AMIX?
Michael Nowak
Lamplighter Software. According to the ad, AMIX is:
* Unix System V compatible
* A multi-user, multi-tasking Unix workalike
* Compatible with Amiga-DOS file structure
* Access Amiga-DOS special functions such as sound and graphics
I called them up and according to the woman on the phone, it is two weeks
from shipping, will cost $395, and will allow Amiga applications to run under
AMIX. They are sending me more information.
Does anyone else have any information on this? It sounds like an interesting
alternative.
------------------------------------------------------------------------------
In Real Life: Michael Nowak "Seek truth from facts."
Via Internet: mi...@ronin.cc.umich.edu - Deng Xiao Ping
Via UUCP: uunet!umix!ronin.cc.umich.edu!mike
Working for but in no way representing the University of Michigan.
------------------------------------------------------------------------------
Michael J. Farren
|In Amazing Computing, Vol 3. Num 4., there is an ad for AMIX from
|Lamplighter Software. According to the ad, AMIX is:
|
| * Unix System V compatible
| * A multi-user, multi-tasking Unix workalike
| * Compatible with Amiga-DOS file structure
| * Access Amiga-DOS special functions such as sound and graphics
|
|I called them up and according to the woman on the phone, it is two weeks
|from shipping, will cost $395, and will allow Amiga applications to run under
|AMIX. They are sending me more information.
|Does anyone else have any information on this? It sounds like an interesting
|alternative.
Two, out of a thousand, questions: 1) How close to a real Unix is it?
Does it provide all of Unix's functionality? All of Unix's utilities?
And 2) Would you buy a Unix from someone who couldn't support it? I
mean REALLY support it? I wouldn't. Not in a billion years. I'd
suggest waiting for C/A's own Unix, since it's already been announced.
Might cost you more, but I would think there's a better chance that
it'll work right, and do what you want.
--
Michael J. Farren | "INVESTIGATE your point of view, don't just
{ucbvax, uunet, hoptoad}! | dogmatize it! Reflect on it and re-evaluate
unisoft!gethen!farren | it. You may want to change your mind someday."
gethen!far...@lll-winken.llnl.gov ----- Tom Reingold, from alt.flame
Morgan W. Jones
>suggest waiting for C/A's own Unix, since it's already been announced.
>Might cost you more, but I would think there's a better chance that
>it'll work right, and do what you want.
That's what I was planning to do, but hmm...
A2000 $1500
A2620 2500 (CSA=1000, 551+4M 32bit static=1500)
A2420 550
A2090A 350
100M 1100
U**x 800
-----
$6800 U.S. = $9000+ Can.
Unless someone convinces me that I'm VERY wrong, I somehow can't see
myself getting this! And it makes me sad because I like the Amiga but
also feel it necessary to get U**x.
>Michael J. Farren | "INVESTIGATE your point of view, don't just
--
Morgan Jones - Bramalea Software Inc. mor...@brambo.UUCP
...!{uunet!mnetor!lsuc!ncrcan, utgpu!telly}!brambo!morgan
"These might not even be my opinions, let alone anyone else's."
Chuck McManis
-> [Amiga UNIX box] == $6800 U.S. = $9000+ Can.
->
->Unless someone convinces me that I'm VERY wrong, I somehow can't see
->myself getting this! And it makes me sad because I like the Amiga but
->also feel it necessary to get U**x.
->Morgan Jones - Bramalea Software Inc. mor...@brambo.UUCP
You are not wrong Morgan. Something to think about, is that people who
make workstations such as Sun are constantly working down the price curve
and trying to make the whizziest system for the least money. IF IT WAS
POSSIBLE TO MAKE A DISKFULL UNIX WORKSTATION FOR < $8K TODAY WE WOULD!
We want to make one because we know people would buy them, but that is
like saying if someone would just make a flying car, rush hour traffic
wouldn't be a problem. Now both of these problems will be solved one
day but that day is not today so all you can do is wait until your
disposable income intersects the cost of a UNIX workstation. I would
be nice but it isn't possible yet. Just like it isn't possible to build
a Cray for less than $100,000. Those would sell too.
--Chuck McManis
uucp: {anywhere}!sun!cmcmanis BIX: cmcmanis ARPAnet: cmcm...@sun.com
These opinions are my own and no one elses, but you knew that didn't you.
Morgan W. Jones
>IF IT WAS
>POSSIBLE TO MAKE A DISKFULL UNIX WORKSTATION FOR < $8K TODAY WE WOULD!
For under $4000 (this is accurate as far as I know) you could run a
386 clone with full SysV. Thus, for less than the cost of upgrading
you could run a concurrent SysV and use the 2000 as a terminal, or as
many terminals when DNET gets ported. That's the reason that I
thought I could be wrong.
>--Chuck McManis
--
Morgan Jones - Bramalea Software Inc. mor...@brambo.UUCP
Chuck McManis
> IF IT WAS POSSIBLE TO MAKE A DISKFULL UNIX WORKSTATION FOR < $8K
Sorry to shout like that.
In article <3...@brambo.UUCP> (Morgan W. Jones) replied:
> For under $4000 (this is accurate as far as I know) you could run a
> 386 clone with full SysV.
Ok folks, I may get flamed for this but hey what's life for anyway.
The problem comes when we get down to brass tacks and define what
a "workstation" is. In my lexicon a workstation has 4 megabytes of
memory, 1M pixels (optionally in color), 80Meg+ of disk space, UNIX*
in some form or another, and a window system in system in some form or
another. Ideally, it also has built in networking.
Intel took the initiative to build this exact beast, given the parts
they made and some of the parts that others made, and came up with a
'386 box with 4 meg of ram, running System V and X11, with an 80Meg
drive, and a Blit board and nice 1K x 1K monochrome monitor. Guess
what? It cost *them* $8,192 , you and me it would cost $11,264 - $12,288
dollars. Sure, you can run UNIX on an AT clone with probably less memory
and an EGA card, but it wouldn't be a workstation in my book. So the
next time someone argues that they can build a workstation for cheap
try meeting those design specs, and see if it is still cheap. This
other thing you have to keep in mind is that it *must* be available
*now*, no fair saying "well, by december this 80Meg drive will only be
$200." ok?
Jerry Whitnell
>In article <49...@sun.uucp> I wrote :
>> For under $4000 (this is accurate as far as I know) you could run a
>> 386 clone with full SysV.
>
>Ok folks, I may get flamed for this but hey what's life for anyway.
>The problem comes when we get down to brass tacks and define what
>a "workstation" is. In my lexicon a workstation has 4 megabytes of
>memory, 1M pixels (optionally in color), 80Meg+ of disk space, UNIX*
>in some form or another, and a window system in system in some form or
>another. Ideally, it also has built in networking.
>
>Intel took the initiative to build this exact beast, given the parts
>they made and some of the parts that others made, and came up with a
>'386 box with 4 meg of ram, running System V and X11, with an 80Meg
>drive, and a Blit board and nice 1K x 1K monochrome monitor. Guess
>what? It cost *them* $8,192 , you and me it would cost $11,264 - $12,288
>dollars.
Intel should shoot their parts suppliers (and somebody should should Intel :-).
Just reading adds in MicroTimes I come up with:
21 mhz 386 clone w/ 1mb $1,995
19" Mono $1,895
80 MB hard disk $ 839
3 mb of memory $ 600 (approximate quess as noone quotes prices)
System V $ 799 (Microport from the Programmer's connection)
X/11 Free
------
Total $6,128
I just saved $2,000 without really trying. Of course I don't get a single
orginization like sun supporting me, but it might be worth $2000 to take
my chances. If you are willing to settle for 800x600 you can drop over
$1000 from the price of the monitor. All prices except Microport come
from the March 88 issue of MicroTimes. Of course, I havn't put this
beast together to see if it all flies together... :-).
>--Chuck McManis
Jerry Whitnell Been through Hell?
Communication Solutions, Inc. What did you bring back for me?
- A. Brilliant
eric townsend
>IF IT WAS
>POSSIBLE TO MAKE A DISKFULL UNIX WORKSTATION FOR < $8K TODAY WE WOULD!
So I guess my AT&T 3b1 w/ Unix Sys V 3.0 (3.51 now available) w/ 3.5Mb RAM,
67Mb HD, etc etc isn't a unix workstation? At my last place of work,
we developed C code for both unix and dos on 3b1's. Think about it.
Developing on a semi-kludge like the 3b1 for a MS-DOS abortion. Poetic,
somehow.
Seriously, with 3b1's as low as they are, I think a great buy would be
an Amiga 500 and a full-blown 3b1 with Sys V 3.51; and just use
the Amiga as a multi-tasking games machine/graphics terminal to the 3b1.
I'll probably get flamed for *that* suggestion. :-)
> >--Chuck McManis
--
Anytime a computer company needs the entire cast of "MASH" to sell it's
newest product line; it's time to take a serious, critical look at that
product line. -- Me.
J. Eric Townsend ->uunet!nuchat!flatline!erict smail:511Parker#2,Hstn,Tx,77007
Peter da Silva
> Ok folks, I may get flamed for this but hey what's life for anyway.
No flames, just a bit of perspective.
> The problem comes when we get down to brass tacks and define what
> a "workstation" is. In my lexicon a workstation has 4 megabytes of
> memory, 1M pixels (optionally in color), 80Meg+ of disk space, UNIX*
> in some form or another, and a window system in system in some form or
> another. Ideally, it also has built in networking.
That's an interesting definition. Let's take the quantitative stuff out
of that and see what we get...
Memory... well, 4 megabytes is nice. I have that much in my Amiga. But
it's not necessary. Let's say you need enough memory that you can run the
rest of the stuff without swapping. Carnegie-Mellon specs 1 Meg for a
workstation, by the way.
Pixels... well, a megapixel display is nice. Part of the 3M definition.
But do you really need that? Unless you're doing CAD work you really only
need enough to fit a couple of "terminal-size" windows on the screen. With
ColorFonts you can even use antialiasing and get nice looking text.
Disk space... well, enough is enough. Let's say you need twice the disk
space that the basic operating system takes up, with a minimum of ten
megs of user space. So this depends again on another parameter: how big
the O/S is.
UNIX in some form of another... I'm tempted to cut this requirement, but
you'll hang me for it. So, how small can UNIX be and still be UNIX? Well,
you need ALL the binaries. The SCO-Xenix distribution for the IBM-PC, with
the manuals, is 23 320K diskettes. That's over 7 megabytes, and makes a
good baseline. No windowing, but the Amiga ROMS have all the layer
manager code in them. A windowing system that takes advantage of them
could be pretty small. Let's be gross, though, and say you need 3 meg
of disk space for the WM and utilities. a 20 Meg disk would do, at a
minimum. But let's double that and say 40 Meg. Sun's sold workstations
with that much disk, anyway.
Windowing... The Amiga has windowing. The Mac has windowing (echoes of
Jonathan Livingston Seagull: "Just flying? A mosquito can do that much!").
You don't need megabytes of RAM or Disk for windowing.
Networking... that's a toughy. But do you really need it for a home
workstation?
Surely you can come up with a machine that uses plain multisync monitors
and runs or under 4 Grand. I suspect that a 2500-type system could do it.
A multisync monitor and flckerFixer instead of the Hedley monitor, and
a less ambitious UNIX than full SVR2.
> Sure, you can run UNIX on an AT clone with probably less memory
> and an EGA card, but it wouldn't be a workstation in my book.
Not everyone has tastes quite that rich. Try these design specs:
1 meg of memory, 1 million pixels resolution, and 1 MIPS. The monitor
to display all those pixels is the hard part, but if you multiply
pixels by bitplanes and take advantage of antialiasing the Amiga
with a 2620 can do it. It's not *that* much of a cheat.
--
-- Peter da Silva `-_-' ...!hoptoad!academ!uhnix1!sugar!peter
-- "Have you hugged your U wolf today?" ...!bellcore!tness1!sugar!peter
-- Disclaimer: These aren't mere opinions, these are *values*.
Todd South
>In Amazing Computing, Vol 3. Num 4., there is an ad for AMIX from
>Lamplighter Software. According to the ad, AMIX is:
>
> * Unix System V compatible
> * A multi-user, multi-tasking Unix workalike
> * Compatible with Amiga-DOS file structure
> * Access Amiga-DOS special functions such as sound and graphics
>
>I called them up and according to the woman on the phone, it is two weeks
>from shipping, will cost $395, and will allow Amiga applications to run under
>AMIX. They are sending me more information.
When you get more information I hope that you will readily post
it to the net. It seems amazing that someone is finally coming out with
Unix on a machine that can truly handle it in the personal market.
Todd South
:-----------------------------------------------------------------------:
| Todd South : Ewa Beach, HI ||| Pacific Proline: (808) 499-2831 2400bd |
| Uucp: {nosc, ihnp4, cacilj, sdcsvax, hplabs!hp-sdd, sun!ihnp4} |
| ...!crash!pnet01!pro-simasd!pro-pac!tsouth |
| ARPA: crash!pnet01!pro-simasd!pro-pac!tso...@nosc.MIL |
| INET: tso...@pro-pac.CTS.COM BITNET: psuvax1!tso...@pro-pac.CTS.COM |
:-----------------------------------------------------------------------:
Eric Green
> In article <49...@sun.uucp> cmcm...@sun.UUCP (Chuck McManis) writes:
>>IF IT WAS
>>POSSIBLE TO MAKE A DISKFULL UNIX WORKSTATION FOR < $8K TODAY WE WOULD!
>
> For under $4000 (this is accurate as far as I know) you could run a
> 386 clone with full SysV. Thus, for less than the cost of upgrading
A '386 clone costs around $4000 for a Unix system, true. But I suspect that
what drives the price of the Sun, etc., upward, is GRAPHICS support, and
NETWORKING support. Networking's a bit cheaper nowdays than it used to be. But
hi-res monitors still are dear. Not to mention that the Sun has faster i/o
than the '386 clone machine does (16 bit bus, blah!).
Add in $2,000 for a professional graphics display, and you're up to $6,000 --
Sun country. And the Ethernet board won't be cheap, either.
But it's an exercise in absurdity. You don't buy a '386 for graphics. People
who run Unix on a'386 are generally doing it so that they don't have to use an
overloaded Vax-780 with 20 simultaneous users running Emacs and Common Lisp.
Quite a different market, compared to the graphics workstation market.
--
Eric Lee Green e...@usl.CSNET Snail Mail P.O. Box 92191
ihnp4!killer!elg Lafayette, LA 70509
"Is a dream a lie that don't come true, or is it something worse?"
SBCS Systems Staff
estimate is a bit low ball, but workable. This is our rough
requirement for a next generation workstation a SUNY/Stony Brook:
10-20 mips main processor
> 10M SP Whetstones FP performance
8 mBytes memory, must be expandable to 512 mBytes or greater
Medium resolution color screen, eg 1280x1024x8, must be
expandable to x24 and graphics rendering processor
FDDI networking, at least. At least 50 mbps end to end throughput.
Fast internal bus, eg 100 mBps or greater.
Real Unix. We're comfortable with 4.3, or SunOS 4.X. SysV not welcome.
We're looking at SGI, Apollo PRISM, etc. Our low end requirement
is roughly Sun-3/60:
3 mips processor
no min FP requirement, '881/'882 preferred
4 mBytes memory, expandable to at least 16 mBytes
Medium res color screen, eg 1024x1024x8
Ethernet networking
Real Unix.
The CMU requirement is just that: it is what CMU will buy. We have
100+ Suns here, and I'll tell you that a 3/50 is pretty well threshold
of pain in current workstations. In such an environment, the current
Amiga makes an awesome terminal with some compute ability. It does
not make a workable Unix machine. At least for the sort of Unix
we run here.
Listen to Chuck. He is giving you the straight story. Real unix
machines cost bucks.
Rick Spanbauer
SUNY/Stony Brook
Chuck McManis
>So I guess my AT&T 3b1 w/ Unix Sys V 3.0 (3.51 now available) w/ 3.5Mb RAM,
>67Mb HD, etc etc isn't a unix workstation?
No, the 3b1 isn't a workstation. It is more of an AT clone class of machine.
Now I am not saying that is bad, however in my previous message I stated
that we were talking workstation and what that meant to me. Yes, you can
get a 3b1, yes you can get an AT clone, no you can't get a workstation.
DMast...@cup.portal.com
In message <12...@sbcs.sunysb.edu>, ro...@sbcs.sunysb.edu writes:
> Listen to Chuck. He is giving you the straight story. Real unix
> machines cost bucks.
>
believe me, there could be such a market)! Ever had the feeling that putting
Unix on PC/XT class machines was an idea slightly ahead of its time??
> Rick Spanbauer
> SUNY/Stony Brook
>
David Masterson
DMast...@cup.portal.com
Darin Johnson
In article <50...@sun.uucp>, cmcmanis%pep...@Sun.COM (Chuck McManis) writes:
>
> Ok folks, I may get flamed for this but hey what's life for anyway.
> The problem comes when we get down to brass tacks and define what
> a "workstation" is. In my lexicon a workstation has 4 megabytes of
> memory, 1M pixels (optionally in color), 80Meg+ of disk space, UNIX*
^^^^
> in some form or another, and a window system in system in some form or
^^ ^^^^ ^^^^ ^^ ^^^^^^^
> another. Ideally, it also has built in networking.
Hmmn, I wonder if DEC knows that they aren't selling workstations
after all :-)
Actually, my definition of a workstation would probably fall pretty
close to a Sun 3/75 (0 Meg disk space). Other Sun versions may have more
power/etc., but the 3/75 fits my requirements:
1) Sit on my desk (well it got used to be; imagine root not having his own
sun :-(
2) The cpu doesn't get bogged down by 'n' other users.
3) I don't have to worry about messing up other users when I
steal the CPU/crash-the-machine/etc.
4) I have full access to disks/cpu on the groups MAIN machine.
Actually, the Amiga would fit my specs if it were; in my office,
had DNET, maybe some more memory (possibly '020).
I can run lots of stuff on my Amiga, and then turn to one of the DNET windows
and run a cpu intensive job on another machine.
Actually, I would class workstations into several categories:
- Inexpensive computers to get a particular job done and to get
those darn administrators off of the VAX (MAC's IBM's)
- Computers to supply windowing/etc. to a larger computer, with
easy access to the larger computers files (and vice-versa).
(The amiga described above would fall into this class)
- Standalone workstations - what you describe above. Generally meant
for one user, although that never happens around here :-)
(includes specialized publishing only or CAD only types)
- Workstations that can be networked/coupled together. For example
Suns with NFS, X, etc. or vaxclusters.
- ? Distributed computers - if anyone actually gets completely
transparent distributed computers, would the individual nodes
(with graphics) be considered Workstations?
Then again, I may be confusing workstations with WorkStations(R)...
--
Darin Johnson (...lll-lcc.arpa!leadsv!laic!darin)
(...ucbvax!sun!sunncal!leadsv!laic!darin)
"All aboard the DOOMED express!"
Michael Portuesi
> *Excerpts from: 23-Apr-88 Re: AMIX? Peter da Si...@sugar.UUC (3262)*
> Memory... well, 4 megabytes is nice. I have that much in my Amiga. But
> it's not necessary. Let's say you need enough memory that you can run the
> rest of the stuff without swapping. Carnegie-Mellon specs 1 Meg for a
> workstation, by the way.
<giggle>
The Sun 3/50 I am typing this message on (running Unix and the Andrew software
environment) has about 4 MB of memory. The IBM RTs at Carnegie Mellon
generally have about 6 MB. Anything less and they would be totally useless.
When the Andrew environment at CMU officially moves to X for its window
manager, they'll be even more constrained since X is a memory pig, much more
than the window manager currently in use (developed at CMU for the Andrew
environment). When that happens, I expect the few Sun-2's around here to
disappear quickly, since they will no longer be able to keep up with what is
demanded of them. Basically, Carnegie Mellon is outright lying when they say 1
MB is sufficient memory for a workstation.
In the Computer Science department, which already runs Mach and X on its
workstations, 10 and 12 MB are common figures.
> Windowing... The Amiga has windowing. The Mac has windowing (echoes of
> Jonathan Livingston Seagull: "Just flying? A mosquito can do that much!").
> You don't need megabytes of RAM or Disk for windowing.
You do if you want to run X on your workstation.
> Networking... that's a toughy. But do you really need it for a home
> workstation?
I don't think Chuck was defining a "home" workstation...he was defining a
workstation. And networking is one of the most useful aspects of a workstation.
--M
Michael Portuesi / Carnegie Mellon University
ARPA/UUCP: mp...@andrew.cmu.edu BITNET: rainwalker@drycas
"Memories are uncertain friends, when recalled by messages" -- OMD, "Messages"
Kent Paul Dolan
Did you catch that? Were you paying attention?
No? Let's try again:
****************************************************************************
* Real Unix. We're comfortable with 4.3, or SunOS 4.X. SysV not welcome. *
****************************************************************************
Still no? One more try:
*******************
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*SysV not welcome.*
*******************
Now? Those are the paying customers speaking guys, and they are NOT
kidding around, and that is NOT an isolated opinion. Think it
through, please? It's your corporate survival, and my IRA retirement
plan. Do it right, OK?
Kent, the man from xanth.
David C. Albrecht
> In article <8...@gethen.UUCP> far...@gethen.UUCP (Michael J. Farren) writes:
> >suggest waiting for C/A's own Unix, since it's already been announced.
> >Might cost you more, but I would think there's a better chance that
> >it'll work right, and do what you want.
>
> That's what I was planning to do, but hmm...
>
> A2000 $1500
> A2620 2500 (CSA=1000, 551+4M 32bit static=1500)
> A2420 550
> A2090A 350
> 100M 1100
> U**x 800
> -----
> $6800 U.S. = $9000+ Can.
>
> Unless someone convinces me that I'm VERY wrong, I somehow can't see
> myself getting this! And it makes me sad because I like the Amiga but
> also feel it necessary to get U**x.
What exactly is your point? Yes, 68020s are expensive. Yes, 100
Megabyte disk drives are expensive. Yes 4M of 32bit static memory is
expensive. I don't know what the A2420 is.
Probably the cheapest high performance Unix boxes available at present
are 80386 clones. The hard disk controller is cheaper (probably $200 or so),
the box itself is cheaper (than the 2000+cpu card combo probably $1000 or so).
A configuration like you outlined in a clone box would probably be in the at
least $5k+ range and it just doesn't come any cheaper. To get much cheaper,
you are going to have to go to lower performance and smaller configurations.
I would expect to pay more for an Amiga based box as they simply don't have
the volume.
In any case, I should hope Unix will run in a smaller configuration than this.
Hey, on the AT&T 7300s it would run on a 1M, 20M machine with about 5M of disk
to spare.
Myself, fortunately I already have a 2000 and a 2090 and will only have to
stare at the sticker on the 2620 and the Unix and decide if they are worth it
to me.
David Albrecht
Lee Sailer
several remote machines simultaneously, and move info to and fro. This
is important to me because, in my work environment, I have to do this a lot.
Also, I see colleagues with less computer chutzpah who are less
productive because they cannot do this. (That is, in this heterogenous
environment, you have to know about 5 communications packages and
command languages to move stuff around.)
Now, in a local environment, with an ethernet board in the Amiga, it
could (I guess) do this. But being limited to one serial port is a problem.
lee
David Herron -- One of the vertebrae
>In article <50...@sun.uucp>, cmcmanis%pep...@Sun.COM (Chuck McManis) writes:
>> The problem comes when we get down to brass tacks and define what
>> a "workstation" is. In my lexicon a workstation has 4 megabytes of
>> memory, 1M pixels (optionally in color), 80Meg+ of disk space, UNIX*
>> in some form or another, and a window system in system in some form or
>> another. Ideally, it also has built in networking.
>it's not necessary. Let's say you need enough memory that you can run the
>rest of the stuff without swapping. Carnegie-Mellon specs 1 Meg for a
>workstation, by the way.
Well .. to give you an example of memory needs ... from memory, the
memory requirement for the Kyoto Common Lisp version of Macsyma is
about 4 megabytes. That's before you try to run any programs in it.
I'd kinda-sorta like to have a Macsyma available to me but my 3b1
won't let me have a program that big. (Haven't looked into WHY yet,
I just know that it won't).
A Meg is minimal (3b1's are slooow with only a meg, but then that
is probably more due to the disk involved rather than the memory).
Regardless of how much memory we spec now, people will want more
in the future.
>Pixels... well, a megapixel display is nice. Part of the 3M definition.
>But do you really need that? Unless you're doing CAD work you really only
>need enough to fit a couple of "terminal-size" windows on the screen. With
>ColorFonts you can even use antialiasing and get nice looking text.
Egads. I suppose I could live with only 2 80x24 windows, but I've really
been spoiled by *really*large* screens. As with memory, disk space, processor
speed and every thing else, I want as much as the market will bear.
It *is* nice that my 3b1 has multiple window available. But they're not
available at the same time. The 3b1 screen is too small.
>Networking... that's a toughy. But do you really need it for a home
>workstation?
YES! How else are you supposed to have, with a home machine, multiple
things going on over your modem at a time? The coming age is one of
even higher speed communication available AT HOME. High enough speed
that you couldn't use it very effectively as "just" a terminal connection.
You'll have enough bandwidth to make something like SLIP or DNET feasible,
allowing you to be doing a terminal session + background file transfers
from home.
--
<---- David Herron -- The E-Mail guy <da...@ms.uky.edu>
<---- or: {rutgers,uunet,cbosgd}!ukma!david, da...@UKMA.BITNET
<---- Windowing... The Amiga has windowing. The Mac has windowing (echoes of
<---- Jonathan Livingston Seagull: "Just flying? A mosquito can do that much!").
Peter da Silva
In article ... mp...@andrew.cmu.edu (Michael Portuesi) writes:
> *Excerpts from: 23-Apr-88 Re: AMIX? Peter da Si...@sugar.UUC (3262)*
> > Memory... well, 4 megabytes is nice. I have that much in my Amiga. But
> > it's not necessary. Let's say you need enough memory that you can run the
> > rest of the stuff without swapping. Carnegie-Mellon specs 1 Meg for a
> > workstation, by the way.
> <giggle>
> [X takes more RAM than] the window manager currently in use...
> Basically, Carnegie Mellon is outright lying when they say 1
> MB is sufficient memory for a workstation.
I guess CMU's wimped out as well.
From what I've hear of Andrew (at Usenix) it's massive overkill. It's
interesting that it seems to do way more than X, yet X is larger.
> > Windowing... The Amiga has windowing. The Mac has windowing (echoes of
> > Jonathan Livingston Seagull: "Just flying? A mosquito can do that much!").
> > You don't need megabytes of RAM or Disk for windowing.
> You do if you want to run X on your workstation.
Why do I want to run X on my workstation? It's a huge monster that doesn't
seem to do any more than Intuition. All I need to do is provide the X toolkit
calls, right? That's the whole point of a portable standard, no?
Personally, I'd rather run NeWS. But if you can emulate X under NeWS I'm sure
Bill Hawes could hack up an X compatibility package in AREXX in a week or so.
:->
> > Networking... that's a toughy. But do you really need it for a home
> > workstation?
> I don't think Chuck was defining a "home" workstation...he was defining a
> workstation.
OK. We have a bunch of intel 310 and 320 boxes on an ethernet with a pretty
transparent common file system, and it doesn't seem to take that much in the
way of resources. An 80286 is in no way shape nor form a workstation class CPU,
but it seems to run Xenix and OpenNet just fine.
And there is Ether for Amy, too.
> And networking is one of the most useful aspects of a workstation.
Funny, the first half dozen workstations I ever had an opportunity to play
around with didn't have any networking. What they had was a big bitmapped
screen, UNIX, and windows.
Peter da Silva
Individuals can't afford 20 grand for a personal computer.
A workstation is a graphically-oriented computer that doesn't keep you
waiting because you don't have a prompt (or because your cursor looks
like a wristwatch).
This requires multitasking. This does not require megapixel displays or
even megabytes of memory.
I agree, by the way. System V is a monstrous bogosity. Of course, 4BSD is
as well. Sometime between 1979 and now everything went to hell...
But at least System V is marginally smaller and it is supported by AT&T.
Stephen Moehle
So far I don't think I have seen anyone say what AMIX is. So.... I know
AMIX is some kind of UNIX look-a-like, but what is it really?
Stephe
{ucbvax,pacbell,hplabs}!well!sjm or well!s...@lll-winken.arpa
"You heard the weirdo man. What is truth?"
Peter da Silva
> Now, in a local environment, with an ethernet board in the Amiga, it
> could (I guess) do this. But being limited to one serial port is a problem.
Yes, the one-serial-port situation is bogus. This is one place I think
Autoconfig is a botch... a serial port is something that'd be a $50 card
if it was as easy to add it to Amy as to an Apple II.
Personally I'd be happy to give up my parallel port for another serial port.
Parallel printers are a historical dreg that has no technical justification.
Peter da Silva
> Regardless of how much memory we spec now, people will want more
> in the future.
I didn't say "let's only have a meg", I said "let's not require people to
spend more than the price of a Yugo for a personal computer", which translates
to "we have to be able to get along with a meg" (yow. Boy was I happy when the
school upgraded to 48K on the computer club's Apple II).
> Egads. I suppose I could live with only 2 80x24 windows, but I've really
> been spoiled by *really*large* screens. As with memory, disk space, processor
> speed and every thing else, I want as much as the market will bear.
The market won't bear super expensive monitors. Once again, let's not spend
more than the price of a small car on a personal computer.
> YES! How else are you supposed to have, with a home machine, multiple
> things going on over your modem at a time?
That's a software problem. I'm talking about "why would you need ethernet
in the home"?
<-- Windowing... The Amiga has windowing. The Mac has windowing (echoes of
<-- Jonathan Livingston Seagull: "Just flying? A mosquito can do that much!").
Yow. I've been promoted to someone's .signature.
Roger Vossler
*True, but you don't have to shoot so HIGH for the HOME Unix market (and,
*believe me, there could be such a market)! Ever had the feeling that putting
*Unix on PC/XT class machines was an idea slightly ahead of its time??
*
*David Masterson
*DMast...@cup.portal.com
Hi Gang,
Seems to me that a good shell and the usual collection of U**X
utilities on top of a good multitasking kernel gives one MOST of
"real unix" without all of the expense (resource, price, etc) and
hassle. Since this already exists for the Amiga, thanks to some very
talented people in both the commercial world and the public domain,
why do we need "real unix"? If one wants to go that far, buy a Sun
and have the real thing. Or have I missed something somewhere?
--
-- Roger Vossler
TRW, Bldg O2-1395, One Space Park, Redondo Beach, CA 90278
BIX: rvossler UseNet: dra...@trwspf.trw.com
ATT: 213.535.2804 ....!trwrb!trwspf!dragon
Tommy Petersson
In article <50...@xanth.cs.odu.edu> ke...@xanth.UUCP (Kent Paul Dolan) writes:
>In article <12...@sbcs.sunysb.edu> ro...@sbcs.sunysb.edu (SBCS Systems Staff) writes:
>****************************************************************************
>* Real Unix. We're comfortable with 4.3, or SunOS 4.X. SysV not welcome. *
>****************************************************************************
Isn't it so that the Berkely Unix will evaporate?
It should follow the path of Xenix and merge into SysV.
This is what I have heard of the future Unix plans. I don't think 4.3
will merge with System V without leaving traces in the resulting
product, but it should be a lot more SysV than Berkely...
Sorry if it's all wrong and I make it harder for some of you to sleep
tonight totally unnecessary, but I would like to know myself. The Unix
we run here includes most of both worlds, but I think it's BSD based.
That may be wrong, too...
Richard Sexton
In article <50...@xanth.cs.odu.edu> ke...@xanth.UUCP (K*nt Pall Dolan) writes:
>>
>> Listen to Chuck. He is giving you the straight story. Real unix
>> machines cost bucks.
>****************************************************************************
>* Real Unix. We're comfortable with 4.3, or SunOS 4.X. SysV not welcome. *
>****************************************************************************
We, Kimasabee ?
>Still no? One more try:
>
> *******************
> *SysV not welcome.*
[evidence of K*nt's editor gone bezerk deleted]
> *SysV not welcome.*
> *******************
>
>Now? Those are the paying customers speaking guys, and they are NOT
>kidding around, and that is NOT an isolated opinion. Think it
>through, please? It's your corporate survival, and my IRA retirement
>plan. Do it right, OK?
Whats this "AH HAS SPOKEN" crap, K*nt ?
SYS V gives the marketing guys a sense of the warm fuzzies, and ya sure
we'd all like BSD, this is a subjective, not objective issue, and becomes
moot in the near future when the two merge.
>K*nt, the overbearing man from xanth.
A BROWN wallet please, eelskin if ya got it.
--
Just a flaming nincompetent poop kinda guy
ric...@gryphon.CTS.COM rutgers!marque!gryphon!richard
Peter da Silva
> Hi Gang,
> Seems to me that a good shell and the usual collection of U**X
> utilities on top of a good multitasking kernel gives one MOST of
> "real unix" without all of the expense (resource, price, etc) and
> hassle.
Sorry, but it doesn't. For an individual running in a non-super account UNIX
is an inherently safe environment. There is very little you can do to the
system as a whole, and if you set up a number of accounts for yourself for
different tasks you can even set up firewalls at the file level. On the other
hand, AmigaDOS is an inherently unsafe environment.
I think it's exteremely unlikely that something like the Amiga, Mac, or PC
viruses could have been implemented on UNIX. Viruses are possible, yes (do
you decode your shar files by hand???), but unless you run one in super you
can't hurt that much... because in UNIX you can't get to anything except
via the operating system.
On the Amiga (or ANY other PC) on the other hand there is no protection. There
is no way to keep viruses or just clumsy users from deleting dh0: (or c:).
Another point is that UNIX is well-documented. Try to figure out just what
Execute() on mi Amiga does from the manual.
> why do we need "real unix"? If one wants to go that far, buy a Sun
> and have the real thing. Or have I missed something somewhere?
And the other point is that individuals shouldn't have to pay the price of a
small car for a decent UNIX system.
Don't get me wrong. AmigaDOS has strengths UNIX lacks. But it's not, and never
will be, UNIX.
eric townsend
| In article <7...@trwspf.TRW.COM>, dra...@trwspf.TRW.COM (Roger Vossler) writes:
| > why do we need "real unix"? If one wants to go that far, buy a Sun
| > and have the real thing. Or have I missed something somewhere?
| And the other point is that individuals shouldn't have to pay the price of a
| small car for a decent UNIX system.
You don't. Buy a fire sale 7300 or 3b1. :-) Actually, buy a used system
from somebody who bought fire-sale priced equipment.
--
Just say "Noh". -- Association for the Advancement of Japanese Theatre.
Girls play with toys -- Real Women skate. -- Powell & Peralta ad
"The truth of an opinion is part of its utility." -- John Stuart Mill
Roger B.A. Klorese
(repeatedly)...
Well, I hope you tell this to the Beserkeley-heads' guru-of-choice, Bill
Whatsisname. Mr. Joy, in an interview with Unix Review, reported that there
would be no SunOS release 5, but rather, System V.4...
---
Roger B.A. Klorese MIPS Computer Systems, Inc.
{ames,decwrl,prls,pyramid}!mips!rogerk 25 Burlington Mall Rd, Suite 300
rog...@mips.COM Burlington, MA 01803
* Your witticism here.* +1 617 270-0613
Peter da Silva
> In article <19...@sugar.UUCP>, pe...@sugar.UUCP (Peter da Silva) writes:
> | And the other point is that individuals shouldn't have to pay the price of a
> | small car for a decent UNIX system.
> You don't. Buy a fire sale 7300 or 3b1. :-) Actually, buy a used system
> from somebody who bought fire-sale priced equipment.
The 3b1 is a perfectly decent machine, unfortunately it's also not being
manufactured. UNIX and Windows seperately are available for well under
2 grand for a base system... but due to relentless upsizing in workstations
you still can't get them together for under 10 grand. You should be able to
get a Sun-1 class machine for 2 grand today, but nobody's making them.
If the Sun-1 was acceptable in 1984, why shouldn't it be acceptable in 1988?
The sort of PC most people are using is still decades behind it.
Hell, the sort of machine most PC owners are stuck with is still decades
behind a PDP-11 running 5th Edition.
David E. Bakken
> *******************
> *SysV not welcome.*
> [etc.]
> *SysV not welcome.*
> *******************
>
> Now? Those are the paying customers speaking guys, and they are NOT
> kidding around, and that is NOT an isolated opinion. Think it
> through, please? It's your corporate survival, and my IRA retirement
> plan. Do it right, OK?
Please note that the details we have heard about the A2500UX are for
European markets, where they demand "pure" SysV. It is entirely
conceivable to me that C= could offer some sort of hybrid or even a
Berkley kernel for the US market. I'm sure C= is well aware of the North
American sentiments as expressed by Kent and others, and they have capable
people in both sales and engineering (their advertising budget is
another story, but I know they have been fighting a war against
bankrupcy, and making good progress lately).
I hope they offer SLIP with the machine, personally.
--
Dave Bakken Boeing Commercial Airplanes (206) 277-2571
uw-beaver!apcisea!hrsw2!bakken
Disclaimer: These are my own views, not those of my employers. Don't
let them deter you from buying the 747 you've been saving hard for.
eric townsend
> If the Sun-1 was acceptable in 1984, why shouldn't it be acceptable in 1988?
> The sort of PC most people are using is still decades behind it.
> Hell, the sort of machine most PC owners are stuck with is still decades
> behind a PDP-11 running 5th Edition.
In the past 6-12mos, I think I've seen a couple of PDP-8's and 11's up for
sale, Real Cheap. Think we could bring back the dead?
Curiosity: Just how does a PDP-8 or PDP-11 being used by one person
rank up side an XT or an AT?
[The author has only seen pictures of PDP-11's, and has never touched
either. :-) ]
LOOK! I didn't include peter's .signature!!!
Brian T. Schellenberger
|Personally I'd be happy to give up my parallel port for another serial port.
|Parallel printers are a historical dreg that has no technical justification.
What, huh?? I've never seen anything shove bits as fast over a serial line
as you can over a parallel line. Try comparing the time to download fonts
to LaserJet II on parallel vs. serial (even at 19.2Kbaud) sometime, and see
if you don't want to keep your parallel port around when you get a laser
printer.
--
--Brian.
(Brian T. Schellenberger) ...!mcnc!rti!sas!bts
. . . now at 2400 baud, so maybe I'll stop bothering to flame long includes.
Jeff Gortatowsky
> 100+ Suns here, and I'll tell you that a 3/50 is pretty well threshold
> of pain in current workstations. In such an environment, the current
> Amiga makes an awesome terminal with some compute ability. It does
> not make a workable Unix machine. At least for the sort of Unix
> we run here.
>
I humbly suggest Sir that you are spoiled. I'd K.I.L.L for a 2/120 at home.
Sure, starting applications takes some time (dbxtool takes longer than *some
time* ;^> ) but it's still real live UNIX with umteen bazzilion
utilities. For a programmer it's almost a nothing else to buy system. Add
in TEX, a few GNU items, X (if you please), VC, and it's almost a nothing
else to buy system for a *user* (whatever THAT is!).
> Listen to Chuck. He is giving you the straight story. Real unix
> machines cost bucks.
>
> Rick Spanbauer
> SUNY/Stony Brook
>
Don't *always* listen to Chuck. Sun charges what it does not only to cover
thier considerabe technology and research, but to keep workstations out
of the "consumer" realm. Further it appears that Sun would give away
CPU's as long as you buy a disk sub-system from them. Check out the price
sheet some time. The disk prices are horrible when you look at the cost
of an SCSI bare drive and a SCSI->ST506 adaptor. ESDI I'm not so sure of.
Imagine my surprise to find out I could get a 3/50 4meg CPU for *less* than
the minimum 142meg shoebox with WangTek tape drive! And here in my part
of Kodak we have more Suns than people to operate them, so I know of what I
speak(we use Suns (not Sun-4) in our product)! Now maybe things have
changed....but I doubt it.
Look CBM. I don't expect a 16.67 no-wait-state 1megapixel by 8bit machine
running 4.3 + X11 with a 140mb disk for under $2000. But how about a
14.32mz 020 with a 764x512x5 (or 6) with a 50meg ST506 disk and VM kickstart
for ~$2500-$3000? Don't go after Sun and ilk. Let them tantilize
us at work, while we laugh and enjoy KS at home.
Oh how I'd have paid that extra 100-150 bucks for at least process protection
on my A-1000 (vintage 10/85)! Listen up CBM!
--
Jeff Gortatowsky .....allegra!rochester!kodak!elmgate!jdg
Eastman Kodak Company
These comments are mine alone and not Eastman Kodak's. How's that for a
simple and complete disclaimer?
Eric Green
> Another point is that UNIX is well-documented. Try to figure out just what
> Execute() on mi Amiga does from the manual.
WHAT??? What have you been smoking????
I have here a complete set of Sys V.2 manuals (the 5-volume set), plus
the old two-volume V7 manual (from ancient days).
And, I also have here, the complete 5-volume Amiga manual set, plus a 2-volume
3rd-party reference manual. And, of course, the autodocs, the equivalent of
the Unix man pages.
As far as readability, usability, etc., go, I find them to be about the same.
Somewhat cryptic, needing a lot of work to dig out gems from, but, eventually,
you can find anything you need. In fact, the Exec and Intuition documentation
probably is above average, compared to the "average" computer manual, in that
some attempt is made at trying to explain the basic structure of the system
(sure, they don't succeed... but at least it's not just a bunch of "man" pages
printed out and bound together, like the Unix manuals).
The exception is, of course, the Dos documentation. Which stinks, Period.
About what you would expect. Throw it in da trash (BPTR's... BLETCH!).
Methinks that the only reason someone might think Unix is better documented,
is if that person is a Unix guru who reads system source code with his
breakfast. That much, at least, you cannot do with AmigaDos (sigh). Source
code may be, indeed, the ultimate documentation (I know that I would never
have been able to figure out "termcap" from the documentation...). But most of
us would be happy to have decent documentation, and leave source code to the
masochists (it's HARD to decipher other people's sources, in case you haven't
noticed -- especially when it's all uncommented, like the Unix system
sources).
--
Eric Lee Green e...@usl.CSNET Snail Mail P.O. Box 92191
ihnp4!killer!elg Lafayette, LA 70509
"Is a dream a lie that don't come true, or is it something worse?"
Peter da Silva
> sale, Real Cheap. Think we could bring back the dead?
If you can get a PDP-11 for cheap, you should probably do it. There'll be
better support for it than your 3b1 :->. Just how cheap is real cheap?
Seriously, there's a guy in town using a PDP-11 running RSX as a BBS. If
you're interested in having a look, send me mail.
> Curiosity: Just how does a PDP-8 or PDP-11 being used by one person
> rank up side an XT or an AT?
A PDP-8 is pretty useless, but an '11 is not a bad little machine. Better
than AT+DOS. Probably not as cool as AT+UNIX.
SBCS Systems Staff
> In article <12...@sbcs.sunysb.edu> ro...@sbcs.sunysb.edu (SBCS Systems Staff) writes:
>
> I humbly suggest Sir that you are spoiled. I'd K.I.L.L for a 2/120 at home.
Jeff, I think you would want to K.I.L.L. the 2/120 once you had
X windows, 3.5 SunOS, dbxtool, a shell or two, etc all running at
once. On the 2/XXX series we have here (even with 3 mBytes) the
system is almost unusably slow once you a just a few Suntools
windows, etc going.
> Don't *always* listen to Chuck. Sun charges what it does not only to cover
> their considerabe technology and research, but to keep workstations out
> of the "consumer" realm. Further it appears that Sun would give away
Oh oh. This sounds like the start of another Sun "the evil empire"
thing. Commdore doesn't spend money doing "considerabe technology
and research" :-) ? Ahem.
> CPU's as long as you buy a disk sub-system from them. Check out the price
> sheet some time. The disk prices are horrible when you look at the cost
> of an SCSI bare drive and a SCSI->ST506 adaptor. ESDI I'm not so sure of.
I have a home 3/50. It cost me $5K. The drive I use is a CDC Wren
155 mByte; cost $1700. For both I paid list price, though I've
seen used 3/50's for $2500 or so. Anways, compare what you get for
$7.1K (adding $150 for Unix+$250 for manuals):
disk not applicable
1152x900 BW non interlaced monitor
15 mHz 68020
4 mBytes 32 bit ram, parity checked
Ethernet/cheapernet controller
DMA SCSI controller
fast MMU
2 serial ports
*real* Unix (think I'll trademark that)
---
$5.2K
Commodore, will field something that looks like this:
disk not applicable
1008x800 BW non interlaced monitor ~$750
14.318 mHz 68020 $1500
1 mByte 32 bit ram
DMA ST506/SCSI controller $350
68851 MMU (1 wait state)
B2000/1 serial port, 1 parallel port ~$1500
SysV Unix (yuck) $800
------
$4900
Of course, to get apples to apples you would have to add
another 3 mBytes 32 bit ram, and Ethernet controller to the
B2000 or a floppy drive, sound, parallel port, blitter, etc to the Sun.
Neither Sun or Commodore manufacture disk drives, so they both
pay the same amount of $$ for a drive. No sense in comparing those.
Anyways, I am not bashing Commodore's efforts to produce and sell
Unix workstations. It just seems to me that they have a long road
ahead of them before they can off goliaths like Sun, Apollo, etc
on their own turf.
> --
> Jeff Gortatowsky .....allegra!rochester!kodak!elmgate!jdg
> Eastman Kodak Company
Rick Spanbauer
SUNY/Stony Brook
Chuck McManis
** Please don fireproof clothing now ... **
In article <9...@elmgate.UUCP> j...@aurora.UUCP (Jeff Gortatowsky) writes:
> Don't *always* listen to Chuck. ...
No, you should never take the words or advice of one person as your sole
source of information. However...
> ... Sun charges what it does not only to
> cover thier considerabe technology and research, but to keep
> workstations out of the "consumer" realm. ...
This is presumptuous and borders on slander. Sun charges what it does
because :
a) The Sun price is often the best cost/performance price on
CPUs anywhere.
b) It pays my salary and overhead.
c) We sell everything we can make and there is no incentive
to lower prices other than "a" above.
You will note that both Scott McNealy and Bill Joy have said publically
that their goal is to get UNIX workstations to the most number of people.
> ... Further it appears that
> Sun would give away CPU's as long as you buy a disk sub-system from
> them. Check out the price sheet some time. The disk prices are
> horrible when you look at the cost of an SCSI bare drive and a
> SCSI->ST506 adaptor. ESDI I'm not so sure of. ...
And guess what ? The only part of the 141 meg "shoebox" that Sun sells,
that is made by Sun, is the casework! Wow you mean disk drive makers
selling into our market can undercut us? You bet they can, they are
selling us the disks too.
> ... Imagine my surprise
> to find out I could get a 3/50 4meg CPU for *less* than the minimum
> 142meg shoebox with WangTek tape drive! And here in my part of Kodak
> we have more Suns than people to operate them, so I know of what I
> speak(we use Suns (not Sun-4) in our product)! Now maybe things have
> changed....but I doubt it.
Basically, it would cost Kodak exactly as much to sell shoeboxes as it
does Sun. There is no financial advantage to remarketing drives, ever
wonder why Commodore doesn't sell them in the US? You can't compete
against Seagate, and Seagate is competing against MiniScribe et al so
they are always driving the price down. When we sell a disk drive we
install it, service it and swap it out if it breaks. If you are
sophisticated to build your own shoe boxes you don't need us. So don't
bother us ok?
** Ok that aside ... **
> Look CBM. I don't expect a 16.67 no-wait-state 1megapixel by 8bit machine
> running 4.3 + X11 with a 140mb disk for under $2000. But how about a
> 14.32mz 020 with a 764x512x5 (or 6) with a 50meg ST506 disk and VM kickstart
> for ~$2500-$3000? Don't go after Sun and ilk. Let them tantilize
> us at work, while we laugh and enjoy KS at home.
> Oh how I'd have paid that extra 100-150 bucks for at least process protection
> on my A-1000 (vintage 10/85)! Listen up CBM!
A reasonable suggestion. I said as much to them when the rumors of a UNIX
machine first appeared. Basically, there are a lot of people crowding into
the 'personal' UNIX workstation market and in a crowded situation such as
that, companies with poor or marginal brand name recognition such as Atari
or Commodore tend to get squeezed out. Whereas sticking with AmigaDOS or
a protected mode version of the same would allow them to distinguish them
selves with being able to compile and run UNIX programs, but you only
need 1 Meg of RAM and 20Meg of hard disk space. [Yes, there is a lot of
work to get from here to there but not much more than that which is required
to get from no UNIX to full UNIX support.]
--Chuck McManis
uucp: {anywhere}!sun!cmcmanis BIX: cmcmanis ARPAnet: cmcm...@sun.com
These opinions are my own and no one elses, but you knew that didn't you.
Matt Dillon
:as you can over a parallel line. Try comparing the time to download fonts
:to LaserJet II on parallel vs. serial (even at 19.2Kbaud) sometime, and see
:if you don't want to keep your parallel port around when you get a laser
:printer.
Ahh hah haa ha.... Man, you haven't seen serial lines. There are
a couple reasons people use serial rather than parallel. When I
design *any* custom communications hardware, I always use serial.
(1) You can do it with a twisted pair, and only two drivers on either
end. Or at worst three wires (full duplex operation). With
full duplex I usually run data communication synchronously.
(2) You can run the things *fast* ... So fast, in fact, that in
many cases one doesn't even need to check the TxRegister Empty
bit... the bits gets shifted out faster than you can write the
port.
It isn't my fault idiot manufacturers are still running their
serial ports at 19.2KB.
-Matt
Peter da Silva
> in article <19...@sugar.UUCP>, pe...@sugar.UUCP (Peter da Silva) says:
> > Another point is that UNIX is well-documented. Try to figure out just what
> > Execute() on mi Amiga does from the manual.
> WHAT??? What have you been smoking????
Water. I never smoke anything stronger than water.
> As far as readability, usability, etc., go, I find them to be about the same.
Here's where the UNIX manuals and include files have the Amiga's beat all
hollow:
Types.
In the AmigaDOS manuals mainly, but even in the RKMs, you can't easily figure
out what the types of everything are. You have to use the examples or the
include files... but even there's there's WAY too much stuff overloaded on
poor old APTR, ULONG, and USHORT. But better manuals are promised.
> Methinks that the only reason someone might think Unix is better documented,
> is if that person is a Unix guru who reads system source code with his
> breakfast. That much, at least, you cannot do with AmigaDos (sigh).
I wish.
The only think I can recall having to be told about in the UNIX manuals is
the damn overloading of the c_cc feilds in termio.
Peter da Silva
> In article <19...@sugar.UUCP> pe...@sugar.UUCP (Peter da Silva) writes:
> |Personally I'd be happy to give up my parallel port for another serial port.
> |Parallel printers are a historical dreg that has no technical justification.
> What, huh?? I've never seen anything shove bits as fast over a serial line
> as you can over a parallel line.
And for most printers this extra bandwidth is completely wasted: they can't
even keep up with 1200 baud.
> Try comparing the time to download fonts
> to LaserJet II on parallel vs. serial (even at 19.2Kbaud) sometime, and see
> if you don't want to keep your parallel port around when you get a laser
> printer.
If I could afford a laser printer I think I could afford a parallel port card.
I mean, what's another couple of hundred bucks out of a couple of grand? In
the meantime I can't:
1) Put two modems on my Amiga.
2) Put a modem and a terminal on my Amiga.
3) Put a cheap EPROM burner and a modem on my Amiga.
4) Put a Midi port and a modem on my Amiga.
... and any combination thereof.
Proposition 1: the more ports your computer has, the better.
Proposition 2: the more versatile the ports on your computer, the better.
Claim: Serial ports are more versatile than parallel ports, because there
is a greater variety of serial devices out there.
Side issue: If there were multiple serial ports, the whole "serial device
debate" would be a non-issue. Personally, I think it is anyway. We
already have a perfectly good namespace for devices.
Conclusion: For me, at least, serial ports are better than parallel ports.
Totally unjustified flame: I have 4 serial ports on my Atari 800, via my 850.
I have 4 joystick ports on it as well. Why does the Amiga have less?
Richard Sexton
>In article <9...@elmgate.UUCP> j...@aurora.UUCP (Jeff Gortatowsky) writes:
>> Don't *always* listen to Chuck. ...
>
>No, you should never take the words or advice of one person as your sole
>source of information. However...
To be perfectly fair, the number of times chuck was wrong is less than
the number of times JG was right.
The worst thing you can say about chuck is, he is a complete wimp
when it comes to spicey food, but so is leo. Dont know what it is
about those NoCal boys....
>> ... Sun charges what it does not only to
>> cover thier considerabe technology and research, but to keep
>> workstations out of the "consumer" realm. ...
My god, what paranoia.
>This is presumptuous and borders on slander. Sun charges what it does
>because :
> a) The Sun price is often the best cost/performance price on
> CPUs anywhere.
.......................... this side of an apollo
(they made me do it chuck, they made me do it. you can give her hell next
--
noalias went. it really wasn't negotiable
ric...@gryphon.CTS.COM rutgers!marque!gryphon!richard
Randell E. Jesup
> disk not applicable
> 1152x900 BW non interlaced monitor
> 15 mHz 68020
> 4 mBytes 32 bit ram, parity checked
parity = 1/8 greater chance of failure :-)
> Ethernet/cheapernet controller
> DMA SCSI controller
> fast MMU
> 2 serial ports
> *real* Unix (think I'll trademark that)
> ---
> $5.2K
> Commodore, will field something that looks like this:
>
> disk not applicable
> 1008x800 BW non interlaced monitor ~$750
grey-scale (7 levels), not BW.
> 14.318 mHz 68020 $1500
> 1 mByte 32 bit ram
4 sorry, the A2500UX mentioned in Hannover was 4 Meg.
> DMA ST506/SCSI controller $350
> 68851 MMU (1 wait state)
> B2000/1 serial port, 1 parallel port ~$1500
Extra serial ports will soon be available, as is ram
expansion, video boards, etc, etc. The 2000 is an
expandable system, unlike the 3/50. Can a 3/50 have more
than the standard 4 Meg memory?
> SysV Unix (yuck) $800
The European Unix market wants Sys V, not BSD. Also, where
did the $800 come from (not that I have any knowlege one
way or the other.)
Also, you get free color video (Sun color workstations are
NOT cheap.)
> ------
> $4900
> Rick Spanbauer
> SUNY/Stony Brook
Don't get me wrong, I like Suns (I've had them on my desk, and
soon will again.) Even a Sun 100U (Sun-1) isn't bad to have, though I HATE
that silly vt100 keyboard. Of course, having 6 or 7 meg memory doesn't
hurt. :-)
I think the more appropriate comparison would be to the lowest-
level Sun-3 with a backplane for expansion, not the (normally) diskless,
unexpandable workstation (plus I thought they were replacing the 3/50 with
the 3/60).
// Randell Jesup Lunge Software Development
// Dedicated Amiga Programmer 13 Frear Ave, Troy, NY 12180
\\// beowulf!lunge!je...@steinmetz.UUCP (518) 272-2942
\/ (uunet!steinmetz!beowulf!lunge!jesup) BIX: rjesup
(-: The Few, The Proud, The Architects of the RPM40 40MIPS CMOS Micro :-)
SBCS Systems Staff
> In article <12...@sbcs.sunysb.edu> ro...@sbcs.sunysb.edu (SBCS Systems Staff) writes:
> parity = 1/8 greater chance of failure :-)
Or looking at it another way, knowing that you do indeed
have a problem.
> grey-scale (7 levels), not BW.
When I say "BW", I mean !color. Greyscale is nice to have
but it isn't a requirement for my "Unix seat".
> Extra serial ports will soon be available, as is ram
> expansion, video boards, etc, etc. The 2000 is an
> expandable system, unlike the 3/50. Can a 3/50 have more
> than the standard 4 Meg memory?
No, but a 3/60 can take up to (I think) 24 meg of
SIMM memory. I seem to remember that the BW version
of 3/60 is around $8K or so. The 3/60 also has a
kludged (ala A1000) expansion connector called the "P4".
>
> The European Unix market wants Sys V, not BSD. Also, where
> did the $800 come from (not that I have any knowlege one
> way or the other.)
>
The $800 is a guess based on what other small machine
vendors are charging for their "full" SysV ports (with
C compiler, etc). Adjust the number as low as you want,
then proceed to add cost for third party software to
patch missing SysV features, eg mature Window software,
networking, FTN, Pascal, etc, etc..
> I think the more appropriate comparison would be to the lowest-
> level Sun-3 with a backplane for expansion, not the (normally) diskless,
> unexpandable workstation (plus I thought they were replacing the 3/50 with
> the 3/60).
No, I was comparing costs of a Unix seat: Amiga -vs- Sun. My
point is that by the time you're done building an Amiga into
a reasonable Unix seat, you've spent about the same bucks
as you would have on a Sun.
A true apples to apples comparison Sun vs Amiga is probably
not possible, as the Sun machines bring higher expansion bus
throughput, more mature Unix, larger amounts of memory, higher&
deeper color, etc vs the Amiga's sound capabilities, blitter,
cheaper expansion peripherals, etc. This is why I constrained
my comments to a generic Unix seat.
Rick Spanbauer
SUNY/Stony Brook
Eric Green
> In article <39...@killer.UUCP>, e...@killer.UUCP (Eric Green) writes:
>> > Another point is that UNIX is well-documented. Try to figure out just what
>> > Execute() on mi Amiga does from the manual.
Hmm. That's a DOS function. DOS is amazingly ill-documented...
>> As far as readability, usability, etc., go, I find them to be about the same.
>
> Here's where the UNIX manuals and include files have the Amiga's beat all
> hollow:
>
> Types.
>
> In the AmigaDOS manuals mainly, but even in the RKMs, you can't easily figure
> out what the types of everything are. You have to use the examples or the
> include files... but even there's there's WAY too much stuff overloaded on
> poor old APTR, ULONG, and USHORT. But better manuals are promised.
The documentation is far from clear (it took me a few days to figure out what
a ViewPort was, for example), but a lot of it is because AmigaDOS is a much
more complex operating system than Unix (SCREECH! BRAKES! What's that kid
saying?!).
Actually, the difference is more conceptual in basis. Unix originally was
designed to be a very high-level operating system with a minimum of system
calls. Most Unix internals have their detail well hidden (the notable
exception is the #$%"$&% TTY driver, sgtty's and sgttyb's and termios and all
that garbage that confuses people who program both AT&T & BSD). On the Amiga,
all that detail is right out in the open. Perhaps because the final details of
the machine probably weren't known until 2 days before shipping :-). But in
the meantime, we get to deal with window structures, views, layers, ports,
nodes, messages, and all that other good stuff, often with several levels of
indirection thrice removed, and it gets confusing.
AmigaDOS (not DOS, but the rest of the thing) is amazingly flexible because of
all that detail out in the open, but the drawback is that the documentation is
a nightmare, and poor Amiga novices like me have a horrible time bootstrapping
to their new machine. For example, I'm still having problems trying to get
text out to the ten little windows that I just opened up :-).
Eric Green
As an aside, I've seen an Amiga shift data out its serial port at
320Kbaud. Wasn't much machine left over for users, though :-).
I suspect that the main reason that "idiot manufacturers" still run their
serial ports at 19.2Kbaud is because most commercially available (= CHEAP)
UARTs are only capable of 19.2Kbaud. For example, the 6551, which is cheap,
has low chip count (add a crystal, bus buffers, and ready to go), and is
easily interfaced with 68xx/65xx/68K-based machines. I suspect there's an
aweful lot of 6502's out there (e.g. I recently discovered that the TVI 910
was 6502-based, when I patched a bug in their firmware for a local
university).
Higher bitrates seem to require either more expensive chips, or custom chips
like in the Amiga.... and, of course, there's always the ULTIMATE in bitrates,
fiber-optic Ethernets at 5 megabaud and up.
But for a low-cost interface for modems and printers, I suspect that 19.2kbaud
is going to hang on for quite a while (at least until 38kbaud equipment
becomes more common).
--
Eric Lee Green {cuae2,ihnp4}!killer!elg
Snail Mail P.O. Box 92191 Lafayette, LA 70509
Matt Dillon
:serial ports at 19.2Kbaud is because most commercially available (= CHEAP)
:UARTs are only capable of 19.2Kbaud. For example, the 6551, which is cheap,
:has low chip count (add a crystal, bus buffers, and ready to go), and is
:easily interfaced with 68xx/65xx/68K-based machines. I suspect there's an
:aweful lot of 6502's out there (e.g. I recently discovered that the TVI 910
:was 6502-based, when I patched a bug in their firmware for a local
:university).
Actually, you can run most UARTS at >19.2KBaud. I currently have
a small home-built computer consisting of a 6502 running at @2Mhz+, several
other IO chips but most noteably two 6551A's running at 76.8KBaud. It is
just a matter of giving the thing a faster crystal. I would be surprised
if you could run x1 UARTS (that require a clock in for synching but do not
require a 16x clock for sampling) at 1MBit easy. Note: read, non-custom
and low cost parts here.
-Matt
Chuck McManis
> I think the more appropriate comparison would be to the lowest-
>level Sun-3 with a backplane for expansion, not the (normally) diskless,
>unexpandable workstation (plus I thought they were replacing the 3/50 with
>the 3/60).
Well Randell, as soon as you can point to an Amiga with a VME bus I will
compare it :-). Lets stop this shall we? Amigas aren't Suns and Suns aren't
Amigas, I have one of each on my desk and find each does some things
better than the other. I wouldn't recommend an Amiga to run UNIX and I
wouldn't recommend a 3/{5|6}0 to do animations.
Matt Dillon
:require a 16x clock for sampling) at 1MBit easy. Note: read, non-custom
:and low cost parts here.
:
: -Matt
Sorry, I meant "I would NOT be surprised..."
-Matt
doug-m...@cup.portal.com
(e.g. no extra charge for C compiler) for $145 for the 80386. I just
saw an ad. Pretty smart of them...this could help keep half-OS from
contaminating too many systems (easy to think of it as a virus, no?)
Doug
---
Doug Merritt ucbvax!sun.com!cup.portal.com!doug-merritt
or ucbvax!eris!doug (do...@eris.berkeley.edu)
or ucbvax!unisoft!certes!doug
SBCS Systems Staff
> In regard to "$800" for Unix...ATT is now advertising *full* Unix
I was thinking of Microport SysV - it sells for $680 or
so for a full system. I would guess that ATT can sell
for that price since they probably don't support it,
and they surely don't have to pay royalities :-).
> ---
> Doug Merritt ucbvax!sun.com!cup.portal.com!doug-merritt
Rick Spanbauer
SUNY/Stony Brook
doug-m...@cup.portal.com
I said:
> In regard to "$800" for Unix...ATT is now advertising *full* Unix
Rick Spanbauer said:
>I was thinking of Microport SysV - it sells for $680 or
>so for a full system. I would guess that ATT can sell
>for that price since they probably don't support it,
>and they surely don't have to pay royalities :-).
I made a mistake, it's not AT&T that's selling 386 Unix, it's Bell
Technologies. The price is $145 for a 2 user license or $275 for
an unlimited user license. The system is the Interactive Systems
port of Unix System V release 3, certified by AT&T.
Includes "all development tools, C Compiler, link kit, vi and
all other Unix utilities, files and programs, including Remote
File System, Streams, and TLI". Apparently the programmer's manuals
and other volumes are separate. This is from their ad on pg 107
of the May 1988 Unix Review.
Interactive Systems themselves have an ad in the same issue for 386/ix,
which sounds like the same port. No price; ad on pg 7.
Contact Bell Technologies at 800-FOR-UNIX (in CA 415-659-9097).
330 Warren Ave. Fremont CA 94539
I have no affiliation etc. Just helping to eradicate the heartbreak
of Half-OS in our lifetimes. :-)
Dave Haynie
> In article <40824UH2@PSUVM>, U...@PSUVM.BITNET (Lee Sailer) writes:
>> Now, in a local environment, with an ethernet board in the Amiga, it
>> could (I guess) do this. But being limited to one serial port is a problem.
> Yes, the one-serial-port situation is bogus. This is one place I think
> Autoconfig is a botch... a serial port is something that'd be a $50 card
> if it was as easy to add it to Amy as to an Apple II.
There's no big deal, hardware-wise, adding an additional serial port. Auto-
config doesn't get into the way, and since such a card would very likely
be of fixed size (probably a 64K jobber), the autoconfig logic for such a
card would be simple. Really. Though if you're adding more than maybe one
or two additional ports, I'd add a smart controller to manage them all and
offload that intensive interrupt activity from the host CPU.
The software aspects of additional serial ports are another matter. You
could just go directly to them, like you would on an Apple II or C64, but
that wouldn't be a very nice thing to try given that we're multitasking.
Most of the talk about additional serial ports has presumed that the serial
drivers would try to be an extension of the current serial.device or some
equally generic device addition would be available. Since the built-in
devices as driven by the Port-Handler are kind of a kludge, this doesn't
just drop nicely into the OS like it should. So there's additional
software complexity that must be considered to do it "right", but not any
real hardware problem. You probably could build a serialx2 card for the
A2000 that retails in the $100 range, maybe even serialx4.
> Personally I'd be happy to give up my parallel port for another serial port.
> Parallel printers are a historical dreg that has no technical justification.
The folks who make those parallel printer cables have been kicking back just
enough money to all us CPU designers to keep these things around....
> -- Peter da Silva `-_-' ...!hoptoad!academ!uhnix1!sugar!peter
> -- "Have you hugged your U wolf today?" ...!bellcore!tness1!sugar!peter
> -- Disclaimer: These aren't mere opinions, these are *values*.
--
Dave Haynie "The B2000 Guy" Commodore-Amiga "The Crew That Never Rests"
{ihnp4|uunet|rutgers}!cbmvax!daveh PLINK: D-DAVE H BIX: hazy
"I can't relax, 'cause I'm a Boinger!"
Peter da Silva
> > Parallel printers are a historical dreg that has no technical justification.
> The folks who make those parallel printer cables have been kicking back just
> enough money to all us CPU designers to keep these things around....
So *that's* why they're so expensive!
--
-- Peter da Silva `-_-' ...!hoptoad!academ!uhnix1!sugar!peter
-- "Have you hugged your U wolf today?" ...!bellcore!tness1!sugar!peter
-- Disclaimer: These may be the official opinions of Hackercorp.