High Res on Model 1/3/4
Skip May
[snip]
:this is in regard to trs80 model 3
:too bad radio shack did not become standard....no help from
:trsdos.
Yes, there was/is a debate over how sucessfull Tandy/Radio Shack actually
was or "was not". They sure shot themselves in the foot quite a bit.
:we have ldos---and really like z80---no overcomplicated etc.
The Z-80 is my processor of choice. Of particular interest to me is
Amateur Radio uses, where the Packet TNC boxs were almost all Z-80 at one
time.
:anyway if you have a floppy disk control card please let me know
:also it would sure be nice to install a hard drive in this computer
:and a graphics card of better resolution (do not know if
:ever existed)
Yes, I do have one controller card comming out of service. I'll have to
look to see if it's already promised. You would actually have to install
the hard disk external to the Model 3/4. I've never seen one done
internal, just by physical size.
Yes, there are/were Hi Res cards for the Model 3 and 4, made by Microdex I
think and tandy both. One model 4 I have has one on board. Sure looks nice
for the graphics that came with the card on disk... that famous space
shuttle picture.
:thanks again (these would rcv. a nice home) :-)
:f...@uswest.net
Others may respond to you from this post on the news group, hope you find
what your looking for.
PS. I plan to respond to all who inquired about my earlier "give-away"
posts. A death of a friend has been taking much of my time.
PSS. I plan to write the new link pages for my home page this weekend. I
think I have most of you in the list. Email me at home if you want to be
sure.
cheers
skipp
home
nosp...@juno.com
the TyRannoSaurus
Skip May <sk...@catbert.ucdavis.edu> wrote in article
<734ne6$r7q$1...@mark.ucdavis.edu>...
> I received this through my trs80 home page...
> [snip]
>
> :this is in regard to trs80 model 3
> :too bad radio shack did not become standard....no help from
> :trsdos.
>
> Yes, there was/is a debate over how sucessfull Tandy/Radio Shack actually
> was or "was not". They sure shot themselves in the foot quite a bit.
They could have done exactly the same as Apple did: bring out a new and
revolutionary system to compete with the then upcoming PC-standard (like
the Apple Mac) and/or extend the life of their popular eight-bit machines
with new technology.
For the first they tried with the Tandy 2000. I think that the fact that it
was partially compatible (on DOS-level) was actually the cause that it
didn't catch on. People could live without compatibility, but not with
partial compatibility, because you never knew where you stood, with the
T2000.
A superior but incompatible system with graphics could have done it.
Imagine a version of Unix/Xenix on the T2000, with an X-windows interface
on it... It could have done what Linux is doing today, and 15 years before
!
For the second they had 2 possibilities: the TRS-80 line and the Coco-line.
The TRS-80 model 4 was slightly improved with the model 4P (compare with
the Apple //c) and 4D, but they failed to look further into the model 4GS
(similar to the Apple //gs) wich could have been a hit, especially in the
educational market. The processors were there (Z-280) the operating system
was (LS-DOS could be ported to a different CPU) and they could have kept
virtual complete compatibility with the model III and 4. Sad that they
didn't.
The coco could have been the subject of the same improvements, and
certainly had a better design to continue to build upon. The operating
system was way ahead of evrybody else (OS-9 was modeled after Unix) and the
CPU was certainly axpandable. With the help of Motorola, compatibility
problems between the 6809 and the 68000 could have been solved. Again, they
could have taken the place of the Amiga. The Coco 3 was a good step, but
why they didn't follow up with the Coco 4 with e.g. a 680xx CPU, a build-in
disk-drive and even better graphics and memory, I don't know.
"Don't try to be an innovator when your experienec comes from selling
batteries for stuffed-animal radio's" I think the editor of 80micro said
that ...
Their biggest impact, but their stupidest decision, was the model 100-line.
Instead of expanding on the idea of the portable computer without moving
parts, they restricted it to a mere word-processor. Imagine a "docking
station" for the T102, where you clip it in and you have floppy-capacity
(even HD, maybe) external monitor and so on. Their VDI was a good step, but
it wasn't an integrated design. Their decision to stop marketing it and
jumping on th MS-DOS bandwagon reduced them to just another clone maker
(with a quirck because of their T1000)
> :we have ldos---and really like z80---no overcomplicated etc.
I see LS-DOS still as the most user-friendly command-line-driven operating
system there is.
> Yes, I do have one controller card comming out of service. I'll have to
> look to see if it's already promised. You would actually have to install
> the hard disk external to the Model 3/4. I've never seen one done
> internal, just by physical size.
Oh, you could fit a HD internally in the model 4, no problem there. But
where would you put the HD-controller board ? I haven't seen the IDE
controller, sold by CN80, but the controller board for the old Tandy HD's
was almost as big as the motherboard !
> cheers
> skipp
Greetings from the TyRannoSaurus
Jan-80
Frank Durda IV
: For the first they tried with the Tandy 2000. I think that the fact that it
: was partially compatible (on DOS-level) was actually the cause that it
: didn't catch on. People could live without compatibility, but not with
: partial compatibility, because you never knew where you stood, with the
: T2000.
: A superior but incompatible system with graphics could have done it.
: Imagine a version of Unix/Xenix on the T2000, with an X-windows interface
: on it... It could have done what Linux is doing today, and 15 years before
: !
Uh, XENIX was sold for the Tandy 2000. It was a failure. The 2000 hardware
was just too unreliable (over two hundred "blue wire" modifications
on the final units shipped), and the graphics hardware was designed in such
a way that speed was never a possibility. Having 100% of the programmers who
developed the machine quit within two months of its introduction didn't help
either. Add this to the 80x186 with no protected memory concept, allowing
one task to clobber another, and a multi-user/multi-tasking system was pretty
much a joke on that platform. Intel didn't help by deliberately re-using I/O
ports for their integrated hardware that were already "known" ports in the
IBM/PC world. Microsoft compounded the error by using INTs and other areas
that Intel had clearly marked as "reserved", so when the 80x186 came along,
things broke. Most people didn't notice all the things breaking until the
80x286, since only Tandy and Olivetti ventured into the 80x186 universe as
a PC platform processor. Since that time, the 80x186 has been doing embedded
applications, like traffic lights, and not trying to be a PC anywhere.
Also, MIT X was incredibly immature at that time, and only ran on two
platforms, neither of which were Intel-based. The Sun SPARC was just
coming on the market, just to give you an idea of where we were in history.
Even the so-called Windows of the day (you can see the Model 2000 ad running
Windows on the back of a Byte magazine starring Bill Gates) shows that
Windows 1.x was text-based in those days, but was in COLOR.
: For the second they had 2 possibilities: the TRS-80 line and the Coco-line.
: The TRS-80 model 4 was slightly improved with the model 4P (compare with
: the Apple //c) and 4D, but they failed to look further into the model 4GS
: (similar to the Apple //gs) wich could have been a hit, especially in the
: educational market. The processors were there (Z-280) the operating system
: was (LS-DOS could be ported to a different CPU) and they could have kept
: virtual complete compatibility with the model III and 4. Sad that they
: didn't.
The problem here was easy. The Tandy buyer for the Model 4 was also the buyer
for the Tandy 1000 in 1984, and he made the decision to kill off all products
under his control except the 1000 line by starving the other products of new
software and advertising. Despite this, Model 4D were still being made as
late as early 1987, and being sold for no less than 4X the manufacturing cost.
: The coco could have been the subject of the same improvements, and
: certainly had a better design to continue to build upon. The operating
: system was way ahead of evrybody else (OS-9 was modeled after Unix) and the
: CPU was certainly axpandable. With the help of Motorola, compatibility
: problems between the 6809 and the 68000 could have been solved. Again, they
: could have taken the place of the Amiga. The Coco 3 was a good step, but
: why they didn't follow up with the Coco 4 with e.g. a 680xx CPU, a build-in
: disk-drive and even better graphics and memory, I don't know.
Again, Tandys marketing decided to kill the Coco when the 1000s got hot, and
the designer who was doing the work on the most advanced Coco ever had his
project canceled when it was nearly done and he quit. Two years later he
came back to and did a cost-reduced version of the Coco2, the machine that to
die, but it also was not advertised and none of the fancy stuff from the
Deluxe Color Computer ever made it to the stores.
: "Don't try to be an innovator when your experienec comes from selling
: batteries for stuffed-animal radio's" I think the editor of 80micro said
: that ...
Actually, there was plenty of innovation, at least until PCs and total
Tandy obedience to Microsoft took over. Tandys R&D had many bright people,
usually squashed under a mass of management incompetence. Despite that,
some good things did leak out. Look at what Tandy did first:
o Serial keyboard interface, Model II, 1979, copied by IBM PC in 1981
o Double-Sided floppies 1982, copied by IBM PC/AT in 1984
o Winchester hard disk systems, Model II, 1981, copied by IBM PC/XT in 1983
o First major computer maker to have a generic card cage design, allowing
for upgrades, including stuff from third parties. Undoubtedly, MITS
and IMSAI did it first, but they were tiny compared to the big players
that existed in 1979, when this appeared in the Model II. Tandys
mistake was trying to keep the card cage "closed", and not scaring
people into not putting non-Tandy cards or parts in TRS-80s.
o First computer maker to bundle application software with new systems, 1983.
Apple copied in 1984. IBM finally did in 1989.
o 12-function key keyboard with Fn keys positioned horizontally, Model 2000,
copied by IBM PS/2 in 1987. This layout was designed by the Scripsit
developers, although the mechanical people shrunk it horizontally
before release. Tandy abandoned this layout in 1985, starting with the
Tandy 1200 and then the 3000, just to be compatible with IBMs vertical
10Fn key layout they had used since the original PC. Two years later,
Tandy had to return to the 12-key horizontal system because IBM PS/2s
had switched to the old Model 2000 design. Today, all computer makers
copy the 2000 keyboard layout.
o First small computer system with 1.2Meg floppies, Model 16, 1982, copied
by IBM PC/AT in 1984.
o First small computers able to run operating systems from several different
vendors, Model I, Model III and Color computer. Vendors include
Tandy, Logical Systems, Apparat, DR, Microware and others. IBM had
multi-vendor support in 1984, when XENIX-86 was released by SCO for
the PC/XT. Apple never had multi-vendor OS support.
o First multiplatform word processing system available (running on Z-80,
68000 XENIX and 8086 MSDOS systems), Scripsit 1980/1983/1986.
(Scripsit runs todays on several mainframe systems as well.)
o First SCSI driver for any device written for a multi-tasking/multi-user
operating system 68000 XENIX, 1984. Other makers developed theirs
in 1986 or later.
o Inexpensive networking and networking software for schools and other small
computer users, Network I, 1979, Network III 1981, Network IV, 1983,
Arcnet, 1983. IBM copied ARCNET by offering SNA in 1983, but it
wasn't cheap.
o Shared hard disk and printer networks for small computers, Network IV, 1983,
copied by IBM 1983.
o Graphics upgrades for small computers, Model III 1979, concept copied by
IBM PC CGA in 1982 (CGA not yet available at PC introduction in 1981)
o First 16-bit (68000) small computer system, Model 16, 1982, copied by
IBM PC/AT in 1984.
o First multi-processor small computer system, Model 16, 1982, copied by
IBM PS/2 in 1989.
o First protected multi-user small computer system under $5,000, Model 16,
with the introduction of XENIX, December 1982. IBM copies with
XENIX-286 for PC/AT in late 1985.
o First portable laptop computer, Model 100, January 1983, copied by IBM and
others in late 1983.
o First computer maker to use Intel 80x186 processor, Model 2000, 1983.
Ok, it wasn't a big hit, but they were first and that knowledge made
Intel change the under-design 80x286 to avoid the problems the 186
had.
o First computer with built-in digital audio system, Tandy 1000 SL/TL, 1988.
SoundBlaster cards become popular 18 months later.
o First company to give away software libraries to operate their digital
sound systems to anybody who wanted them, 1988.
o First retail/mass-sold personal computer system for under $700, the original
TRS-80, 1977. This price barrier is beaten by Tandy over and over
again over the years, but IBM never comes close until 1983, with the
IBM PCJr, and it can only manage a $1199 price for the stripped unit.
o First computer with bundled/integrated modem, Model 100, 1983.
o First computer maker to offer 16-color 640x400 graphics, Model 2000, 1983,
better than IBMs 320x200x4 CGA. IBM responded with EGA 640x350x16,
1984.
o First and only company to ever use Microsoft Modular Windows, 1992. Yeah,
this was a big disaster and Microsoft now denies it was ever a
product, but Tandy did it first.
There are probably a hundred other things I didn't think of at this moment.
The real death of innovation in the personal computer industry began in August
of 1981 when the IBM PC came out. It just took several years for innovation
of an obvious type to be completely killed at Tandy.
: Their biggest impact, but their stupidest decision, was the model 100-line.
: Instead of expanding on the idea of the portable computer without moving
: parts, they restricted it to a mere word-processor. Imagine a "docking
: station" for the T102, where you clip it in and you have floppy-capacity
: (even HD, maybe) external monitor and so on.
The VDI was a docking station, although not as streamlined as the ones you
are used to today. Hey, when you are first, you don't always get it right.
Plans for hard disk support were underway, but with the arrival of small
winchester drives that could go in the laptop itself, the product was
not developed further. The VDI did support an external monitor as well as
floppies and other goodies. There was also pressure from Microsoft to develop
laptop designs that could run MS-DOS, even back in 1985.
I don't even consider this (the VDI) in the top 100 "stupid decisions" Tandy
made. There were plenty of other boners. "VIS" probably ranks as #1, simply
because it was a major factor in Tandys decision to go out of the computer
business. Several management selections probably fill-out the bulk of the
top ten.
Frank Durda IV - only these addresses work:|"If you make it go out of focus,
<uhclem.dec98%nemesis.lonestar.org> | it will become clearer."
| - A real Tandy Manager speaks on
This Anti-spam address expires Dec. 31st | problem solving.
Copr. 1998, ask before reprinting.
Gary K
II mother board, quad uart boards, and the 300 baud modem cards? or the one
that was supposed to go against the phillips game player system? (It always
annoyed me they used the same acronym twice. 'Course, they've used the same
catalog numbers twice for totally different products also.)
I'll vouch for Frank, he knows more about Tandy Computers than I, and I
worked in various parts of Tandy Computer support in Ft Worth for almost 20
years (Ok, I started before the Model I came out, and left after the
Computer City build your own program started.)
I forgot about the 2000 Xenix episode, just a well! Vividly remember the
2K blue wires, RFI mods, bad Power Supply pots (or was it the torque seal?),
WD1010 alignments and dead bug mods. (Did Frank have to work at TBP at
nights?) 2K would have had a chance if the PC AT had not come out 6 months
later...
Now what's this about the ROM cartridge slots on the Tandy 1000? :}
Gary K
the TyRannoSaurus wrote in message
<01be1607$cc44ddc0$Loca...@news.tornado.be.tornado.be>...
>Hi,
>
>After reading this, I concede that you are far better informed about Tandy
>history than I am. (Wich is not that difficult) Have you been working for
>them ?
>
>Just a few remarks:
>
>> : A superior but incompatible system with graphics could have done it.
>> : Imagine a version of Unix/Xenix on the T2000, with an X-windows
>interface
>> : on it... It could have done what Linux is doing today, and 15 years
>before
>> : !
>>
>> Uh, XENIX was sold for the Tandy 2000.
>
>Never got to the European catalogs :)
>
>> It was a failure. The 2000 hardware
>> was just too unreliable (over two hundred "blue wire" modifications
>> on the final units shipped),
>
>That's odd. I never found the 2000 unreliable...
>
>> and the graphics hardware was designed in such
>> a way that speed was never a possibility.
>
>This is WAY before the graphic accelerator cards, of course.
>
>> Having 100% of the programmers who
>> developed the machine quit within two months of its introduction didn't
>help
>> either. Add this to the 80x186 with no protected memory concept,
>
>Isn't that a function of the operating system ? There is/was even a version
>of Unix on a 8086 (Minix ?)
>
>> Intel didn't help by deliberately re-using I/O
>> ports for their integrated hardware that were already "known" ports in
>the
>> IBM/PC world.
>
>Yeah, that was the reason why you couldn't use a standard DOS on a 80186.
>
>> Microsoft compounded the error by using INTs and other areas
>> that Intel had clearly marked as "reserved", so when the 80x186 came
>along,
>> things broke.
>
>Do you mean, Microsoft XENIX or MS-DOS ?
>
>> Most people didn't notice all the things breaking until the
>> 80x286, since only Tandy and Olivetti ventured into the 80x186 universe
>as
>> a PC platform processor.
>
>So did Philips an Apricot. (Dutch and English PC-makers, in those days)
>
>> Also, MIT X was incredibly immature at that time, and only ran on two
>> platforms, neither of which were Intel-based.
>
>I thought that X was older than that. Thanks for the information.
>
>> The Sun SPARC was just
>> coming on the market, just to give you an idea of where we were in
>history.
>
>Yeah.
>
>> Even the so-called Windows of the day (you can see the Model 2000 ad
>running
>> Windows on the back of a Byte magazine starring Bill Gates) shows that
>> Windows 1.x was text-based in those days, but was in COLOR.
>
>What issue of Byte ?
>
>> : For the second they had 2 possibilities: the TRS-80 line and the
>Coco-line.
>>
>> Despite this, Model 4D were still being made as
>> late as early 1987, and being sold for no less than 4X the manufacturing
>cost.
>
>80micro called the profit margin on the 4D ridiculous, or something.
>
>> Again, Tandys marketing decided to kill the Coco when the 1000s got hot,
>and
>> the designer who was doing the work on the most advanced Coco ever had
>his
>> project canceled when it was nearly done and he quit.
>
>Who was it ?
>
>Didn't he sell his project to someone else or got into business for himself
>?
>
>> Two years later he
>> came back to and did a cost-reduced version of the Coco2,
>
>The Coco 3 ?
>
>> the machine that to
>> die, but it also was not advertised and none of the fancy stuff from the
>> Deluxe Color Computer ever made it to the stores.
>
>Sad.
>
>> : "Don't try to be an innovator ...
>>
>> Look at what Tandy did first:
>>
>> o Double-Sided floppies 1982, copied by IBM PC/AT in 1984
>
>The PC also had DS drives. Supported by DOS 2.0. The first DOS for the AT
>was 3.20
>
>> o First computer maker to bundle application software with new systems,
>1983.
>> Apple copied in 1984. IBM finally did in 1989.
>
>Huh ? Do you mean the 100 or the 1000 ?
>
>> o 12-function key keyboard with Fn keys positioned horizontally, Model
>2000,
>> copied by IBM PS/2 in 1987.
>
>No: the AT and PC/XT could be ordered with 12-function keyboards. (AFAIK:
>1984) But they were, so you are right anyway, later to come out with them
>than Tandy with the T2000 keyboard.
>
>> Two years later,
>> Tandy had to return to the 12-key horizontal system because IBM PS/2s
>> had switched to the old Model 2000 design.
>
>They continued to supply all 1000-series computers with the 12-function
>keyboard, but they switched to an IBM layout with the SL/TL-series. And the
>3000 NL also had an extended layout.
>
>> Today, all computer makers
>> copy the 2000 keyboard layout.
>
>The IBM AT layout...
>
>> o First small computers able to run operating systems from several
>different
>> vendors, Model I, Model III and Color computer.
>
>But there was never any support from Tandy for these vendors. Only Logocal
>Systems was supported on the model III for their harddisk-OS, and later
>taken in as writer for TRSDOS 6.x.
>
>Actually, early-on, Tandy NEVER supported third-party software.
>
>> o Graphics upgrades for small computers, Model III 1979, concept copied
>by
>> IBM PC CGA in 1982 (CGA not yet available at PC introduction in 1981)
>
>Hmmm. The model III graphics card was a graphics add-on to the computer,
>while the CGA was a seperate video-card, supporting both text and graphics.
>
>> o First 16-bit (68000) small computer system, Model 16, 1982, copied by
>> IBM PC/AT in 1984.
>
>Wasn't the IBM PC 16-bit also ?
>
>> o First multi-processor small computer system, Model 16, 1982, copied by
>> IBM PS/2 in 1989.
>
>Right ! How many people know that the 16 had 2 CPU's, a 68000 AND a Z-80 ?
>
>> o First portable laptop computer, Model 100, January 1983, copied by IBM
>and
>> others in late 1983.
>
>Epson had a small portable computer before that, but the sceen was so tiny
>(4x20 or so) that you could hardly use it. You could say that the 100 was
>the first real multifunctional useful laptop.
>
>> o First computer with bundled/integrated modem, Model 100, 1983.
>
>The model 100 is their biggest achievement, I think...
>
>> o First computer maker to offer 16-color 640x400 graphics, Model 2000,
>1983,
>> better than IBMs 320x200x4 CGA. IBM responded with EGA 640x350x16,
>> 1984.
>
>Originally, EGA was 64 K and allowed only 640*350*4 !
>
>> There are probably a hundred other things I didn't think of at this
>moment.
>> The real death of innovation in the personal computer industry began in
>August
>> of 1981 when the IBM PC came out.
>
>Hear, hear !!! Microcomputer-architecture is Dullsville compared to 20
>years ago.
>
>> : Their biggest impact, but their stupidest decision, was the model
>100-line.
>> : Instead of expanding on the idea of the portable computer without
>moving
>> : parts, they restricted it to a mere word-processor.
>
>The WP-2.
>
>> : Imagine a "docking
>> : station" for the T102, where you clip it in and you have
>floppy-capacity
>> : (even HD, maybe) external monitor and so on.
>>
>> The VDI was a docking station, although not as streamlined as the ones
>you
>> are used to today.
>
>Why didn't they re-design it later on ?
>
>> The VDI did support an external monitor as well as
>> floppies and other goodies.
>
>I know. It was virtually a computer on it's own.
>
>> There was also pressure from Microsoft to develop
>> laptop designs that could run MS-DOS, even back in 1985.
>
>Anybody around here who's still saying the DOJ's case against MS is
>unjustified ?
>
>> I don't even consider this (the VDI) in the top 100 "stupid decisions"
>Tandy
>> made.
>
>I was referring to axing the T100/102-line. Their success today still
>proves there is a market for it. Nowadays it could be produced with a
>640x200 colorscreen and use PCMCIA-cards as storage for programs and data,
>and still be compatible with the original 100.
>
>> There were plenty of other boners. "VIS" probably ranks as #1, simply
>> because it was a major factor in Tandys decision to go out of the
>computer
>> business.
>
>I don't get it. WHat is VIS ?
>
>> Frank Durda IV
the TyRannoSaurus
After reading this, I concede that you are far better informed about Tandy
history than I am. (Wich is not that difficult) Have you been working for
them ?
Just a few remarks:
> : A superior but incompatible system with graphics could have done it.
> : Imagine a version of Unix/Xenix on the T2000, with an X-windows
interface
> : on it... It could have done what Linux is doing today, and 15 years
before
> : !
>
> Uh, XENIX was sold for the Tandy 2000.
Never got to the European catalogs :)
> It was a failure. The 2000 hardware
> was just too unreliable (over two hundred "blue wire" modifications
> on the final units shipped),
That's odd. I never found the 2000 unreliable...
> and the graphics hardware was designed in such
> a way that speed was never a possibility.
This is WAY before the graphic accelerator cards, of course.
> Having 100% of the programmers who
> developed the machine quit within two months of its introduction didn't
help
> either. Add this to the 80x186 with no protected memory concept,
Isn't that a function of the operating system ? There is/was even a version
of Unix on a 8086 (Minix ?)
> Intel didn't help by deliberately re-using I/O
> ports for their integrated hardware that were already "known" ports in
the
> IBM/PC world.
Yeah, that was the reason why you couldn't use a standard DOS on a 80186.
> Microsoft compounded the error by using INTs and other areas
> that Intel had clearly marked as "reserved", so when the 80x186 came
along,
> things broke.
Do you mean, Microsoft XENIX or MS-DOS ?
> Most people didn't notice all the things breaking until the
> 80x286, since only Tandy and Olivetti ventured into the 80x186 universe
as
> a PC platform processor.
So did Philips an Apricot. (Dutch and English PC-makers, in those days)
> Also, MIT X was incredibly immature at that time, and only ran on two
> platforms, neither of which were Intel-based.
I thought that X was older than that. Thanks for the information.
> The Sun SPARC was just
> coming on the market, just to give you an idea of where we were in
history.
Yeah.
> Even the so-called Windows of the day (you can see the Model 2000 ad
running
> Windows on the back of a Byte magazine starring Bill Gates) shows that
> Windows 1.x was text-based in those days, but was in COLOR.
What issue of Byte ?
> : For the second they had 2 possibilities: the TRS-80 line and the
Coco-line.
>
> Despite this, Model 4D were still being made as
> late as early 1987, and being sold for no less than 4X the manufacturing
cost.
80micro called the profit margin on the 4D ridiculous, or something.
> Again, Tandys marketing decided to kill the Coco when the 1000s got hot,
and
> the designer who was doing the work on the most advanced Coco ever had
his
> project canceled when it was nearly done and he quit.
Who was it ?
Didn't he sell his project to someone else or got into business for himself
?
> Two years later he
> came back to and did a cost-reduced version of the Coco2,
The Coco 3 ?
> the machine that to
> die, but it also was not advertised and none of the fancy stuff from the
> Deluxe Color Computer ever made it to the stores.
Sad.
> : "Don't try to be an innovator ...
>
> Look at what Tandy did first:
>
> o Double-Sided floppies 1982, copied by IBM PC/AT in 1984
The PC also had DS drives. Supported by DOS 2.0. The first DOS for the AT
was 3.20
> o First computer maker to bundle application software with new systems,
1983.
> Apple copied in 1984. IBM finally did in 1989.
Huh ? Do you mean the 100 or the 1000 ?
> o 12-function key keyboard with Fn keys positioned horizontally, Model
2000,
> copied by IBM PS/2 in 1987.
No: the AT and PC/XT could be ordered with 12-function keyboards. (AFAIK:
1984) But they were, so you are right anyway, later to come out with them
than Tandy with the T2000 keyboard.
> Two years later,
> Tandy had to return to the 12-key horizontal system because IBM PS/2s
> had switched to the old Model 2000 design.
They continued to supply all 1000-series computers with the 12-function
keyboard, but they switched to an IBM layout with the SL/TL-series. And the
3000 NL also had an extended layout.
> Today, all computer makers
> copy the 2000 keyboard layout.
The IBM AT layout...
> o First small computers able to run operating systems from several
different
> vendors, Model I, Model III and Color computer.
But there was never any support from Tandy for these vendors. Only Logocal
Systems was supported on the model III for their harddisk-OS, and later
taken in as writer for TRSDOS 6.x.
Actually, early-on, Tandy NEVER supported third-party software.
> o Graphics upgrades for small computers, Model III 1979, concept copied
by
> IBM PC CGA in 1982 (CGA not yet available at PC introduction in 1981)
Hmmm. The model III graphics card was a graphics add-on to the computer,
while the CGA was a seperate video-card, supporting both text and graphics.
> o First 16-bit (68000) small computer system, Model 16, 1982, copied by
> IBM PC/AT in 1984.
Wasn't the IBM PC 16-bit also ?
> o First multi-processor small computer system, Model 16, 1982, copied by
> IBM PS/2 in 1989.
Right ! How many people know that the 16 had 2 CPU's, a 68000 AND a Z-80 ?
> o First portable laptop computer, Model 100, January 1983, copied by IBM
and
> others in late 1983.
Epson had a small portable computer before that, but the sceen was so tiny
(4x20 or so) that you could hardly use it. You could say that the 100 was
the first real multifunctional useful laptop.
> o First computer with bundled/integrated modem, Model 100, 1983.
The model 100 is their biggest achievement, I think...
> o First computer maker to offer 16-color 640x400 graphics, Model 2000,
1983,
> better than IBMs 320x200x4 CGA. IBM responded with EGA 640x350x16,
> 1984.
Originally, EGA was 64 K and allowed only 640*350*4 !
> There are probably a hundred other things I didn't think of at this
moment.
> The real death of innovation in the personal computer industry began in
August
> of 1981 when the IBM PC came out.
Hear, hear !!! Microcomputer-architecture is Dullsville compared to 20
years ago.
> : Their biggest impact, but their stupidest decision, was the model
100-line.
> : Instead of expanding on the idea of the portable computer without
moving
> : parts, they restricted it to a mere word-processor.
The WP-2.
> : Imagine a "docking
> : station" for the T102, where you clip it in and you have
floppy-capacity
> : (even HD, maybe) external monitor and so on.
>
> The VDI was a docking station, although not as streamlined as the ones
you
> are used to today.
Why didn't they re-design it later on ?
> The VDI did support an external monitor as well as
> floppies and other goodies.
I know. It was virtually a computer on it's own.
> There was also pressure from Microsoft to develop
> laptop designs that could run MS-DOS, even back in 1985.
Anybody around here who's still saying the DOJ's case against MS is
unjustified ?
> I don't even consider this (the VDI) in the top 100 "stupid decisions"
Tandy
> made.
I was referring to axing the T100/102-line. Their success today still
proves there is a market for it. Nowadays it could be produced with a
640x200 colorscreen and use PCMCIA-cards as storage for programs and data,
and still be compatible with the original 100.
> There were plenty of other boners. "VIS" probably ranks as #1, simply
> because it was a major factor in Tandys decision to go out of the
computer
> business.
I don't get it. WHat is VIS ?
Frank Durda IV
: I forgot about the 2000 Xenix episode, just a[s] well! Vividly remember the
: 2K blue wires, RFI mods, bad Power Supply pots (or was it the torque seal?),
: WD1010 alignments and dead bug mods.
It was fingernail polish or "Loc-tite". Either one ate the guts out of the
pots and the power supply voltages would suddenly shoot way up or down,
usually with fatal results.
: (Did Frank have to work at TBP at nights?)
Yes, as a matter of fact, several times. The longest was debugging the fact
that Texas Instruments had recalled two months production of certain 74ALS
parts and our factory had decided that pulling the defective parts off the
boards they had already built AND returning the unused stock would be too
much bother, so they continued stuffing the bad parts and didn't tell anyone
in R&D about the recall notice, and then they wondered why all the machines
paniced when you tried to run XENIX on them... Everybody was soldering
for a few days...
It even continued after we became AST. Five days after they let the entire
hardware development staff go, including the modem designers, the factory was
shut-down because of a modem problem, and I ended-up living out there for a
week trying to figure out why the system could not find the modems in the
new product line computers they started building that week.
Hi, Gary! (Back at Tandy, Gary used to try to explain to the service
centers how to service the stuff the factories were producing, that didn't
necessarily match what we designed. We'd find this out about two hours after
Gary did.
If I remember right, Gary also tried to get the service center in Australia
to quit opening the hard disk bubbles themselves in a dusty warehouse they
used for a repair centre. We never got them to stop doing that, but they
did send photos of a thing they built that used wood clothesline clips to
hold the hard disk heads out of the way while they had the platters out...)
Frank Durda IV - only these addresses work:|"Make it Green, Glow-in-the-Dark,
<uhclem.dec98%nemesis.lonestar.org> | and Submersible, and you'll have
| the computer that everybody will
This Anti-spam address expires Dec. 31st | buy." - Tandy Marketing.
Frank Durda IV
[2]Hi,
[2]After reading this, I concede that you are far better informed about Tandy
[2]history than I am. (Wich is not that difficult) Have you been working for
[2]them ?
I worked in Tandy Research and Development from 1982 until it was sold in 1993
and worked on hundreds of projects, including Model 16, Model III, Model 100,
MC-10, Model 4, Model 4P, LDOS, TRSDOS-6, TRSDOS-6 Hard disk, TRSDOS-6
international, Model 2000 BIOS (finished bits original programmers left
undone), fast graphics for Model 2000 BASIC, re-wrote all Z-80 drivers and
utilities for XENIX, ported drivers to 80286 for XENIX-286 and later XENIX-386,
finished the digital sound and music programs for the 1000 SL/TL, wrote
the streaming audio engine for the Digital Sound Toolkit (used by Disney and
other game vendors), set-up Tandys CD-ROM mastering system (in 1991, this
was hard and expensive), spent a year on VIS, worked on voice/speakerphone
modem designs, etc. I even designed the network and office numbering system
in the Tandy Technology Square facility, a seven story, 210,000ft building
where R&D moved in 1991. I did work on the 3000, 4000, and all the 1000
series units that had digital sound, 2500, and so on. Then I did modems.
The reason the 4P ROM has international messages in there is because I
personally pushed for them, even though we had a devil of a time getting
anyone to agree on what they should say...
[0]the TyRannoSaurus (Ja...@tornado.be.nospamplease) originally wrote:
[0]A superior but incompatible system with graphics could have done it.
[0]Imagine a version of Unix/Xenix on the T2000, with an X-windows interface
[0]on it... It could have done what Linux is doing today, and 15 years before
[0]!
[1]I replied with:
[1]Uh, XENIX was sold for the Tandy 2000.
[2]the TyRannoSaurus (Ja...@tornado.be.nospamplease) responded with:
[2]Never got to the European catalogs :)
I think it only appeared in one of the six-month catalogs in the US.
You could order it though for quite a while, and I know we had a cement
company in England that was using it, as they were always having trouble.
(Cause: cement dust inside the machine would grab water out of the air
causing random strangeness and very scratched backup media.)
[1]It was a failure. The 2000 hardware
[1]was just too unreliable (over two hundred "blue wire" modifications
[1]on the final units shipped),
[2]That's odd. I never found the 2000 unreliable...
That's what the hardware people said, until we put XENIX on it. Then
it would crash constantly. The fact that XENIX was doing lots more stuff
and would complain if things went strange (DOS would keep going or just
hang) meant that we found all sorts of timing problems. Trying to execute
applications while devices were generating interrupts or using DMA was
fatal until you applied a few dozen modifications. DOS is completely
single-tasking, although we later found the DOS background print spooler
could also cause similar crashes on the unmodified 2000.
Over the 68000, 286, 386 and 486 years, we would drag out a floppy that had
written on it "EVILDISK", and start it on the new machine under XENIX. Newly
developed machines would invariably crash consistently and the hardware people
would spend months looking for obscure system timing problems. The factory
also used EVILDISK to shake-out flaky machines coming off the line. EVILDISK
was just a shell script that did copies and compares of files simultaneously.
The hardware people hated the disk, because the same machine would run DOS
just fine. Of course, their definition was frequently "boot and get to a DOS
prompt and do a DIR command without crashing". Later we found that machines
that failed on EVILDISK also would not pass the Novell certification suite,
confirming what XENIX had been saying all along - something is wrong with the
hardware. If EVILDISK survived, the machine would almost always pass the
costly Novell certification test. Since the QA group got complete say over
what tools they used to test new hardware, EVILDISK was one of the first
things tried.
[1]and the graphics hardware was designed in such
[1]a way that speed was never a possibility.
[2]This is WAY before the graphic accelerator cards, of course.
No, I was comparing CGA to the 2000, competiting technology of the day.
To update all the pixels that made up the 320x200 video screen on CGA, an
application would have to write to 64,000 bytes. On the 2000, the 640x400
screen was four times larger, but you had to write to TWELVE times as many
bytes. This was because of the way the video was organized, and unless you
wanted the screen to flash through several other colors before you got to
the color you wanted, you had to this extremely expensive operation.
To update a single pixel on CGA, EGA or VGA in the most common modes, you
would have to write a single byte, or at worst, read a byte, update part of
the value in a register (AND/OR opcodes), and write it back. Simple,
straightforward, and fast.
To update a single byte on the 2000, you would have to OUT a byte to a
port to select plane 1, read the byte, update the one bit for the pixel
you wanted to change (AND/OR opcodes), write the byte back, OUT a byte to a
port to select plane 2, read the byte, update the one bit for the pixel you
wanted to change (AND/OR), write that, OUT a byte to a port to select plane 3,
read, update, write, OUT a byte to select plane 4, read, update, write
and FINALLY you are done. That was for ONE pixel. If you are doing
anything other than painting a rectangular-shapped area (including the entire
screen) the same color, the 2000 video hardware design resulted in insanely
slow video updates. Drawing characters or lines that didn't go straight up
or horizontal were extremely expensive and impossible to optimize.
The bit-per-plane design of the 2000 graphics system is what makes graphical
operations so slow.
[1]Having 100% of the programmers who
[1]developed the machine quit within two months of its introduction didn't
[1]help either. Add this to the 80186 with no protected memory concept,
[2]Isn't that a function of the operating system ? There is/was even a version
[2]of Unix on a 8086 (Minix ?)
What good is having the concept in the OS if the hardware can't enforce it?
Any application running under MINUX can crash the box, just like a
badly-behaved DOS application running in a "protected" DOS-box under Windows
can make that system die, or corrupt it. Since applications can't be
trusted to stay within the limits the OS may specify, the hardware protection
is required to give what the OS wants some teeth. Multi-user systems simply
cannot run predictably and safely on systems without protected hardware
capabilities. The 68000 didn't have it either, but the Model 16/6000 had
circuitry that implemented protected memory regions for the processor.
[1]Intel didn't help by deliberately re-using I/O
[1]ports for their integrated hardware that were already "known" ports in the
[1]IBM/PC world.
[2]Yeah, that was the reason why you couldn't use a standard DOS on a 80186.
Actually, you could, you just had to pick the right version. The 2000
had a Microsoft-specified BIOS system that was not compatible with what IBM
was doing in the PCs, which is the main reason you can't go get DOS 5 or
whatever and make it go on the 2000.
[1]Microsoft compounded the error by using INTs and other areas
[1]that Intel had clearly marked as "reserved", so when the 80186 came
[1]along,
[1]things broke.
[2]Do you mean, Microsoft XENIX or MS-DOS ?
MS-DOS used forbidden things, and it was discussed in Dvoraks columns
quite frequently in 1983/1984.
[1]Also, MIT X was incredibly immature at that time, and only ran on two
[1]platforms, neither of which were Intel-based.
[2]I thought that X was older than that. Thanks for the information.
Parts of X can be traced back to the Multics project, circa 1970. I actually
have a document from that year where people at MIT were using the term "GUI"
to describe what they were doing. However, an Intel port didn't appear
until the mid-eighties, in the form of Motif.
[1]Even the so-called Windows of the day (you can see the Model 2000 ad running
[1]Windows on the back of a Byte magazine starring Bill Gates) shows that
[1]Windows 1.x was text-based in those days, but was in COLOR.
[2]What issue of Byte ?
Sorry, don't remember off the top of my head. It was the back-cover, as
I had a photocopy of it on my cube for years, with all those stickers you
get from Ziff Davis ads stuck all over it, ie "FREE GIFT" stuck on Bills
shirt, etc. I am pretty sure it was in 1984.
[1]Despite this, Model 4D were still being made as
[1]late as early 1987, and being sold for no less than 4X the manufacturing
[1]cost.
[2]80micro called the profit margin on the 4D ridiculous, or something.
Yeah, well you should have seen Wayne Greens advertising rates if you want to
see high profit margins. Kettle black.
[1]Again, Tandys marketing decided to kill the Coco when the 1000s got hot, and
[1]the designer who was doing the work on the most advanced Coco ever had his
[1]project canceled when it was nearly done and he quit.
[2]Who was it ?
He's around, he can speak up if he wants. He lives at AMD these days.
[2]Didn't he sell his project to someone else or got into business for himself
[2]?
There was this pesky thing about agreeing to not compete against Tandy
for N years after you left that Tandy wanted everybody to sign.
Tandy would have sued just to prove a point. Such a project would never
would have happened.
[1]Two years later he
[1]came back to and did a cost-reduced version of the Coco2,
[2]The Coco 3 ?
Yes.
[1]o Double-Sided floppies 1982, copied by IBM PC/AT in 1984
[2]The PC also had DS drives. Supported by DOS 2.0. The first DOS for the AT
[2]was 3.20
Let me clarify - I meant STANDARD EQUIPMENT, not optional. IBM didn't have
them standard until the PC/AT. 360K floppies were an option on the original
PC and XT.
[1]o First computer maker to bundle application software with new systems, 1983.
[1] Apple copied in 1984. IBM finally did in 1989.
[2]Huh ? Do you mean the 100 or the 1000 ?
Following the "big launch" of January 1983, every subsequent machine had at
least one application bundled. Most were pretty useless, but the machines
did come with application software. The Model 4 was first, and came with a
little mailing list program that was written in BASIC. The Model 100 was
second. Then something was included on the 2000 (pie chart BASIC demo?).
Then came Deskmate, on the Model 4 and then to PCs and beyond.
[1] Today, all computer makers
[1] copy the 2000 keyboard layout.
[2]The IBM AT layout...
Nope, the ORIGINAL IBM AT layout had F-keys in a vertical placement on the
left side of the keyboard, just like the XT and original PC. IBM switched to
the 2000 layout a year or so later, definitely by the time the PS/2 hit the
US market.
[1]o First small computers able to run operating systems from several different
[1] vendors, Model I, Model III and Color computer.
[2]But there was never any support from Tandy for these vendors. Only Logocal
[2]Systems was supported on the model III for their harddisk-OS, and later
[2]taken in as writer for TRSDOS 6.x.
Your point is? Mine was that Tandy platforms had OS choices years before
any other hardware maker did, regardless of whether they were supported by
the hardware maker.
[2]Actually, early-on, Tandy NEVER supported third-party software.
No one said they did. Plus, I was talking about third-party OS availability,
not applications.
Tandy *did* license a lot of third-party software, and proceded to sell and
support it. TRSDOS 6 is just one example. DoubleDuty, 13 Ghosts, ...
[1]o Graphics upgrades for small computers, Model III 1979, concept copied by
[1] IBM PC CGA in 1982 (CGA not yet available at PC introduction in 1981)
[2]Hmmm. The model III graphics card was a graphics add-on to the computer,
[2]while the CGA was a seperate video-card, supporting both text and graphics.
Again, your point is? Add-on graphics is what I said was the feature.
IBMs upgrade was from monochrome to color, so long as cyan and magenta
were two of the colors you wanted... Tandys was from low-res block graphics
(included with the system) to pixel-controllable graphics.
[1]o First 16-bit (68000) small computer system, Model 16, 1982, copied by
[1] IBM PC/AT in 1984.
[2]Wasn't the IBM PC 16-bit also ?
I was comparing external data paths, and the 8088 had 8-bit external paths.
If you want to compare processor register size, then Tandy had a 32-bit
computer on the market in 1982, while IBM had a 16-bit. Tandy was still
first with the bigger system. Measure it either way.
The 16 supported 768K of RAM while the original PC only listed an
upgrade path to a maximum of 256K.
[1]o First multi-processor small computer system, Model 16, 1982, copied by
[1] IBM PS/2 in 1989.
[2]Right ! How many people know that the 16 had 2 CPU's, a 68000 AND a Z-80 ?
Did you know some other party built a board for the Model II/12/16/6000
with an 8088/8087 on it? Again, Tandy had the choices first.
[2]Why didn't they re-design it [100 VDI] later on ?
The industry moved forward, Tandy bought GRiD and phased-out the non-PC
laptops very quickly after that.
[1]There was also pressure from Microsoft to develop
[1]laptop designs that could run MS-DOS, even back in 1985.
[2]Anybody around here who's still saying the DOJ's case against MS is
[2]unjustified ?
Read about VIS, below.
[1]I don't even consider this (the VDI) in the top 100 "stupid decisions"
[1]Tandy made.
[2]I was referring to axing the T100/102-line. Their success today still
[2]proves there is a market for it. Nowadays it could be produced with a
[2]640x200 colorscreen and use PCMCIA-cards as storage for programs and data,
[2]and still be compatible with the original 100.
Not if you can't license the code inside from Microsoft anymore.
[1]There were plenty of other boners. "VIS" probably ranks as #1, simply
[1]because it was a major factor in Tandys decision to go out of the
[1]computer business.
[2]I don't get it. WHat is VIS ?
Ah, since you were in Europe, you were spared VIS. I am speaking of the 1992
version of VIS. Tandy re-used the name VIS three times for products between
1977 and 1992. The final VIS was a joint product between Microsoft and
Tandy. Our job was to invest lots of time and money, and then turn away so
that Microsoft could stab us.
VIS (Mark III, code named Gryphon) was a TV-top system targeted to compete
with (and hopefully destroy) the Philips CD-Interactive system. Like CD-I,
it could function as an audio CD-player, but its main purpose was to run
software stored entirely on CD-ROM. The titles tended to be "edu-tainment",
games, talking books, etc.
The VIS platform consisted of a CD-ROM drive, 1Meg of ROM, 1Meg of RAM in a
PC-model, meaning you got 640K plus 64K High Memory Area usable RAM.
A SVGA video system output video in NTSC (switchable to PAL-I) on RF, RCA
and S-video outputs. The same RAMDAC present in the original Sensation!
computer (came out the same year) allowed for high-color YUV video modes
and other goodies. The unit also came with a slot for a Dallas Semiconductor
memory card, capable of storing application-specific stuff, like high-scores
in games, etc. The unit was shipped with a single wireless hand controller,
but supported two, or one wired and one wireless controller, or one wireless
and one PS/2 keyboard or PS/2 mouse (never documented). A slot in the back
supported a standard DB-9 serial port, and supposedly would support a MPEG-I
(MPEG-II had not yet arrived on the scene) video decoder for movie playback.
The base processor was a 80286 running at 12MHz. The sound hardware was
compatible with Adlib Gold (same as Sensation!), a company that went out of
business just a few months prior to the release of VIS/Sensation!
One application was embedded in VIS, and that was a CD-ROM player application.
The product came with two CD-ROMs. One was Comptons encyclopedia, a port
of the MPC version they had been selling for the past year. The other
disc was a sampler, mainly designed to run in the stores as a self-running
demonstrator, with animated characters that talked and talked until the store
staff turned the speakers off.
Due to Microsofts vision of what was later called "Windows everywhere", the
entire project strategy relied on a new product called "Modular Windows",
a version of Windows 3.1 that could run from ROM and could to a limited extent
have parts removed to make a smaller system. A ROM MS-DOS layer also existed
in the unit. The Microsoft code name for Modular Windows was "Haiku", and
some of the staff at Microsoft were photographed by some magazine wearing
shirts that had the code name on it. (Hi Mike, Victor and all!)
Microsofts plan all along was to argue that because the CD-Is OS-9 operating
system environment was "non-Windows", and such required a special development
environment, the lack of Windows was the reason there were few titles for CD-I
(but the titles they had were very good). Modular Windows would attack by
making it "simple" to port existing Windows 3.x applications to Modular
Windows platforms and overwhelm CD-I in the marketplace by having lots more
titles (150 were planned for the original roll-out - by comparison, CD-I had
50 after two years in production).
Of course, most of the multimedia applications for Windows were really
lousy at this time, with 75% of them being someone reading a childrens book
with click-n-talk illustrations. Microsoft and Tandy figured that they could
still wipe CD-I off the map by using quantity in place of quality, even
though the childrens reading book was less than 20% of the CD-I title catalog.
John Roach, CEO of Tandy, was clearly not getting the full picture on
what VIS was supposed to do, because he consistently indicated that he had
been told that we were building a system to compete against Nintendo.
In fact, on several occasions Roach asked the executive in charge of the
project if this "thing" was going to be better than Nintendo and was assured
that it would be.
Of course, anybody with an atom is computer savvy back then would know
instantly that a SVGA 12MHz 286 isn't going to run rings around a Nintendo
with a hardware video scrolling and other accelerations, nor will a bunch
of loosely written C code be faster than hand-coded machine code, which is
what the Nintendo and Sega games used in those game-cartridge days.
The other reality was that you just couldn't "easily port" Windows
applications to this platform. The memory constraints were extremely tight,
no way to swap, and a lot of programmers discovered for the first time how
sloppy their code really was. The "development" PC systems had more RAM
and concealed their coding sins until they got prototype hardware. The
worst problem of all was that the applications written for Windows default
640x480 VGA video looked absolutely awful on NTSC video, with massive
interlace flicker, despite MONTHS of warnings to people that they would have
to deal with this. They assumed we were making it sound worse than it was,
because Microsoft had promised it would easy. (Duh.) Applications used
fonts that were too small to not be fuzzy on TV, or small icons with several
sharply contrasting colors next to one another (a no-no in NTSC), still
picture detail exceeded NTSC bandwidth causing massive color flicker,
particularly those line-drawings like you find in dictionaries, such as
Comptons, which was the application that was to be bundled with the product.
Before release, several hundred photos were just dropped from Comptons,
rather than reprocess them.
Then we discovered why Nintendo/Sega had constantly moving backgrounds and
never used a solid white background: it avoids warping the steel shadow mask
present in most TV sets, which causes large purple and green splotches to
appear on TVs within minutes. But guess what Color the background was in
Windows 3.x? BRILLIANT WHITE! Since we had no multi-plane video support
like Nintendo or CD-I had plus the fact that Microsoft would not change to
a color other than bright white for the Windows background and control color,
the VIS system "avoided" the warping problem by lowering all video brightness
by 20%. It really didn't fix the problem, it just made it less obvious.
Oh, ALL the apps had to lower the pallete values themselves - the hardware
could not be changed at this late date to do it electronically.
The other major problem was that Modular Windows revealed the true evils
of the underlying Windows architecture, things that Microsoft had hidden with
programs like SMARTDRV, that could not exist on this platform. For example,
when Windows 3.x starts, it opens, reads and closes the SYSTEM.INI file
75 separate times. The big improvement between Windows 3.0 and 3.1 was the
addition of SMARTDRV, which cached SYSTEM.INI and WIN.INI to the exclusion
of all else just to make these re-reads go faster. Now, if you put
SYSTEM.INI on a CD-ROM, and you have no hardware cache in the drive and no
SMARTDRV buffering, things get real slow. Microsoft said there was no way to
fix this "architectual artifact", and Windows '98 still does it today, but
this inefficency is hidden away by the '98 equivalent of SMARTDRV.
The prime example of the sluggishness of Modular Windows was the game
Kings Quest V, which was ported to the Modular Windows platform. From
disc insertion until the point here the game asked "Have you played this game
before?", it took exactly FIVE MINUTES on the Modular Windows platform.
During those five minutes, you spend 4:30 looking at a Windows hourglass
logo and little else. It went out the door this way. A processor emulator
allowed us to see lots of Windows re-reads of SYSTEM.INI and other INI files
during this hourglass period, plus lots of DLL reloads due to tight memory.
By comparison, a port of the Sherlock Holmes games (DOS program that
ran in 320x200x256) was very good, because the video was run in a "tricked"
interlace mode, there was ZERO flicker, and the because DOS took about
150msec to boot, the game was running and talking within 12 seconds. This
really annoyed Microsoft as they didn't want *any* DOS applications on
the platform, but the DOS applications were the only ones that ran fast.
Even the CD-player application Microsoft provided was written in Windows and
took 20 seconds to boot from ROM. Duh.
On finding how slow Kings Quest was, Tandy considered not selling any product
that took more than 90 seconds to boot, and found they would have less than
ten titles (all DOS based) out of the original 60 if they imposed this limit.
This idea was abandoned.
VIS went on the US and Canada markets in November 1992. Exactly 256 VIS
systems were sold between Thanksgiving and Christmas in ALL of North America,
despite having two units each in all 7,000 stores. In January 1993, all work
on the VIS-II system was canceled, and many of the prople working on the
project were let go. Shortly thereafter, Microsoft publicly denied that
Modular Windows was ever a product, claiming it was just a "Concept".
Tandy didn't think that was funny, since we had about 50,000 of those
"concepts" sitting in stores and warehouses. VIS units eventually found
their way to Home Shopping Club and other fire-sale venues. Tandy lost
$75 million on VIS, as of July 1st 1993.
There is a lot more to the VIS story, like Microsoft deliberately breaking
Modular Windows compatiblity three days after the VIS ROMs were sent to
be masked. This forced third-party software makers to decide whether to
follow VIS or follow Modular Windows. Either path ended-up leading to death,
since Modular Windows wasn't even a supported (or listed) Microsoft product
in six months, and Tandy was selling the VIS units for $40 less than the
manufacturing price in April of 1993 to see if it was possible to sell any
units at all. Since third-party software makers did not pay Tandy royalties
(although that was the original plan for VIS), that loss was not made-up
elsewhere.
A lot of third-party software vendors that invested everything in VIS
went out of business or got out of the multi-media market. Curiously,
Microsoft has since brought to market games and other applications almost
identical to the best of those from these defunct companies...
Two years later, Microsoft introduced "Windows CE". Guess what? Yep,
that's Modular Windows with a few new gadgets, but all the old problems.
Frank Durda IV - only these addresses work:|"The Knights who say "LETNi"
<uhclem.dec98%nemesis.lonestar.org> | demand... A SEGMENT REGISTER!!!"
|"A what?"
This Anti-spam address expires Dec. 31st |"LETNi! LETNi! LETNi!" - 1983
Jeff Lemke
> made. There were plenty of other boners. "VIS" probably ranks as #1, simply
> because it was a major factor in Tandys decision to go out of the computer
> business. Several management selections probably fill-out the bulk of the
> top ten.
>
> Frank Durda IV - only these addresses work:|"If you make it go out of focus,
> <uhclem.dec98%nemesis.lonestar.org> | it will become clearer."
> | - A real Tandy Manager speaks on
> This Anti-spam address expires Dec. 31st | problem solving.
> Copr. 1998, ask before reprinting.
What's "VIS"?
Jeff Lemke
Leonard Erickson
> I forgot about the 2000 Xenix episode, just a well! Vividly remember the
> 2K blue wires, RFI mods, bad Power Supply pots (or was it the torque seal?),
> WD1010 alignments and dead bug mods. (Did Frank have to work at TBP at
> nights?) 2K would have had a chance if the PC AT had not come out 6 months
> later...
I recently picked up a couple of 2000 systems. I'm considering getting
*really* crazy and seeing if DR-DOS will port to it. :-)
--
Leonard Erickson (aka Shadow)
sha...@krypton.rain.com <--preferred
leo...@qiclab.scn.rain.com <--last resort
Leonard Erickson
>> either. Add this to the 80x186 with no protected memory concept,
>
> Isn't that a function of the operating system ? There is/was even a version
> of Unix on a 8086 (Minix ?)
The OS can only do so much. A *proper* multi-tasker requires a CPU that
has the *hardware* ability to restrict acces to ranges of ports and RAM.
>> o Graphics upgrades for small computers, Model III 1979, concept copied
> by
>> IBM PC CGA in 1982 (CGA not yet available at PC introduction in 1981)
>
> Hmmm. The model III graphics card was a graphics add-on to the computer,
> while the CGA was a seperate video-card, supporting both text and graphics.
I think he goofed on that one. The Model III came out in 79, but the
hi-res board wasn't available for at least a year or two after that.
The *Model II* hi-res board came out sooner.
>> o First 16-bit (68000) small computer system, Model 16, 1982, copied by
>> IBM PC/AT in 1984.
>
> Wasn't the IBM PC 16-bit also ?
Not really. By the then existing standard terminology, the PC was an
*8* bit system, because the 8088 only had an 8-bit bus (it takes *two*
read/write cycles to pass 16-bit data to or from the CPU).
The 8086 was the "same" chip, except it had a 16-bit bus, and thus
could run noticeably faster. But very few systems used it, because it
required more expensive chips to interface to the CPU.
>> o First computer maker to offer 16-color 640x400 graphics, Model 2000,
> 1983,
>> better than IBMs 320x200x4 CGA. IBM responded with EGA 640x350x16,
>> 1984.
>
> Originally, EGA was 64 K and allowed only 640*350*4 !
No, the original EGA did do 640x350x16. You just had to pay for the
extra RAM (which went on a daughterboard on the IBM EGA cards).
Leonard Erickson
> [2]Didn't he sell his project to someone else or got into business for
> himself
> [2]?
>
> There was this pesky thing about agreeing to not compete against Tandy
> for N years after you left that Tandy wanted everybody to sign.
> Tandy would have sued just to prove a point. Such a project would never
> would have happened.
Of course *now*, such contracts are generally held to be restraint of
trade IIRC. <sigh>
> [1] Today, all computer makers
> [1] copy the 2000 keyboard layout.
>
> [2]The IBM AT layout...
>
> Nope, the ORIGINAL IBM AT layout had F-keys in a vertical placement on the
> left side of the keyboard, just like the XT and original PC. IBM switched to
> the 2000 layout a year or so later, definitely by the time the PS/2 hit the
> US market.
Only if you are talking about the F-key placement. Otherwise there are
considerable differences between the 2000 layout and the "standard"
102-key layout.
> [1]o First small computers able to run operating systems from several
> different
> [1] vendors, Model I, Model III and Color computer.
>
> [2]But there was never any support from Tandy for these vendors. Only Logocal
> [2]Systems was supported on the model III for their harddisk-OS, and later
> [2]taken in as writer for TRSDOS 6.x.
>
> Your point is? Mine was that Tandy platforms had OS choices years before
> any other hardware maker did, regardless of whether they were supported by
> the hardware maker.
Well, the PC *did* have UCSD p-system and CP/M-86. But since you had to
buy them thru IBM, I guess that doesn't really count.
And in any case, given the *large* price differential between PC-DOS
($50?) and the other two ($150-200?), it's no wonder that PC-DOS was so
much more common.
Eric J. Korpela
Leonard Erickson <sha...@krypton.rain.com> wrote:
>The OS can only do so much. A *proper* multi-tasker requires a CPU that
>has the *hardware* ability to restrict acces to ranges of ports and RAM.
I dare you to say that in comp.sys.amiga.advocacy. :)
Eric
--
Eric Korpela | An object at rest can never be
kor...@ssl.berkeley.edu | stopped.
<a href="http://sag-www.ssl.berkeley.edu/~korpela">Click for home page.</a>
Leonard Erickson
>> Yes, I do have one controller card comming out of service. I'll have to
>> look to see if it's already promised. You would actually have to install
>> the hard disk external to the Model 3/4. I've never seen one done
>> internal, just by physical size.
>
> Oh, you could fit a HD internally in the model 4, no problem there. But
> where would you put the HD-controller board ? I haven't seen the IDE
> controller, sold by CN80, but the controller board for the old Tandy HD's
> was almost as big as the motherboard !
Well, I'd install a pair of 1/3rd height 5.25 drives. That'd fill the
lower bay. Then in the upper bay, put a couple of 3.5 drives on-edge,
That leaves room for a 3.5 IDE drive *and* more than enough room
"behind" them for the controller card.
Leonard Erickson
> In article <981123.183835...@krypton.rain.com>,
> Leonard Erickson <sha...@krypton.rain.com> wrote:
>>The OS can only do so much. A *proper* multi-tasker requires a CPU that
>>has the *hardware* ability to restrict acces to ranges of ports and RAM.
>
> I dare you to say that in comp.sys.amiga.advocacy. :)
Unless they are still using 68000s rather than the later 680x0 chips,
they've *got* such a CPU chips.
Eric J. Korpela
Leonard Erickson <sha...@krypton.rain.com> wrote:
>kor...@islay.ssl.berkeley.edu (Eric J. Korpela) writes:
>
>> In article <981123.183835...@krypton.rain.com>,
>> Leonard Erickson <sha...@krypton.rain.com> wrote:
>>>The OS can only do so much. A *proper* multi-tasker requires a CPU that
>>>has the *hardware* ability to restrict acces to ranges of ports and RAM.
>>
>> I dare you to say that in comp.sys.amiga.advocacy. :)
>
>Unless they are still using 68000s rather than the later 680x0 chips,
>they've *got* such a CPU chips.
Last I heard, (which I admit has been a while), AmigaDOS does not use
memory protection capabilities even on CPUs which have the capability.
In fact I seem to remember that later Amigas used a 68030 with the memory
protection capabilities removed.
The only Amiga I recall with protected memory was the Unix version.
Charles Stephens
LE> kor...@islay.ssl.berkeley.edu (Eric J. Korpela) writes:
>> In article <981123.183835...@krypton.rain.com>,
>> Leonard Erickson <sha...@krypton.rain.com> wrote:
>>> The OS can only do so much. A *proper* multi-tasker requires a CPU that
>>> has the *hardware* ability to restrict acces to ranges of ports and RAM.
>>
>> I dare you to say that in comp.sys.amiga.advocacy. :)
LE> Unless they are still using 68000s rather than the later 680x0 chips,
LE> they've *got* such a CPU chips.
There seems to be confusion about memory protection (and along with
that virtual memory) versus multitasking. Either can be present in a
system without requiring the other. Most UNIX variants used to run on
PDP's just fine with out a MMU.
Before the 68K MMU's were available, Sun designed their own. We ended
up just designing the whole CPU and out came SPARC.
cfs
--
Charles F. Stephens = cfs AT eng.sun.com
Software Psychic and Illuminary =
Solaris Network Sustaining = "We don't make mistakes, we make
Solaris Software = happy accidents."
Sun Microsystems, Inc. = -- Bob Ross
Menlo Park, California, USA =
Eric J. Korpela
Charles Stephens <dev...@dobbs.eng.sun.com> wrote:
>
>Before the 68K MMU's were available, Sun designed their own. We ended
>up just designing the whole CPU and out came SPARC.
For that matter, Tandy designed it's own MMU for the 68K, too. I have
to say that Sun did a much better job, though. Tandy's MMU pretty much
gave processes a separate (but limited) address space, but didn't support
paging, etc.
Frank Durda IV
: There seems to be confusion about memory protection (and along with
: that virtual memory) versus multitasking. Either can be present in a
: system without requiring the other. Most UNIX variants used to run on
: PDP's just fine with out a MMU.
It got off track. Originally I said that "MULTI-USER" systems were not
feasible on systems that could not offer protection from user A nuking
(or spying) on user B. Because I also mentioned "multi-tasking" nearby,
the two were apparently taken to be a package deal. You can have
multi-tasking without having the protections against other users that serious
multi-user systems demand. See Windows 3.x/'95/'98 and its blue-screens
for an example of "multi-tasking" without "multi-user" protection.
MMUs and/or limit/offset registers solved the problem of protecting user A
from the programs user B ran. Early examples of limit/offset registers were
the UNOS systems from Charles Rivers (1980?) and the Tandy Model 16 (1982),
both of which used a MC68000 processor that did have a concept of user and
supervisor mode, but no on-board circuitry to enforce user mode. This was
usually accomplished with offset registers that added to the memory address
specified by the CPU when in user mode, and limit registers that did a
comparison of a maximum memory location against the actual address being
accessed to see if it was within bounds.
Since the 68000 did not have the concept of restartable instructions, a
partially-completed execution that got a limit register trap could not
resume, so paging or growable stacks were not possible. The 68010 had
restartable instructions, as it saved more processor state on the supervisor
stack frame.
I forget when Sun came along, but I am fairly certain Charles Rivers was
doing M-T/M-U small microcomputer systems (not minis) long before anybody
else.
By the way, Tandy originally planned to ship UNOS in late 1982, and were only
a few weeks away from doing just that when Microsoft gave the Tandy Merchies
a deal they could not refuse. UNOS was dumped and Microsoft XENIX was put in
its place.
What you would consider "modern" paged/blocked memory protection systems
appear as far back as the DECsystem-10/DECsystem-20 era, using TOPS-10 and
TOPS-20, both derived from Multics work at MIT in the early 1970s. These
systems had some better virtual memory concepts than the VAX-11/780
did when it came out years later.
In the Intel universe, the 80286 was the first processor to provide
any sort of memory protection model.
Frank Durda IV - only these addresses work:|"The Knights who say "LETNi"
<uhclem.dec98%nemesis.lonestar.org> | demand... A SEGMENT REGISTER!!!"
|"A what?"
This Anti-spam address expires Dec. 31st |"LETNi! LETNi! LETNi!" - 1983