certain bookstores that carry
hi-tech books.
-dwight-
*** Reading replies to 8085 ***
#: 8120 S9/UNIX for the PC
20-Nov-86 07:31:55
Sb: #Special Topic Intro.
Fm: Darryl Okahata 75206,3074
To: Dwight Leu [Microport] 76703,4266 (X)
Thanks! Do you know the specs of the machine used to run
Microport UNIX (e.g., clock speed, avg. disk access time, etc.)?
Also, using the C compiler, can you have single data arrays larger
than 64K? I'm asking this because I'm wondering if it's possible to
port GNU Emacs to MP UNIX. It probably can't be ported, but it would
be very nice if it could (I imagine you'd have to rewrite large chunks
of GNU Emacs).
-- Darryl
*** There is a reply:
8144
#: 8144 S9/UNIX for the PC
20-Nov-86 17:29:42
Sb: #8120-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Darryl Okahata 75206,3074 (X)
Any AT clone that runs DOS should be able to run our software. I know
of machines that run it from 6-10 Mhz (286). The Toshiba disk drives
(86MB) are dynamite; they have a 23 ms access time (I have one on my
own system). Televideo, Bell Technologies, and Nexus all have very nice
hardware for running several users at the same time.
Unfortunately, you'll have to wait for the V.3 release, for
very large data arrays. This should be out first quarter next year. As
far as Emacs goes, someone was kind enough to port micro Emacs, and
give us a copy. We'll be assembling public domain floppies in the near
future, and these will be availible for just the cost of distributing
them.
-dwight-
*** There is a reply:
8162
*** Reading replies to 8085 ***
*** More ***
#: 8162 S9/UNIX for the PC
20-Nov-86 22:01:22
Sb: #8144-Special Topic Intro.
Fm: Darryl Okahata 75206,3074
To: Dwight Leu [Microport] 76703,4266 (X)
Thanks for the info. I've played around with MicroEmacs and have
found it to be good as far as IBM PC editors go, but poor as far as
UNIX editors go. On UNIX, MicroEmacs is slow in screen updating; vi
updates the screen much faster. MicroEmacs is also missing regular
expressions, not to mention the (mock)lisp programming language.
GNU Emacs, on the other hand, has more of the features that I like
in a UNIX editor (screen updates are still slower than those of vi,
though). It's got regular expressions and the integrated lisp
compiler/interpreter (?). It's incredibly awesome to be able to assign
a lisp function to a key (not to mention being able to interactively
debug the function). In case I haven't been too clear, GNU Emacs has a
"version" of the lisp language built into it. Unfortunately, GNU Emacs
was written for the 680x0-based UNIX systems and makes a few MAJOR
assumptions about how memory is organized. For example, to initially
configure GNU Emacs, the newly compiled version is run, which then
loads the set of lisp functions that you want it to have, after which
it writes a copy of itself from memory (?) to disk. It is this copy
which is then run as "GNU Emacs"; the original version is no longer
needed and can be removed. Not a very good programming practice, in my
opinion.
-- Darryl
*** Reading replies to 8085 ***
#: 8161 S9/UNIX for the PC
20-Nov-86 21:21:37
Sb: #Special Topic Intro.
Fm: Jim Foy 75735,1505
To: Dwight Leu [Microport] 76703,4266 (X)
Does your version of UNIX run on a Tandy 3000?
*** There is a reply:
8175
#: 8175 S9/UNIX for the PC
21-Nov-86 07:49:14
Sb: #8161-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Jim Foy 75735,1505 (X)
Yep. It runs just fine on the Tandy 3000. In fact, we tested out one of
the original releases at our local Radio Shack store.
-dwight-
*** There is a reply:
8179
*** Reading replies to 8085 ***
*** More ***
#: 8179 S9/UNIX for the PC
21-Nov-86 15:05:03
Sb: #8175-#Special Topic Intro.
Fm: Jim Foy 75735,1505
To: Dwight Leu [Microport] 76703,4266 (X)
What languages does the Microport Unix support? Any versions of Basic,
such as MBasic that runs on SCO Xenix?
*** There is a reply:
8187
*** Reading replies to 8085 ***
*** More ***
#: 8187 S9/UNIX for the PC
21-Nov-86 19:03:27
Sb: #8179-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Jim Foy 75735,1505 (X)
Fortran 77 is included in the software development package. I don't
know what other languages will run on it, but the general rule is
anything that runs on the 6300+ UNIX is binary compatible with our
system. And ATT has a huge book listing all of the applications that
run on the various architectures under UNIX. According to them, this
applications book is availible at bookstores (such as B Dalton's) and
your friendly neighborhood library. Sorry I don't know anything further
on this, but I only use C.
-dwight-
*** There is a reply:
8202
*** Reading replies to 8085 ***
*** More ***
#: 8202 S9/UNIX for the PC
21-Nov-86 21:24:42
Sb: #8187-#Special Topic Intro.
Fm: Jim Foy 75735,1505
To: Dwight Leu [Microport] 76703,4266 (X)
For those of us who are not gurus, please explain exactly what "binary
compatable" means. DOes that mean you can load and go, or does the
source code have to be re-compiled, or what? Thanks.
*** There is a reply:
8233
*** Reading replies to 8085 ***
*** More ***
#: 8233 S9/UNIX for the PC
22-Nov-86 14:34:09
Sb: #8202-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Jim Foy 75735,1505 (X)
"Binary compatible" means that you can take a program that's been
compiled on one machine, under one operating system, and move it over
to another machine and have it run immediately, without recompiling the
source.
In our case, all of the application software that runs on ATT's
6300+ runs directly on our System V/AT. This means that you can buy a
6300+ application package, and use it on our software.
-dwight-
*** Reading replies to 8085 ***
#: 8166 S9/UNIX for the PC
21-Nov-86 01:06:37
Sb: #Special Topic Intro.
Fm: Denis Hennessy 73277,1203
To: Dwight Leu [Microport] 76703,4266 (X)
dwight,
Can you tell me whether your unix will support the PC Network?
If so, can it be a server to other DOS/UNIX machines? if not, can
DOS Merge run the PC Network program (server) as a task with unix still
available. Also can DOS Merge run MS Windows?
Thanks
Denis
*** There is a reply:
8176
#: 8176 S9/UNIX for the PC
21-Nov-86 07:53:32
Sb: #8166-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Denis Hennessy 73277,1203 (X)
Our current product doesn't support PC Net. Under DOS MERGE, PC Net and
Windows *may* work; but we're still in the process of bringing the
product into Beta release as of yet. So, the basic answer is "Please
stay tuned". We should know about both of these in January.
-dwight-
*** Reading replies to 8085 ***
#: 8190 S9/UNIX for the PC
21-Nov-86 20:02:52
Sb: #Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Levi Thomas (Sysop) 76703,4060 (X)
Well, since the questions have slowed down a bit, and since I only have
a couple more days left in this forum, I'd like to discuss a few things
that have been on my mind lately. This is with the expressed hopes of
stirring things up here.
When I was at the hacker's conference a few weeks back, at one
of the last discussions there seemed to be a general agreement on
something that I felt was incredibly off-base.And that is the software
world is several years behind the hardware technology. BULL! I didn't
feel like saying anything then, but this misconception bothered me, and
my not pointing out this fallacy has bothered me even more so now. So
I'd like to correct this idea.
The truth of the matter is that as far as the microcomputer
world goes, the hardware technology is only now starting to catch up
with the software technology. What has blinded people to this is that
they are merely thinking in terms of yesterday's software technology;
which for microcomputers has been DOS 3.x.
The IBM AT benchmarks in at the low end of a VAX 750. The 386
has 3/4 the power of a VAX 8600 (as measured in dhrystones). Folks, the
software to run these types of machines has been around for years! It's
only been recently that microcomputers have had the power to run this
kind of software.
Of course the obvious case in point is UNIX. It has only been
with the advent of the AT that one has had a decent UNIX engine at a
relatively cheap price (i.e. less than 3 grand, including software).
But my arguement here goes beyond UNIX, it applies to other OS's that
are running on mini's and mainframes. When microcomputer's become as
capable as the larger machines all one has to do is to port that
software technology to the micro's. In term's of the current state of
the art (software and hardware) we have availible very sophisticated
tools *right now* for micro's.
(cont'd next message)
*** There are replies:
8192, 8220
#: 8192 S9/UNIX for the PC
21-Nov-86 20:19:21
Sb: #8190-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Dwight Leu [Microport] 76703,4266 (X)
Lee Feldenstein made a comment that I feel is going to leave him, and
so many other 386 developers out in the technological boonies. His
comment was something like "it's going to take 2-3 years to develop the
software that is going to run on (his) new hardware". If one stays in
the mental rut of 8086 land, he is absolutely correct; and in fact 2-3
years may be optimistic.
What he, and so many others don't realize is that the software
is availible *today*. Graphics, networking, applications are all
availible, or are being ported *now*. And anyone who doesn't make use
of this fact is going to get left way behind by one's competitors.
Well, it's time for me to get off of my soapbox on this.
Comments anyone?
-dwight-
*** There is a reply:
8204
*** Reading replies to 8085 ***
*** More ***
#: 8204 S9/UNIX for the PC
21-Nov-86 21:55:58
Sb: #8192-#Special Topic Intro.
Fm: Ray Duncan [DDJFORUM] 76703,4265
To: Dwight Leu [Microport] 76703,4266 (X)
Yeah, I have a comment. After using five different implementations of
UNIX, all I have to say is I can't imagine any more inappropriate
operating system for a personal computer than UNIX.
*** There are replies:
8216, 8226
*** Reading replies to 8085 ***
*** More ***
#: 8216 S9/UNIX for the PC
22-Nov-86 07:46:07
Sb: #8204-#Special Topic Intro.
Fm: Jim Foy 75735,1505
To: Ray Duncan [DDJFORUM] 76703,4265 (X)
ANd I can't believe that any serious multiuser application could be put
onto a micro without UNIX. True, it's not real friendly to the
programmer--he must be able to understand a fairly complex set of
commands. But the tradeoff is that he can design a very friendly
application for an unsophisticated user. OS's such as MS-DOS are
light-years behind UNIX.
*** There are replies:
8221, 8244
*** Reading replies to 8085 ***
*** More ***
#: 8221 S9/UNIX for the PC
22-Nov-86 10:37:38
Sb: #8216-#Special Topic Intro.
Fm: Bob Peterson 76703,532
To: Jim Foy 75735,1505 (X)
Jim,
Well, let's see. The following are _multiuser_ microcomputer
operating systems that are _not_ Unix or Unix clones, or Unix-like:
MBOS (A COBOL-oriented system)
PDOS (A real-time system)
RM/COS (Another COBOL-oriented system)
Pick (Database oriented)
Mirage (Initially to support APL, now with BASIC, FORTRAN, and
Pascal)
I can _easily_ imagine doing a multiuser microcomputer application
without Unix!!! Now, if you want multitasking to support multiple
terminals running a single application, the list of non-Unix support
software grows significantly.
Don't fall into the trap of assuming that Unix, the most common
multiuser OS on micros, is the only (or best) of its type. Also don't
forget that Intel isn't the only company whose cpu products are
incorporated into personal computers.
*** There is a reply:
8236
*** Reading replies to 8085 ***
*** More ***
#: 8236 S9/UNIX for the PC
22-Nov-86 15:31:20
Sb: #8221-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Bob Peterson 76703,532 (X)
I hope that I haven't started a UNIX vs other OS's donnybrook; and the
same applies to the Intel hardware. I obviously have my own beliefs and
biases in both of these areas. I was using UNIX and the 386 as a
specific example of how software technology is *ahead* of hardware
technology; and I believe it will be for the foreseeable future. As
microcomputers get more sophisticated, they can use the software
technology currently running on larger systems. And that it doesn't
make sense to limit one's self to yesterday's tools. Granted, they'll
probably be around forever (look at Fortran!). But to limit oneself
like this will prove to be a significant hinderance if you're serious
about market penetration using new hardware technology, whatever your
product is.
What I love about the hi-tech arena is that the rules change
every three years, when new technology is brought out. What I object to
is people who insist on using old tools and ideas on new systems. I
claim that developers who stick to these won't stand a chance in the
long run. This just seems incredibly obvious to me, and I believe that
history has shown it over and over again. Yet there seems to be a
certain reluctance to embrace new capabilites. Case in point being DOS
5.0; a horrible example of trying to teach an old dog new tricks. I
would like to see people in this industry being innovative, and bold
enough to come up with new approaches and ideas. Yet this is woefully
lacking today. Why? What does it take to get people using new concepts?
Is it simply cost-effectiveness? Perhaps marketing? I really don't
know; and am trying to find out by stirring up some comments on our
industry's muddling along, and what I perceive to be misguided notions
among people who should be leading, and not sticking to comfortable old
notions.
Oh well, I'm climbing down off of my soapbox. Is there anyone
out there who sees that I'm missing something, or am off-base on this?
I'd really like to know.
-dwight-
*** There are replies:
8238, 8247, 8260
*** Reading replies to 8085 ***
*** More ***
#: 8238 S9/UNIX for the PC
22-Nov-86 17:36:59
Sb: #8236-#Special Topic Intro.
Fm: Chris Dunford [IBMNET] 76703,2002
To: Dwight Leu [Microport] 76703,4266 (X)
I'm sorry, have you seen DOS 5.0? I didn't realize it was available
yet, for that kind of judgement.
*** There is a reply:
8295
*** Reading replies to 8085 ***
*** More ***
#: 8295 S9/UNIX for the PC
24-Nov-86 22:54:40
Sb: #8238-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Chris Dunford [IBMNET] 76703,2002 (X)
I have some very good contacts with what's happening with DOS 5.0. What
leads me to believe in their information is that it's consistant from
different sources. It's definately not availible yet, and of course the
final judgement will have to await it's appearance. But if I understand
the thinking up at Microsoft, I'd be willing to bet that the only thing
that will carry it is a huge marketing effort, and not it's technical
merits.
-dwight-
*** There is a reply:
8304
*** Reading replies to 8085 ***
*** More ***
#: 8304 S9/UNIX for the PC
25-Nov-86 17:20:07
Sb: #8295-#Special Topic Intro.
Fm: Lee Penn 70140,274
To: Dwight Leu [Microport] 76703,4266
I feel that the force of DOS has been carried by those many programs
that have been written that require that the user only be able to fine
the reset button. The average USER USER does not want to even know
about the Op sys. XENIX/UNIX et al is definitely not for this person.
For me, the UNIX environment is nervana. The multiscreen function with
SCO (and UNIX 5) makes this machine absolutely vital to my business.
AND I do have spreadsheet/word processing/ and file processing all
available at a quick flick of the ALT Fkey. Dos seem painfully archaic
whenever I user. UNIX on MICROS FOREVER!
*** There is a reply:
8307
*** Reading replies to 8085 ***
*** More ***
#: 8307 S9/UNIX for the PC
25-Nov-86 18:02:16
Sb: #8304-#Special Topic Intro.
Fm: Bill Hersh 73117,3320
To: Lee Penn 70140,274
Lee:
I agree with you that us programming types have no problems with
crytpic operating systems that need to be mastered to be useful. But if
computer use ever reaches the point that we dream of, with machines on
almost every desk, then we will be a definite minority, with the
majority of people wanting to use those easy-to-use programs and
operating systems.
Bill
*** There is a reply:
8312
*** Reading replies to 8085 ***
*** More ***
#: 8312 S9/UNIX for the PC
25-Nov-86 20:26:12
Sb: #8307-#Special Topic Intro.
Fm: Jim Foy 75735,1505
To: Bill Hersh 73117,3320 (X)
The main thing missing from UNIX is, as you say, ease of use. I
suspect that we will ultimately see another shell--one that is less
cryptic for the user. I guess Microsoft tried this approch with vsh,
but it is not the answer. Various applications come with a user shell
that handles the most important jobs, but fame and glory await the guru
that developes one for all purposes.
*** There is a reply:
8324
*** Reading replies to 8085 ***
*** More ***
#: 8324 S9/UNIX for the PC
26-Nov-86 10:23:34
Sb: #8312-Special Topic Intro.
Fm: Bryan Headley/Tandy 70007,2060
To: Jim Foy 75735,1505
Jim - I personally like Telexpress's XMENU.
Bryan
*** Reading replies to 8085 ***
*** More ***
#: 8247 S9/UNIX for the PC
22-Nov-86 22:41:29
Sb: #8236-#Special Topic Intro.
Fm: Ray Duncan [DDJFORUM] 76703,4265
To: Dwight Leu [Microport] 76703,4266 (X)
I'm surprised at your comment on DOS 5.0. People who know something
about the system aren't allowed to discuss it. So are you just
guessing, or violating your nondisclosure agreements?
*** There is a reply:
8296
*** Reading replies to 8085 ***
*** More ***
#: 8296 S9/UNIX for the PC
24-Nov-86 22:59:05
Sb: #8247-Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Ray Duncan [DDJFORUM] 76703,4265
My beliefs on DOS 5.0 come from different sources associated with
Microsoft. What leads me to believe their information is that they
confirm the approach that Microsoft is taking. So I guess that you
could say that I'm not guessing, or violating my nondisclosure
agreement (as I have none in this area), but from violating other
people's non-disclosure agreements <grin>. But of course *I* would
never say this!
-dwight-
*** Reading replies to 8085 ***
*** More ***
#: 8260 S9/UNIX for the PC
23-Nov-86 12:41:59
Sb: #8236-#Special Topic Intro.
Fm: Bob Peterson 76703,532
To: Dwight Leu [Microport] 76703,4266 (X)
Dwight,
It is not my intention to fuel a "my OS is better than yours"
discussion. Such religious wars are not productive. I don't know of
an OS that meets what I'd like to have! And that includes Mac's
Finder, Unix (on anyone's box, with anyone's extensions), MS-DOS, or
even what I'm using right now, the UCSD p-System.
Your statements about wishing to use new tools and your (apparent)
position that Unix is such a tool are a very interesting contrast.
I also don't believe that moving software from minicomputers to
personal computers is (generally) the answer to the personal computer
users' needs. There are, of course, exceptions, ie, situations when
running Unix on a personal computer is exactly the right solution.
As I noted above, our opinions of Unix' position with respect to
"state of the art" operating systems differ. And that's OK. I have no
urgent need to convert you to my view, only to state my opinion for the
benefit of discussion.
I'd like to see more powerful, more sophiscated personal computers be
used to support not minicomputer system software, but system tools that
are appropriate to the work a personal computer user is doing. This
implies, to me, software that goes far beyond the capabilities present
in most minicomputer operating systems, perhaps even to the point of
the operating system as a distinct piece of software disappearing.
I certainly agree that taking a narrow view of how personal computers
should be operated, by limiting one's view to existing tools and
methods, is the road to minimum performance, and a backward-looking
position to take. On the other hand, I don't expect to see 1-2-3
running in native mode on a 386 for at least a year!
I have enjoyed this week. I hope you've found enough value here to
get a Microport account on Compuserve.
Here's my .signature file:
Hardcopy and Electronic Addresses:
Bob Peterson Usenet: ...!ut-sally!im4u!ti-csl!peterson
P.O. Box 1686 Compuserve: 76703,532
Plano, Tx 75074 (214) 995-6080 (work) or (214) 596-3720 (ans.
machine)
*** There are replies:
8261, 8297
*** Reading replies to 8085 ***
*** More ***
#: 8261 S9/UNIX for the PC
23-Nov-86 13:03:53
Sb: #8260-Special Topic Intro.
Fm: Levi Thomas (Sysop) 76703,4060
To: Bob Peterson 76703,532 (X)
Dwight will be with us on his current account for another week so we
can keep this fire going for a bit longer.
--Levi
*** Reading replies to 8085 ***
*** More ***
#: 8297 S9/UNIX for the PC
24-Nov-86 23:30:11
Sb: #8260-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Bob Peterson 76703,532 (X)
There are several interesting points that you bring up here. First, I'd
agree that a number of the ideas offered would be desirable. But what
is the quickest way to implement them? I believe that this can be done
by making use of the software technology that is used to exploiting the
power of sophisticated hardware, rather than trying to mimic this
capability with more limited technology.
For this reason, I disagree completely with the notion that
software technology is behind hardware technology for micros. I've been
rather disappointed that no one has disagreed with me; there seemed to
be so much agreement on this at the hackers conference (where's Lee
Felsenstein when I need him <grin>!).
Also, I'm not sure that you'll ever see 1-2-3 running in native
mode on the 386, except under DOS 3.x. To do so would be to ignore the
capabilities of the 386. In Virtual Mode, though, is another matter.
It's currently running on our implementation of DOS-MERGE, and on
multiple serial consoles.
Finally, it looks like you'll have to put up with me for
another week. Levi was kind of enough to let me extend my access to
CIS. Thanks Levi! I'll probably be off though, from Thursday through
Saturday, and returning Sunday.
-dwight-
*** There is a reply:
8309
*** Reading replies to 8085 ***
*** More ***
#: 8309 S9/UNIX for the PC
25-Nov-86 19:58:06
Sb: #8297-Special Topic Intro.
Fm: Bob Peterson 76703,532
To: Dwight Leu [Microport] 76703,4266
Dwight,
Part of my position is that minicomputer software doesn't take
advantage of microcomputer hardware! Specifically, how much
minicomputer (or larger) software works with bit-mapped and/or color
dislays? What I see happening is that such capabilities are utilized
by various vendors grafting something onto a minicomputer software
base.
Heck, Unix doesn't even do a decent job of utilizing a video display
terminal! The command interface assumptions with respect to hardware
haven't really changed since Unix began! (I'm talking here about Unix
as delivered by AT&T or Berkley.) No standard shell (sh csh ksh) makes
decent use of a VDT. The user interface assumes nothing more than a
TTY-33!!! Hardware technology made that sort of interface obsolete
years ago.
Sun has made some progress in retrofitting utilization of a large,
bitmapped screen to a Unix core. But the basic assumptions of the
TTY-33 are still there. And, at this point, Sun is just beginning to
distribute their stuff to the general marketplace, ie, other vendors'
hardware.
For example, why are Unix pipes still linear? I'd like to take the
output of the "tee" command and connect _BOTH_ output streams to two
other standard inputs!
I've been trying to disagree with your assertion that personal
computer software technology is ahead of personal computer hardware. I
find most personal computer operating system pretty bad, mostly because
they are modeled on minicomputer systems not designed as single user,
multitasking systems.
Why shouldn't I expect 1-2-3 to run in native mode on a '386? Didn't
Lotus admitt that 640K isn't enough when they changed 1-2-3 to access
"extended" memory? Doesn't this imply that 1-2-3 _should_ be running
in an environment that doesn't require a kludge? And isn't the '386
linear addressing just such an environment?
*** Reading replies to 8085 ***
*** More ***
#: 8244 S9/UNIX for the PC
22-Nov-86 22:23:53
Sb: #8216-#Special Topic Intro.
Fm: Ray Duncan [DDJFORUM] 76703,4265
To: Jim Foy 75735,1505 (X)
Exactly my point, a multiuser operating system has no business on a
personal computer.
*** There is a reply:
8298
*** Reading replies to 8085 ***
*** More ***
#: 8298 S9/UNIX for the PC
25-Nov-86 00:05:25
Sb: #8244-#Special Topic Intro.
Fm: Dwight Leu [Microport] 76703,4266
To: Ray Duncan [DDJFORUM] 76703,4265
"A multiuser OS has no business on a personal computer"?? At least I've
managed to touch someone's nerve <grin>! I suppose one could go down
the drain in various technical debates on this. And personally, it
strikes me as rather dogmatic. So let me give you one extremely
important reason why the single-user OS is a thing of the past (not
that *I'm* dogmatic :-)).
Take a look at simple economics, that driving forcce of the
marketplace. In my opinion, there is no greater waste of money than a
bunch of AT's running DOS. The simple fact is that for around $800 *per
user* you can automate an entire office! To my knowledge, there is no
other alternative which is as cost effective as an AT running 9 users
on System V/AT. A dealer can get an 8-10 Mhz AT for around $1K, 80 Mb
(23 ms access time) Toshiba for around another $1K, 3.5 Mb Ram for less
than $500. A multiport board (8 users) will run less than $500. And
let's say a mono card and monitor for about another $500. Throw in 8
terminals for around $300 each. With our complete software package,
that brings the price up to less than $6.5K. For 9 users, that's around
$725 per user (and my price estimates are slightly high). And there's
no networking needed!
I defy *anyone* to show me a more cost effective solution!
Given this type of price/performance ration, the single user OS is
going to have a great deal of trouble competing in the long term.
-dwight-
*** There are replies:
8302, 8310, 8322
*** Reading replies to 8085 ***
*** More ***
#: 8302 S9/UNIX for the PC
25-Nov-86 05:57:59
Sb: #8298-Special Topic Intro.
Fm: Chuck Forsberg Omen Tech 70007,2304
To: Dwight Leu [Microport] 76703,4266
Where's a good place to get that price on the Tosh HD and the 8 port
serial board? 8 ports is just about right for one power user :-)
*** Reading replies to 8085 ***
*** More ***
#: 8310 S9/UNIX for the PC
25-Nov-86 19:59:01
Sb: #8298-#Special Topic Intro.
Fm: Bob Peterson 76703,532
To: Dwight Leu [Microport] 76703,4266
Dwight,
"So let me give you one extremely important reason why the
single-user OS is a thing of the past ..." Your reasons are all well
and good for those environments that are I/O bound at the mass storage
interface.
However, there _are_ applications that are compute bound, either
number crunching or at the user interface. In this sort of environment
the best solution I'm aware of is diskless workstations accessing a
central file server via a high-speed link. (A local area network is
_one_ example of such a link.)
I know of an environment of around 100 personal computers with color
graphics. Users interact with the displays, using significant amounts
of cpu time to update the screen _and_ to check the updates for
legality. The displays are required to respond within a few tenths of a
second. The database lives on an IBM mainframe. At various points in
the process, the workstation exchanges data with the mainframe at high
speed. "High speed" means the mainframe transaction should complete in
0.4 seconds or faster. Such an environment could _not_ be economically
implemented using a timeshared host because of the computational
requirements. The personal computers run MS-DOS! The personal
computers _replaced_ IBM color terminals in order to improve response
time.
These users, and their management, do not consider the machines
running MS-DOS to be a waste of money! I assert that even when scaled
down to 8 workstations, a single central machine will not provide
adequate cpu power. To be more specific, would you be prepared to force
8 engineers to share a single CPU to run AutoCAD??? Most AutoCAD users
feel a PC is too slow, much less 1/8 of an AT!
*** There is a reply:
8314
*** Reading replies to 8085 ***
*** More ***
#: 8314 S9/UNIX for the PC
25-Nov-86 21:25:46
Sb: #8310-#Special Topic Intro.
Fm: tom genereaux 76470,32
To: Bob Peterson 76703,532 (X)
I use a microvax II as a personal workstation for some of the things
that I do - and it's too slow! A small Cray should be just about
right...
Tom G.
*** There is a reply:
8325
*** Reading replies to 8085 ***
*** More ***
#: 8325 S9/UNIX for the PC
26-Nov-86 10:24:39
Sb: #8314-Special Topic Intro.
Fm: Bryan Headley/Tandy 70007,2060
To: tom genereaux 76470,32
Tom - There you are! My care package never arrived... (How's Germany
nowadays?)
Bryan
*** Reading replies to 8085 ***
*** More ***
#: 8322 S9/UNIX for the PC
26-Nov-86 06:20:22
Sb: #8298-Special Topic Intro.
Fm: Russ Ranshaw (Wiz-10) 70003,3076
To: Dwight Leu [Microport] 76703,4266
But the multi-user system is definately NOT a PERSONAL Computer! The
idea behind the PC is that each individual has a system unto
him-/her-self. This allows each individual to have an environment that
is potentially configurable for specific needs. Admitedly there are
situations where multi-TASKING is desirable (I've not encountered on
with my PC), but multi-USER is not necessary. The down-side is that in
order to use a multi-user system, multiple terminals are required. And
terminals are not free. For not significantly more $$$ than a good
terminal one can have a lot of PC power. For those applications which
require multi-user access (accounting, engineering, other large data
bases), an individual's PC can also have access to a central system,
either through a multi-user system or a LAN. And what about the fact
that central computers take down an entire operation when they fail? I
thought the move to PC's on every desk was a step in the right
direction.
Lest the facts get lost, I have over 26 years of experience in the
computer field. I cut my computer teeth on an IBM 650, and my intimate
use of computers covers a range of some 10 or 12 large systems. I have
been a proponent of time sharing for many years. I don't believe the
need for multi-user systems will vanish. But I am also a supporter of
PC's on every desk. In our environment (CIS software development),
most programmers have some kind of PC. We typically do our editing
locally, upload the CHANGES to the files, and compiler on the
mainframe. This has cut down the mainframe requirements significantly.
Editing on-line takes a lot of resources, and the capabilities can
never match the speed/efficiency of local editing on the PC, where
everyone can use the editor of his/her choice.
Let's not be myopic and try to keep a larger perspective.
*** Reading replies to 8085 ***
*** More ***
#: 8226 S9/UNIX for the PC
22-Nov-86 10:43:25
Sb: #8204-#Special Topic Intro.
Fm: Bill Hersh 73117,3320
To: Ray Duncan [DDJFORUM] 76703,4265 (X)
Ray:
I love your "Advanced MS-DOS" book, and agree that it is a superb
reference for MS-DOS. Got a slightly more personal question. There is
a small blurb at the back of the book stating you are a physician,
speicalizing in neonatology. Are you still practicing medicine? I know
you are in California; have you ever hooked up with Shortliffe and his
group at Stanford or Blois and his group at UCSF? Are you interested
in computer applications specific to medicine?
I myself am in my last year of Internal Medicine residency at
University of Illinois Hospital in Chicago. I am, however, quite
interested in computer applications in clinical medicine, and will most
likely be with Blois at UCSF in their Medical Informatics Fellowship
next summer.
Bill
*** There is a reply:
8246
*** Reading replies to 8085 ***
*** More ***
#: 8246 S9/UNIX for the PC
22-Nov-86 22:32:47
Sb: #8226-#Special Topic Intro.
Fm: Ray Duncan [DDJFORUM] 76703,4265
To: Bill Hersh 73117,3320 (X)
I still practice but the amount varies, I cover vacations for some
neonatologists in this area and also work weekends and cover nights for
them. My fellowship days were before the days of official Medical
Informatics fellowships & suchlike. But I did spend about half of my
residency and fellowship implementing a data management/on-line
reference/physician assistant system for the Neonatal Intensive Care
unit at Cedars-Sinai in LA. It was based on a Z-80 system with a hard
disk running MP/M II and served 4 terminals and a Printronix 300 LPM
printer. Was a lot of fun and was used around the clock by the
residents and neonatology staff for about 4 years. Since I left there
the software has been ported to a VAX I understand! (they needed more
computing power to acquire data frotient monitors in realtime etc.)
*** There is a reply:
8253
*** Reading replies to 8085 ***
*** More ***
#: 8253 S9/UNIX for the PC
23-Nov-86 05:41:59
Sb: #8246-#Special Topic Intro.
Fm: Bill Hersh 73117,3320
To: Ray Duncan [DDJFORUM] 76703,4265 (X)
Ray:
Interesting. Are you working on any medical computing projects
currently, or do you plan to?
Bill
*** There is a reply:
8272
*** Reading replies to 8085 ***
*** More ***
#: 8272 S9/UNIX for the PC
23-Nov-86 20:55:24
Sb: #8253-Special Topic Intro.
Fm: Ray Duncan [DDJFORUM] 76703,4265
To: Bill Hersh 73117,3320 (X)
We have done quite a bit of work on an NICU data management system
based on IBM PC/ATs and Ethernet. However, I don't know when and if we
will ever try to market it.