One tangent of my Atari What if:

[Kalvan]

I've Decided that as the window of opportunity for Microsoft to retaliate

with the MSX standard over the Windows Lawsuit is rather short (one year

max) before Macintosh sales exceed Apple IIs, I think that there's a

different butterfly flap in the timeline I can use:

 

Bill Gates gets wind of the MSX project in late 1981-1982, and when Don

Estridge is talking to him about an idea to conquer the home market as

thuroughly as the PC 5150 did the office, Bill says that he'd like to put

him in touch with a guy in Microsoft's Japanese office named Nishi Kazuhiko

since he might have just the thing to do the job, and one thing leads to

another. Of course, as word gets around to the suppliers, some of them, and

their clients, decide they want in too. In some cases (like Tandy and Texas

Instruments) it represents another bite at the home computer apple (pun

intended) in the face of past failing hardware. And in at least one case

(Coleco) it represents an opportunity to hitch yourself to a rising star in

the computing world. Due to the increased American input into the standard,

and the greater need for innovation, the spec will change somewhat from OTL.

Will the "IBM PCjr Compatibility Standard" be able to catch Atari, Apple,

and Commodore? In a world without Windows, and possibly VGA and Sound

Blaster, will the MSX standard hasten the development of the Zilog Z380 and

Ez80 and go on eclipse the PC standard, relegating Intel to server chips,

hard and solid state drives, RAM, and microcontrollers? And how will this

development effect the competition?

 

American MSX Computers in the Atari Lives Timeline:

 

MSX 1 (released in 1982):

 

Specification:

 

CPU: Zilog Z80A, 3.58 Mhz

ROM: 48 K consisting of 16K BIOS, 16K Microsoft Basic, MSX-DOS 1.0 (Floppy

Disc Control)

RAM: 8K Minimum. 128 Maximum through memory paging. Most computers possessed

32-64K

GPU: Texas Instruments TMS9918 series 16kb VideoRAM

Sound Processor: Texas Instruments SN76489

Most computers were packed with a 32/64/128k 5 ½ " Floppy Disc Drive or some

form of tape drive.

 

IBM PC jr.

Radio Shack/Tandy TSR-80 Model 2

Texas Instruments TI250, TI250/4, TI250/4A

Coleco ADAM

Sharp LG-80

Panasonic CF-2000, CF-2700

Phillips VG 8000, VG 8010, VG 8020

Spectravision SV-728, SV-738

 

MSX 2: (Released 1985)

 

Specification:

CPU: Zilog Z800, At least 3.58-7.66 Mhz

ROM: 64-96 K Consisting of MSX BIOS2.0 32K, Microsoft Basic 16K, MSX DOS 2.0

(floppy disc control), and optionally MSX-Audio BIOS

RAM: 64K Minimum, Frequently 128 K, Machines of up to 512K were made

(Maximum Theoretical Capacity of 4 Meg)

GPU: Texas Instruments/Yamaha TMS/YV9938 series. Video RAM: Variably 64-198K

Sound Processor: Choice of Texas Instruments AS 2000 series or Yamaha WM2149

Clock Chip: Western Design 65C19

3.5": Single Sided Double Density/ Double Sided Single Density disc drive

common.

 

Canon V Series

IBM HS/2

Radio Shack/Tandy TSR-180

Panasonic FS-A1

Phillips VG/NMS 8220-8280

Spectravision SV-838

Texas Instruments TI350

 

MSX 3 (Released 1987):

 

Specification:

CPU: Zilog Z800/Z280, at least 5.73 Mhz (Up to 14.25 Mhz)

ROM: 64-96 K Consisting of MSX BIOS 3.0, MSX DOS 3.0, Microsoft Basic or

Microsoft Q-Basic

RAM: 64K Minimum, 128-256k common. Machines of up to 640k were produced.

Theoretical Memory Capacity of 16 Meg

GPU: Texas Instruments/Yamaha TMS/YV9958 Video RAM 198 K

Sound Processor: Choice of Texas Instruments AS 2000 series or Yamaha WM2149

Clock Chip: Western Design 65C19

3.5" Double Sided Double Density/ Double Sided High Density disc drive

Mouse or Joystick Common

 

Canon H Series

IBM HS/3

Radio Shack/Tandy TSR-280

Panasonic FS-A2

Phillips VG/NMS 8420-8680

Sony HB-XV

Texas Instruments TI500

 

MSX 4: (Released in 1990)

 

Specification:

 

CPU: Zilog Z380, Ascii R800, 7.16 Mhz (Up to 28.6 Mhz)

ROM: 48 K: MSX BIOS 4.0

RAM: 256 K Minimum. Machines of up to 2 Meg were produced

Theoretical Memory Capacity of 16 Meg

GPU: Texas Instruments/Yamaha TMS/YV9958. Optional Texas Instruments

TMS9995, Optional Blitter Upgrade

Sound Processor: Choice of Texas Instruments AS 2000 series, Texas

Instruments AS 3000 Series, Yamaha WM2149, Yamaha YM2413

PCM Subprocessor: 8 Bit Single Channel, 16 kHz Max Sampling Rate

MIDI in/out (Sony and Panasonic computers only)

Clock Chip: Zilog Z8 Microcontroller

3.5" Double Sided High Density/Single Sided Extended Density disc drive

4 megabyte hard drive: Loaded with MSX DOS 4.0 (bitcode compatible with

MS-DOS and IBM PC-DOS 5.0) and either Microsoft QBasic, Microsoft Visual

Basic or Microsoft Visual LOGO.

 

IBM HS/4

Panasonic FS-A3/FS-A4

Phillips VG-8800-8880

Radio Shack/Tandy TSR-380

Sony VAIO

Texas Instruments TI750

 

Here's the Side-by-Side Spec for the 1982

Christmas Shopping Season:

 

Apple IIe

 

 

Microprocessor

 

6502 or 65C02 running at 1.023 MHz

8-bit data bus

Memory

 

64 KB RAM built-in

16 KB ROM built-in

Expandable from 64 KB up to 1 MB RAM or more

Video modes

 

40 and 80 columns text, white-on-black, with 24 lines¹

Low-Resolution: 40×48 (16 colors)

High-Resolution: 280×192 (6 colors) *

Double-Low-Resolution: 80×48 (16 colors)

Double-High-Resolution: 560×192 (16 colors) *

*effectively 140×192 in color, due to pixel placement restrictions

 

¹Text can be mixed with graphic modes, replacing either bottom 8 or 32 lines

of graphics with 4 lines of text, depending on video mode

 

Audio

 

Built-in speaker; 1-bit toggling

Built-in cassette recorder interface; 1-bit toggle output, 1-bit

zero-crossing input

Expansion

 

Seven Apple II Bus slots (50-pin card-edge)

Auxiliary slot (60-pin card-edge)

Internal connectors

 

Game I/O socket (16-pin DIP)

RF modulation output (4-pin Molex)

Numeric keypad (11-pin Molex)

External connectors

 

NTSC composite video output (RCA connector)

Cassette in/out (two 1/8" mono phono jacks)

Joystick (DE-9)

 

Retail Price: Typically $850

 

Commodore 64:

 

Microprocessor CPU:

MOS Technology 6510/8500 (the 6510/8500 being a modified 6502 with an

integrated 6-bit I/O port)

Clock speed: 0,985 MHz (PAL) or 1,023 MHz (NTSC)

Video: MOS Technology VIC-II 6567/8562 (NTSC), 6569/8565 (PAL)

16 colors

Text mode: 40×25 characters; 256 user-defined chars (8×8 pixels, or 4×8 in

multicolor mode); 4-bit color RAM defines foreground color

Bitmap modes: 320×200 (2 unique colors in each 8×8 pixel block),[25] 160×200

(3 unique colors + 1 common color in each 4×8 block)[26]

8 hardware sprites of 24×21 pixels (12×21 in multicolor mode)

Smooth scrolling, raster interrupts

Sound: MOS Technology 6581/8580 SID

3-channel synthesizer with programmable ADSR envelope

8 octaves

4 waveforms: Triangle, Sawtooth, Variable pulse, Noise

Oscillator synchronization, ring modulation

Programmable filter: high pass, low pass, band pass, notch filter

Input/Output: Two 6526 Complex Interface Adapters

16 bit parallel I/O

8 bit serial I/O

Time of Day clock (TOD)

16 bit cascadable timers

RAM:

64 kB (65 536 bytes), of which 38 kB minus 1 byte (38 911 bytes) were

available for BASIC programs

512 bytes color RAM

Expandable to 320 kB with Commodore 1764 256 kB RAM Expansion Unit (REU);

although only 64 kB directly accessible; REU mostly intended for GEOS. REUs

of 128 kB and 512 kB, originally designed for the C128, were also available,

but required the user to buy a stronger power supply from some third party

supplier; with the 1764 this was included. Creative Micro Designs also

produced a 2 MB REU for the C64 and C128, called the 1750 XL. The technology

actually supported up to 16 MB, but 2 MB was the biggest one officially

made. Expansions of up to 16 MB were also possible via the CMD SuperCPU.

ROM:

20 kB (9 kB BASIC 2.0; 7 kB KERNAL; 4 kB character generator, providing two

2 kB character sets)

 

Retail Price: Typically $599

 

Atari 1000

Processor Atari Ten Bit 7802 "Sylvia"

Clock speed 2.15 MHz

RAM 48 KB (80K 1000XL)

ROM 32 KB

O.S. version: AtariOS 2.1 (Based on the OTL XL OS, but torture tested to be fully backward compatible

AtariSoft Basic

Parallel Bus Interface Yes (1000XL Only)

Colors 16 colors with 16 intensity level each

Sound 4 voices, 3.5 octaves (POKEY chip)

Text modes Five text modes, max: 40 x 24, min: 20 x 12

Graphics modes 12 graphic modes, maximum: 320 x 192 pixels

Co-processor GTIA (video output, sprites), POKEY (sound, I/O), ANTIC (video, display lists)

Keyboard Full stroke keyboard

function keys (Reset, Option, Select, Start, Help, F1, F2, F3, F4)

I/O ports RGB, cartridge, Expansion port, Tape, RS232, Joystick

Size / Weight

Build-in media modem, speech synthesiser chip

 

Retail Price: $450 (1000), $625 (1000XL)

 

MSX: (IBM PC jr & Compatibles)

 

Specification:

 

CPU: Zilog Z80A, 3.58 Mhz

ROM: 48 K consisting of 16K BIOS, 16K Microsoft Basic, MSX-DOS 1.0 (Floppy

Disc Control)

RAM: 8K Minimum. 128 Maximum through memory paging. Most computers possessed

32-64K

GPU: Texas Instruments TMS9918 series

16kb VideoRAM

Text modes: 40×24 and 32×24

Resolution: 256×192 (16 colours)[11]

Sprites: 32, 1 color, max 4 per horizontal line

Sound Processor: Texas Instruments SN76489

Four Sound Channels

Most computers were packed with a 32/64/128k 5 ½ " Floppy Disc Drive or some

form of tape drive.

 

Retail Price: Low of $325 Radio Shack/Tandy TSR 80 Model 2-36-3001 16K High

of $675 (IBM PCjr Deluxe, 128 K, Standard Disc Drive) Coleco ADAM (80 K RAM

but Standard Printer, 2 packed in joysticks, second cartridge slot

compatible with the ColecoVision, and special expansion ports to link with

the ColecoVision and its periphrals.)

 

Questions? Comments? Flames? Contributions?

Edited December 22, 2009 by Kalvan

[Kalvan]

Oh, yes, and the memory location bugs that plagued OTL's MSX 1 standard were found and fixed before it came time to print motherboards.

[JamesD]

I'm not sure I get the IBM PCjr reference. It was supposed to be a home friendly IBM PC and I don't see MSX replacing it.

 

1981 was too late.

 

For example... Tandy introduced the Model II in '79, the Model III in '80, and the CoCo in '80. That means Tandy started to develop the hardware a year or two before that. By the time MSX design got anywhere Tandy had sold hundreds of thousands of machines and the path for those machines was set.

The Tandy 1000 was introduced in 1984 (late '83?) and was a PCjr clone of sorts. That is the only machine that might have been replaced from Tandy and MSX offered little advantage over Tandy's existing models to justify supporting it.

 

BTW, a 10 bit processor never would have worked out for a home couputer due to all RAM being set up around 8 or 16 bit CPUs. It might have worked in a videogame with little RAM. A 12 bit CPU might have had a chance but you would have to almost do away with the 16 bit processors which had been around for years.

 

I think the only way 8 bits would have survived is if it had proved too expensive to produce 16 bit chips or if speeds dropped as chip die size increased.

If 8MHz 8 bit cpus had been cheap and widely available by 1981 then 8 bits might have survived longer.

The only thing I see coming of either scenario is RISC becoming the dominant technology as adoption of the 8088/68000 chips would have been delayed and RISC would have had a die size/speed advantage.

[Kalvan]

I'm not sure I get the IBM PCjr reference. It was supposed to be a home friendly IBM PC and I don't see MSX replacing it.

 

It was also more expensive than all its competitors save the Apple IIe, came out without color standard (and the only color option was crumby CGA), no sprite manipulation, sound hardware that was all beeps and boops, that buggy remote control keyboard, and with only 64k of RAM in 1984, was merely middle of the pack in memory capacity and about to get trounced by Amiga, ST, and Apple IIgs. Without Windows 3.0, Sound Blaster, or the XGA/Super VGA standard, but with a readily available cheaper home computer with truely competitive graphics and sound features able to run most contemporary PC-DOS/MS DOS software (memory, drive capacity, and aftermarket hardware issues not withstanding), the PC jr as we know it made about as much sense as hawking cut down Digital Equipment, Hewlett-Packard, or Silicon Graphics workstations for the home (Even less sense, actually. Those were known for their graphics and sound hardware, and were quite capable of running UNIX).

 

edit: no, I take that back, IBM created the EGA standard for PC jr. But it still meant only 16 colors in 1984!

Edited December 23, 2009 by Kalvan

[Kalvan]

I'm not sure I get the IBM PCjr reference. It was supposed to be a home friendly IBM PC and I don't see MSX replacing it.

 

1981 was too late.

 

For example... Tandy introduced the Model II in '79, the Model III in '80, and the CoCo in '80. That means Tandy started to develop the hardware a year or two before that. By the time MSX design got anywhere Tandy had sold hundreds of thousands of machines and the path for those machines was set.

The Tandy 1000 was introduced in 1984 (late '83?) and was a PCjr clone of sorts. That is the only machine that might have been replaced from Tandy and MSX offered little advantage over Tandy's existing models to justify supporting it.

 

The CoCo ran all sound and graphics off the CPU until Model 3 in 1985, by which time it was an absolute irrelevance. The Tandy 1000 was every bit as expensive as the PC jr (perhaps more so), and until the hard drive came along in 1986, it couldn't do anything its home computer competition couldn't (except run MS-DOS applications).

 

Why wouldn't Tandy jump on the MSX bandwagon if it meant getting something that could finally run with Commodore in time to far better compete during the 1982 price war, and due to running the same BASIC as the Model I and CoCo, was fully backward compatible with it (aside from possibly the cartridges)?

Edited December 23, 2009 by Kalvan

[Kalvan]

But aside from the Texas Instruments IT99/4(A), Mattel Aquarius and of course, the OTL PC jr (all dismal failures), no one was selling 16 bit home computers until the Tandy 1000, Amiga, ST, and Apple IIgs in 1985-6, because they couldn't afford to put in enough good memory, sound and graphics to make them worthwhile to the comsumer for the money.

[Kalvan]

 

I think the only way 8 bits would have survived is if it had proved too expensive to produce 16 bit chips or if speeds dropped as chip die size increased.

If 8MHz 8 bit cpus had been cheap and widely available by 1981 then 8 bits might have survived longer.

The only thing I see coming of either scenario is RISC becoming the dominant technology as adoption of the 8088/68000 chips would have been delayed and RISC would have had a die size/speed advantage.

 

But Zilog already had the Z800 out, so there was a 16 bit upgrade path available for the architecture. Perhaps if Zilog had'nt marched down the Z8000/Z80000 trap and instead developed what would have been the Ascii R800 in 1982, and the Z380 (but based on R800's RISC architecture) in 1985.

[vcsdream]

But aside from the Texas Instruments IT99/4(A), Mattel Aquarius and of course, the OTL PC jr (all dismal failures), no one was selling 16 bit home computers until the Tandy 1000, Amiga, ST, and Apple IIgs in 1985-6, because they couldn't afford to put in enough good memory, sound and graphics to make them worthwhile to the comsumer for the money.

 

Sinclair QL?

[Kalvan]

But aside from the Texas Instruments IT99/4(A), Mattel Aquarius and of course, the OTL PC jr (all dismal failures), no one was selling 16 bit home computers until the Tandy 1000, Amiga, ST, and Apple IIgs in 1985-6, because they couldn't afford to put in enough good memory, sound and graphics to make them worthwhile to the comsumer for the money.

 

Sinclair QL?

 

It may have been Linus Torvalds' first computer, but it wasn't available in the U.S. and continued to use the Microdrive without refining its speed or reliability, upping its capacity to keep up with disc drives, or adding a random seek function.

 

In short it was about as successful as all the above.

[Kalvan]

But aside from the Texas Instruments IT99/4(A), Mattel Aquarius and of course, the OTL PC jr (all dismal failures), no one was selling 16 bit home computers until the Tandy 1000, Amiga, ST, and Apple IIgs in 1985-6, because they couldn't afford to put in enough good memory, sound and graphics to make them worthwhile to the comsumer for the money.

 

And even the Tandy 1000 only had MS-DOS compatibility as an advantage.

[JamesD]

The PCjr at least fit with IBM, MSX was in a totally different direction and had no relation at all to IBM establishing themselves as the standard. The MSX machine would never run MS-DOS apps. If the PCjr had been more like that Tandy 1000 it might have succeeded. Remember, it was basically a DOS PC with better graphics and a sound chip. If IBM had concentrated on making it cheaper to produce instead of creating the bizarre expansion setup and chicklet keyboard (a primary reason for failure of any computer in the US) it may have succeeded. Just replacing the keyboard would have probably more than doubled sales. It was also introduced in 1983, not 1984.

 

The TI99 sold well but it could not compete on price with the C64. Frankly, neither would MSX which is why MSX wasn't very successful outside of Japan and non-existent in the worlds largest computer market, the US.

 

 

All the Tandy machines you called irrelevant were successful and some people didn't seem to care about sprites and the rest.

The CoCo sold in the millions and Tandy said it was their best selling machine. Tandy never released actual sales data after the Model I but one quote placed the number of CoCos sold by the mid 80s at over 2 million. Since the CoCo continued to sell until the early 90s that put it on pace to sell similar numbers as the Atari 8 bits. Far from irrelevant. The CoCo was also cheap to produce allowing it to survive the price war.

 

If you actually look around you'll find the CoCo 3 version of Donkey Kong pretty much stomps all the rest and it does it without hardware sprites or a sound chip.

 

The CoCo's Extended Color Basic was commonly recognized by magazines as being the best version of BASIC. You could learn BASIC on the CoCo, then load up GW-BASIC on the PC and it's syntax was almost identical. No other 8 bit can make such a claim.

 

Too bad Motorola dropped development of their advanced graphics chip. It was pretty much the CoCo 3 GIME chip but with sprites and other features well beyond that of the CoCo 3. It also could have been available years before the CoCo 3, which would have made Tandy capable of running head to head with the C64 on graphics but with the much better 6809 CPU. The Motorola chip was also designed to interface to the 68000 CPU, making it possible for Tandy to bring out a direct competitor to the Atari ST and Amiga. But Motorola dropped the ball by dropping development of the chip. I'm guessing they didn't like the price point required to compete with the C64 and Motorola had always been fond of large margines on their sales.

[Kalvan]

The PCjr at least fit with IBM, MSX was in a totally different direction and had no relation at all to IBM establishing themselves as the standard. The MSX machine would never run MS-DOS apps.

 

So why did everyone and his uncle I looked up say that that MSX-DOS was bitcode compatible with the MS-DOS command interpreter until 6.0? Admittedly, there was no standard hard drive (but then there wasn't on the PC jr), and memory was limited to 128K for the MSX 1 standard, but that's still enough to run Lotus 123 until 1985.

[JamesD]

But aside from the Texas Instruments IT99/4(A), Mattel Aquarius and of course, the OTL PC jr (all dismal failures), no one was selling 16 bit home computers until the Tandy 1000, Amiga, ST, and Apple IIgs in 1985-6, because they couldn't afford to put in enough good memory, sound and graphics to make them worthwhile to the comsumer for the money.

 

Sinclair QL?

 

It may have been Linus Torvalds' first computer, but it wasn't available in the U.S. and continued to use the Microdrive without refining its speed or reliability, upping its capacity to keep up with disc drives, or adding a random seek function.

 

In short it was about as successful as all the above.

The QL sold just over 100,000 machines (from what I could find) which is not even close to the Tandy 1000, Amiga, ST or IIgs.

The IIgs had the lowest sales of those machines but it certainly had many times the sales of the QL.

The others totaled in the millions over their lifetimes.

[Kalvan]

The QL sold just over 100,000 machines (from what I could find) which is not even close to the Tandy 1000, Amiga, ST or IIgs.

The IIgs had the lowest sales of those machines but it certainly had many times the sales of the QL.

The others totaled in the millions over their lifetimes.

 

I was referring to Aquarius and PC jr.

[JamesD]

So why did everyone and his uncle I looked up say that that MSX-DOS was bitcode compatible with the MS-DOS command interpreter until 6.0? Admittedly, there was no standard hard drive (but then there wasn't on the PC jr), and memory was limited to 128K for the MSX 1 standard, but that's still enough to run Lotus 123 until 1985.

Sorry to burst your bubble but I think they just mean the format of the disk is compatible.

The two machines use totally different CPUs! 8088 vs Z80.

An MSX machine would not run Lotus 123 unless lotus rewrote it for the Z80.

[roland p]

MSX 3 (Released 1987):

 

...

 

MSX 4: (Released in 1990)

I don't know any msx-3 or 4 after the msx-2. That would be the msx turbo-r right?

Edited December 23, 2009 by roland p

[JamesD]

The QL sold just over 100,000 machines (from what I could find) which is not even close to the Tandy 1000, Amiga, ST or IIgs.

The IIgs had the lowest sales of those machines but it certainly had many times the sales of the QL.

The others totaled in the millions over their lifetimes.

 

I was referring to Aquarius and PC jr.

Ok... Aquarius probably. I think the PCjr still outsold the QL by at least 2 to 1 in spite of the fact that it was a flop... which should tell you just how big of a flop the QL was. If the QL had gone with a floppy, a better keyboard, and didn't have issues at launch it probably would have sold pretty well.

[JamesD]

MSX 3 (Released 1987):

 

...

 

MSX 4: (Released in 1990)

I don't know any msx-3 or 4 after the msx-2. That would be the msx turbo-r right?

I think those are part of the "what if" he is proposing.

Since he appears to think MSX could somehow run MS-DOS software, I think that blows the entire idea.

[Kalvan]

MSX 3 (Released 1987):

 

...

 

MSX 4: (Released in 1990)

I don't know any msx-3 or 4 after the msx-2. That would be the msx turbo-r right?

 

Basically

[Kalvan]

Well then, JamesD, if what you say is true (which I consider to be assuming a spherical chicken of uniform density for the moment), would it still be possible for Microsoft (among others) to support the MSX project by porting over Multiplan (the predecessor to Excel), Project, Word, and as many of their computer language products as they can manage with interopterable file syatems with their MS-DOS counterparts so as to let someone bring their work home with them? Perhaps offer a package discount?

[JamesD]

Well then, JamesD, if what you say is true (which I consider to be assuming a spherical chicken of uniform density for the moment), would it still be possible for Microsoft (among others) to support the MSX project by porting over Multiplan (the predecessor to Excel), Project, Word, and as many of their computer language products as they can manage with interopterable file syatems with their MS-DOS counterparts so as to let someone bring their work home with them? Perhaps offer a package discount?

Spherical chicken?? Um... yeah... whatever.

 

I've programmed a lot of different chips in assembly.

The 8088 supports 16 bit numbers and large amounts of memory better than the Z80.

It's registers are also a little more general purpose in nature than the Z80 so it supports compilers better.

Ultimately, that makes developing for MSX more difficult than for MS-DOS.

However, it would certainly be possible to port many MS-DOS programs to MSX.

MS-DOS is similar to CP/M, and MSX DOS has most of the CP/M calls if I remember right.

I think you'll find that Multiplan was first introduced on CP/M and there is a version for MSX.

 

Now, having already said it would be more difficult to port programs to MSX, there are C compilers and assemblers that can automatically make use of the Hitachi 64180 and later Z180 MMU. (the Z180 duplicates the 64180 but with a more Z80 compatible pinout)

If MSX had used the 64180/Z180 from day one, then supporting larger amounts of memory would have been that much easier.

However, the Z180 MMU only supports 1MB of address space. That means RAM + BIOS can't exceed 1MB.

So, as long as your MS-DOS program only needs 512K, it would probably be easy to port to an MSX Z180 system.

There are some other added bonuses to using one of these CPUs.

Built in I/O ports, DMA, timers, reduced execution time of instructions, and added instructions that make compiler output more efficient.

I have no idea what year the 64180 was introduced so it may or may not have been possible.

<edit> I forgot to mention higher clock speeds. There was also a 64180 upgrade unit for the TRS-80 Model 1 so it might have been available in time for MSX.

Edited December 26, 2009 by JamesD

[JamesD]

http://www.atarimagazines.com/compute/issue56/107_1_MSX_IS_COMING.php

[Kalvan]

http://www.atarimagazines.com/compute/issue56/107_1_MSX_IS_COMING.php

 

But that was in OTL. Rewind that back three years to the 1982 price war and replace the Z80 with a Z180, and the results would have been rather different. My timeline has Bill Gates get wind of the project long before he did in real life, and he and Phillip Estridge put their full backing behind it.

Edited December 27, 2009 by Kalvan

[JamesD]

http://www.atarimagazines.com/compute/issue56/107_1_MSX_IS_COMING.php

 

But that was in OTL. Rewind that back three years to the 1982 price war and replace the Z80 with a Z180, and the results would have been rather different. My timeline has Bill Gates get wind of the project long before he did in real life, and he and Phillip Estridge put their full backing behind it.

I don't think it's just a matter of Bill Gates getting wind of it sooner, I think it's a matter of the project starting sooner.

Bill probably heard about the project within 6 months or less of it's beginning. I'm sure they approached him about it early on.

If MSX had been available before the price war... it would have given the C64 a run for the money on numbers.

 

However... the TI chipset requires more chips to build a system. Same problem the Colecovision had.

Without an updated chipset competing on price would have been difficult.

And remember, TI was still in the game before the price war so pricing on the chipset would have been favorable to TI.

 

MSX was also based heavily on CP/M, which started dropping in popularity with the introduction of MS-DOS.

MS-DOS was introduced in 1981. Once MS-DOS hit, PC compatibles started leading sales.

If you combined all the clones they actually outsold the C64.

That should tell you what MSX would have been up against.

I think for MSX to become a major player, it would have had to actually beat the IBM PC to market.

That would have given it time to get a foothold in the market.

But since it depends on MS-DOS format for it's disks, I find that highly unlikely.

 

Given those considerations, I think to do what you want requires MSX to be another Z80 8 bit machine from 1980 and MSX 2 would have been from 1982 with the Z180. MSX 2 would have to require support for 4MHz but could allow for higher MHz.

 

When you start doing this sort of what if session you have to consider how other companies would react.

Motorola would be pushed to finish it's new graphics chipset that supported higher resolutions, color palette, sprites, etc...

Commodore would have been pushed to improve the C64 design before it's release. Maybe they clock it at 2MHz.

Atari might have been pressured to improve it's chipset or go under.

 

Development of 16/32 bit machines may have been pushed forward to offer an alternative.

Apple dumps the Lisa and introduces the Mac earlier.

If the Mac hits earlier, it competes more with MS-DOS... and pretty soon the entire picture changes.

Suppose the Motorola chipset hits and the first Mac is color.

Now the Mac has larger market share and Microsoft is caught sleeping.

The Motorola chip was designed to work with the 6809/68000... so Tandy introduces the CoCo 2 with the new chip.

Now OS-9 Level 2 and the CoCo 2 is to the Mac what MSX is to MS-DOS.

Or Tandy introduces their own 68000 machine based on OS-9 68K.

Then they have a low end machine with a migration path to a high end system but with a more powerful OS than MS-DOS or MSX.

 

Change the past at some point and everything that follows changes.

What you propose may work or it may have unexpected consequences.

[Kalvan]

However... the TI chipset requires more chips to build a system. Same problem the Colecovision had.

Without an updated chipset competing on price would have been difficult.

And remember, TI was still in the game before the price war so pricing on the chipset would have been favorable to TI.

 

But the TI99/4(A)'s layout was so absurd it reads like someone was doing this just to humor someone in a very high-up position, either that or setting him up to be a fall guy. I mean, only 1/2k of the 16 bit RAM that the chip could directly process? That's small enough to be a cache. And no way to expand the memory without permanantly cracking open the shell other than that exerable SuperBASIC cartridge? That's inexcusable.

 

In my position, if I were in charge of TI, I would have jumped on the MSX badwagon and never looked back.

Edited December 28, 2009 by Kalvan