Stupid Decisions (or What Were They Thinking?)

[dudeslife]

atari 5200 controllers.

 

commodore +4, 16, 128

 

amiga 600

[Retrospect]

The Texas TI - 99 .... a machine I dearly love ... but for these reasons it failed to gain momentum in the UK....

 

The computer was given a TMS-9900 when it should have had some other chip they were planning, which failed at a late stage, so in the end it ended up with the 16-bit chip instead the 8-bit, on an 8-bit data bus, with no direct access to the Cassette deck ... now, you might think, what's the tape deck got to do with anything? Well, I'll tell ya .... that's why it failed, in Britain ... you couldn't save, or load, any Machine Code onto the tapes ... that had to be on Disk only (or cartridge) .... the TI was selling in a country where even the ZX Spectrum with it's .... keyboard thing .... could have M/C games on a tape, which would cost around a fiver each.

 

Ti games were getting made on tape in TI basic and Ext. Basic costing upwards of a tenner a-piece .... and they did not sell well.

 

To get the disk drive you had to fork out for the disk controller as well, which between them in 1981 would cost you upwards of 800 quid.... then there's the 32k expansion too.

 

Another thing was the joysticks .... two rather obscure sticks on one lead.

 

Having said all that, for my own reasons I love the TI 99 but from a marketing point of view it was doomed over here.

[Jess Ragan]

The C= +4 simply shouldn't have existed. What a waste of resources (time and money being two big ones) and materials when you think about it. C64 could have been turned into a simple crippled +4 with a single cartridge.

 

They were gonna make a dumbed down C64 called the Max, right? I remember playing its version of Wizard of Wor on an emulator... holy crap was it some awful stuff.

 

Anyway... yeah, the +4 line was hugely pointless. The system case is pretty attractive but there's not much else about it that justifies its existence. Commodore had spread itself pretty thin by 1985, between acquiring the Amiga brand, releasing the +4, supporting the C64, and trying desperately to kill the VIC-20 (with little success).

[am1933]

MATTEL AQUARIUS.

[BassGuitari]

MATTEL AQUARIUS.

And you can thank the Keyboard Component for that, since the only reason Mattel even bothered with the Aquarius (and the ECS) was to get the FTC off their back about little things like "false advertising" and "fraud."

 

As it is, it's a cute system but certainly was never going to be a contender BITD. The hardwired power brick was a bad idea though. I hate those. (See also: Channel F, Astrocade)

[JamesD]

Since the Aquarius was mentioned it reminded me of several machines... which reminded me of several others from the same period.

 

The Panasonic JR-200. In many ways it was a valiant attempt. 32K, built in sound, joystick ports, parallel port, expansion port, etc... Most of the specs were ok but it had no bitmapped graphics, a limited palette and a chicklet keyboard. It was too expensive to compete with the VIC20 and TS-1000, but not capable enough to compete with more expensive machines. But hey, it's non-standard BASIC was fast!

The NEC PC-6001/Trek. Lets create a Tandy CoCo work alike but with a Z80 clone CPU and support for 8 colors on screen at once in raster graphic modes. They even created an expansion interface to add upgrade cartridges like the CoCo. Not the worst idea here but why create a custom chip to control the 6847 and attach an external character generator ROM rather than just creating a custom video chip with better capabilities in the first place? The worst part... they didn't isolate the Z80 from video RAM so the added wait states were horrible in higher graphics resolutions. If they had even isolated part of RAM and the cartridge port addresses, the machine would have been ok for a sub $400 machine but it was too expensive to keep up with price cuts. It had some success in Japan though and led to better machines from NEC.

VTech VZ/Laser 110/200/300 series. This was aimed at the ZX-80 derivatives so from that standpoint it's ok, it has a better keyboard, color and sound. But In Europe it had competition from more capable machines at the same price like the Oric 1. The architecture is actually based on the TRS-80 Model I but with a 6847 for video output and some ROM patches. It's sort of what the CoCo would have been had Radio Shack used a Z80 instead of the 6809. But why license Microsoft BASIC and have so many commands disabled yet still in ROM? Why not allow for video RAM expansion so people could use the higher res graphics modes the 6847 already supports? They didn't even add that with the 300 which had a better keyboard. None of the chips are socketed either so hacking it requires solder work. If they had even socketed the ROMs and the 6847, owners could have plugged in a ROM with commands re-enabled and a video mod with more video RAM. As is, it's more like the MC-10 than the CoCo only more expensive to build. Had it been introduced somewhere between '77 (when the 6847 was released) and '80 (when the CoCo was released) it might have had a chance. People recognizing it as a TRS-80 at heart would have probably adopted the little beast but by '83 it wasn't enough... well... except in Australia and New Zealand. I guess that market was largely ignored by everyone but Dick Smith Electronics.

MC-10. Another machine targeted at the ZX-80 series. Memory map decoding is so incomplete it's ridiculous and built in RAM should be at the bottom or top of the memory map. Then BASIC might have easily supported up to 40K of RAM. With 8K and a small revision to the way the 6847 was wired it could have supported higher resolution graphics modes out of the box. Why didn't they socket the ROM? It was designed to be upgradable to a larger one on the motherboard. The internal memory mapping could be disabled by an expansion cartridge so it might have even been possible to disable the ROM and hide video RAM in order to run the FLEX OS with a 51 or 64 character wide screen using graphics like had been done with the CoCo. If they didn't want it to compete with the CoCo all they had to do was leave it as a cassette only system. If they were worried about that, why did they add the serial port in the first place?

Oric 1. The machine itself is fairly capable for the budget market; it's almost like an Apple II but with a better screen memory map and a built in single AY chip Mockingboard. But you can't disable the ROM to enable the underlying internal RAM without plugging in a disk controller. It seems to me the 6522 didn't even have all it's ports used so there was no excuse for that requirement. No joystick port? But there was an unused port on the AY chip so why not put in a connector? And have you ever tried typing on that keyboard? Odd shaped stiff keys that are worse to type on than chicklet keys. The Atmos fixed the keyboard issue and some ROM bugs but why not fix the memory issue? You have 16K of RAM than can't be used on a tape only system. Oh, and no vertical blank interrupt just like on the Apple II.

Jupiter Ace. It's like an enhanced ZX-80 running on Forth in 1984. Why did this ever exist at all? B&W?

And as for the TS-1000 (ZX-81)... people wanted TRS-80 Model I capabilities at a cheaper price and got a machine that didn't even have a power switch.

BTW, what did some companies have against power and reset switches?

Edited March 21, 2015 by JamesD

[MarkO]

atari 5200 controllers.

 

commodore +4, 16, 128

 

amiga 600

 

I don't know about Atari 5200 Controllers, but the Commodore +4 and 16 and the Amiga 600 yes.

 

The C128.. That is what every Good Computer Company should Do!!! Make your, New, Advanced Machine, BACKWARD COMPATIBLE with your Old Established Machine..

 

C64 ---> C128

Apple ][e ---> Apple ][GS

 

MarkO

[jhd]

Wow -- there were way-more non-Coco systems that used the 6847 video chip than I had ever realised! A Z-80 and a 6847 make for an interesting hardware combination.

 

As for the MC-10, it should be added that the tokenized BASIC format was (slightly) incompatible with the Coco. Yet another stupid design decision in a machine plagued with them.

 

I recall a letter to Hot Coco (IIRC) about possibly using the MC-10 as a smart terminal attached to a Coco using OS-9 for a second user. That may have been why the serial port was originally included. (The editors opined that there was simply not enough system resources to support two simultaneous users, and the question was never addressed further in a future issue.)

[JamesD]

Wow -- there were way-more non-Coco systems that used the 6847 video chip than I had ever realised! A Z-80 and a 6847 make for an interesting hardware combination.

 

As for the MC-10, it should be added that the tokenized BASIC format was (slightly) incompatible with the Coco. Yet another stupid design decision in a machine plagued with them.

 

I recall a letter to Hot Coco (IIRC) about possibly using the MC-10 as a smart terminal attached to a Coco using OS-9 for a second user. That may have been why the serial port was originally included. (The editors opined that there was simply not enough system resources to support two simultaneous users, and the question was never addressed further in a future issue.)

That's not all the 6847 systems.

There's the 6800 based APF Imagination Machine... which just deserves an all around WTF? It has awkward joysticks, a massive keyboard/computer expansion, to do anything with the computer part requires a bunch of expansion carts and even then it was pretty limited. I think it's biggest problem is it wasn't very good as a video game or as a computer. Had they put more effort into either direction it had to do better than it did but it was pretty limited color and sound wise vs the VCS. The ability to play cassette audio on one channel while loading a program on the other was an original idea though. Other than Flex systems, it's one of the only 6800 based personal computers. The 6800 required a more complex clock circuit and had longer cycle times per instruction than it's derivatives. When you combine a slow clock with the long cycle times you get a slow computer.

And then there's the Acorn Atom which was a 6502 based system. It's what were they thinking issue is they hadn't planned a fully expanded system ahead of time so the memory map ended up a bit of a mess with all sorts of optional expansion ROMs, optional hardware and you end up with little RAM for BASIC I think it was originally only intended as a prototype since some of the designers supposedly didn't want to release it as is. Against other personal computers available at the time in the Europe the Atom was ok. It supported hi-res graphics, had optional color, a good keyboard and sound for a decent price. The Atom pretty much laid the groundwork for the BBC Micro firmware. Without that Acorn might not have won the BBC contract and without the BBC Micro, there probably wouldn't have been an ARM CPU. I give it some slack because it was available as a kit but it was well ahead of other kit machines and Acorn didn't continue to try to market it once more capable machines were released. It was discontinued a year before the VZ, MC-10 and TREK were introduced.
That brings up my other "what were they thinking" issue with the Atom which is more of a "why didn't they?" If you think about it, Acorn could have released an updated compact version with chicklet keyboard to compete with the ZX-80 market in '81. At that time such a machine might have made sense and they could have racked up a lot of sales before the Spectrum and Oric hit the market. With the 6502 it would have been significantly faster than the ZX-80 and possibly just as cheap or cheaper to produce. The Atom is powerful enough to run a recent port of Elite pretty well, something the ZX-80/81 would never be good at. Then you have to think of the possibility of Timex partnering with Acorn instead of Sinclair. The 6847 was already designed for NTSC and a US version would actually be quicker to develop than the TS-1000 and cheaper to produce than it's Euro parent. Oh the implications... that has the makings of a serious "what if" topic.

The different tokens of MC-10 BASIC were easy to overcome. Just save the program with the A (ASCII) option.

 

This brings up some of the biggest what were they thinking issues from that time.
The CoCo was originally designed to be an Agriculture data terminal in '77-'78 but it wasn't released for a while.

It had lots of expensive parts for '78 including a 6809 so it made more sense as a computer but... the CoCo wasn't released until '80 and it was so stripped down you could only write BASIC programs around 3K in size.
Why does a a data terminal need a 6809, the extra graphics modes the SAM adds to the 6847, etc... ? Talk about overkill.
If you are doing it to end up with a computer, why not release it in '78 when hi-res graphics, color and sound are rare?

And Motorola had three years before the CoCo was released to improve on the 6847, so why didn't they? They could have at least added lowercase characters but that didn't happen until later versions of the CoCo 2 ('85?).

The MC-10 was a simple design and was perfect for use as a data terminal. It would have been cheap even in '77. It didn't need any more ROM than terminal software (4K?), it needed to hook to a TV, it *may* have needed cassette to save data, it didn't need hi-res graphics, who cares how the memory is mapped for a dedicated data terminal and it needed a serial port which was already part of the 6803 instead of being a bit banger like the CoCo. It didn't even require a lot of typing so the small keyboard isn't much of an issue.
Talk about getting things backwards.