I'm including this October 1994 article as, of all the technology writing that I've done in the past thirty years, I'm proudest of the forecast that I made in the second-to-last sentence: Someone, someday, will make an enormous amount of money from marketing a low-cost UNIX computer.
With both Apple and Google making highly successful products built on top of UNIX, there's no question that this prediction came true!
With both Apple and Google making highly successful products built on top of UNIX, there's no question that this prediction came true!
Requiem for the TT
The ST and TT are obsolete - and that's official. William Hern mourns the unfulfilled potential of the TT.
I presume that by now everyone will have heard about the financial report from Atari that described its ST and TT ranges of personal computers as "technically obsolete and non-competitive". While it's hardly surprising for the ST to be described in such a way (in this age of PCs with clock speeds of 40MHz or more, how can a computer with a puny 8MHz processor be described as anything else?), I do feel sorry for the TT. If this is to be the end of the TT, then it will depart without ever having fulfilled its rich potential.
Right from the start, the TT faced problems.It was late to the market and for much of its existence supplies of it were short which made it difficult for dealers to promote it hard. It was not as competitively priced as the ST was when it was first released. And, like all recent Atari products, it suffered from a lack of advertising.
However the major reason for the failure of the TT in my opinion was the choice of primary operating system. The single-tasking TOS was fine for the ST but the TT was capable of so much more. What it needed was an operating system that could make full use of its substantial computing power.
Originally the plan had been to give the TT exactly that. Before the TT's launch, Atari announced that they were porting UNIX System V to it. This was tremendously exciting news - there had never been a truly low-cost UNIX computer before.
Unfortunately Atari then seemed to lose interest in UNIX. The project languished unfinished within the company as delivery dates slipped steadily. Eventually it was cancelled, with only a few copies having been passed out to a select group of developers.
Quite why Atari chose to drop the development is unclear to this day. My guess is that the company severely underestimated the amount of work involved in porting the operating system to the TT. Costs eventually rose to such a point that the company couldn't see how it could ever be profitable for them.
There was also the probable issue of UNIX's heavy system requirements - the OS needs a lot of RAM and hard disk space. Running it on a non-hard disk system is unimaginable.
Whatever the reasons, cutting the UNIX development was a real blow to the TT's prospects. Without that operating system, the TT could really never be anything other than a go-faster ST.
UNX would have been the ideal operating system to take advantage of the TT's special features. UNIX offers full pre-emptive multi-tasking, allowing many programs to be run at once, all oblivious to each other's existence. Thanks to a technique known as virtual memory, hard disk space can be used as pseudo-RAM, giving the illusion that the computer has far more memory than it actually has.
And forget about TOS's eight-plus-three-character filename limit - in practical terms there is no limit to how long a UNIX filename can be.
The reasons for the advantages of UNIX over TOS can be seen by looking at their roots. TOS was built upon CP/M, an operating system designed for small eight-bit personal computers with extremely limited memory. By contrast, UNIX was designed for running on powerful multi-user mini-computers with megabytes of memory space.
UNIX has long been popular in academia and there is plenty of good software available. The range of UNIX utilities is staggering and allows almost any data processing task to be accomplished. I've mentioned before the ST ports of the GNU software series (see the March 1993 column). Good as the ports are, they don't compare to the UNIX implementations.
Porting UNIX to the TT would have meant thousands of programs would have been available for it overnight - all that would have needed to be done would be to compile and link them on the TT.
The power and flexibility of the UNIX system does come at a price. The most common criticism made of UNIX is that it is unfriendly to use. Skeptics will cite the cryptic commands as an example of why UNIX will be unable to cross over into the mainstream computer market.
Now there's a certain element of truth to this allegation. With commands named rm, ls and chmod, it's not hard to understand why new users have difficulty at first. If anyone else does get round to sitting those monkeys down at their typewriters to being work on their first Shakespearian sonnet, you can be pretty sure that the first thing they'll manage to type successfully will be a valid UNIX command.
However great strides forward have been made in recent years to tame UNIX and make it easier to use. Anyone who has used one of the NeXT machines, with its innovative UNIX-based NeXTSTEP operating system will be able to confirm that it is as easy to use as a PC or a Mac or a ST.
The other criticism made of UNIX is that it is too resource demanding for a home computer. That certainly used to be true but the advances in computer technology mean that home computers are becoming ever more powerful. UNIX is viable on a 68030 or higher processor equipped machine, with sixteen megabytes of memory and about two hundred megabytes of hard disk space. Given that both processing power and memory storage of the average computer doubles every two years, it won't be very long at all before this is a typical configuration for a home computer.
It's tempting to speculate where the TT might be today had Atari persevered with the development of System V. A low-cost UNIX workstation would be attractive to a large number of professionals. It would also be popular among universities which previously had to purchase expensive Sun or Hewlett-Packard machines. If Atari had been able to market the TT as a sub one thousand pound UNIX workstation then they would have attracted a lot of attention in the market.
Using the single-tasking TOS on a TT is like driving a Ferrari only in first gear. The only way to show the TT's true talents would have been with an industrial-strength operating system such as UNIX. Someone, someday, will make an enormous amount of money from marketing a low-cost UNIX computer. It's a shame that it probably won't be Atari.
William Hern
Right from the start, the TT faced problems.It was late to the market and for much of its existence supplies of it were short which made it difficult for dealers to promote it hard. It was not as competitively priced as the ST was when it was first released. And, like all recent Atari products, it suffered from a lack of advertising.
However the major reason for the failure of the TT in my opinion was the choice of primary operating system. The single-tasking TOS was fine for the ST but the TT was capable of so much more. What it needed was an operating system that could make full use of its substantial computing power.
Originally the plan had been to give the TT exactly that. Before the TT's launch, Atari announced that they were porting UNIX System V to it. This was tremendously exciting news - there had never been a truly low-cost UNIX computer before.
Unfortunately Atari then seemed to lose interest in UNIX. The project languished unfinished within the company as delivery dates slipped steadily. Eventually it was cancelled, with only a few copies having been passed out to a select group of developers.
Quite why Atari chose to drop the development is unclear to this day. My guess is that the company severely underestimated the amount of work involved in porting the operating system to the TT. Costs eventually rose to such a point that the company couldn't see how it could ever be profitable for them.
There was also the probable issue of UNIX's heavy system requirements - the OS needs a lot of RAM and hard disk space. Running it on a non-hard disk system is unimaginable.
Whatever the reasons, cutting the UNIX development was a real blow to the TT's prospects. Without that operating system, the TT could really never be anything other than a go-faster ST.
UNX would have been the ideal operating system to take advantage of the TT's special features. UNIX offers full pre-emptive multi-tasking, allowing many programs to be run at once, all oblivious to each other's existence. Thanks to a technique known as virtual memory, hard disk space can be used as pseudo-RAM, giving the illusion that the computer has far more memory than it actually has.
And forget about TOS's eight-plus-three-character filename limit - in practical terms there is no limit to how long a UNIX filename can be.
The reasons for the advantages of UNIX over TOS can be seen by looking at their roots. TOS was built upon CP/M, an operating system designed for small eight-bit personal computers with extremely limited memory. By contrast, UNIX was designed for running on powerful multi-user mini-computers with megabytes of memory space.
UNIX has long been popular in academia and there is plenty of good software available. The range of UNIX utilities is staggering and allows almost any data processing task to be accomplished. I've mentioned before the ST ports of the GNU software series (see the March 1993 column). Good as the ports are, they don't compare to the UNIX implementations.
Porting UNIX to the TT would have meant thousands of programs would have been available for it overnight - all that would have needed to be done would be to compile and link them on the TT.
The power and flexibility of the UNIX system does come at a price. The most common criticism made of UNIX is that it is unfriendly to use. Skeptics will cite the cryptic commands as an example of why UNIX will be unable to cross over into the mainstream computer market.
Now there's a certain element of truth to this allegation. With commands named rm, ls and chmod, it's not hard to understand why new users have difficulty at first. If anyone else does get round to sitting those monkeys down at their typewriters to being work on their first Shakespearian sonnet, you can be pretty sure that the first thing they'll manage to type successfully will be a valid UNIX command.
However great strides forward have been made in recent years to tame UNIX and make it easier to use. Anyone who has used one of the NeXT machines, with its innovative UNIX-based NeXTSTEP operating system will be able to confirm that it is as easy to use as a PC or a Mac or a ST.
The other criticism made of UNIX is that it is too resource demanding for a home computer. That certainly used to be true but the advances in computer technology mean that home computers are becoming ever more powerful. UNIX is viable on a 68030 or higher processor equipped machine, with sixteen megabytes of memory and about two hundred megabytes of hard disk space. Given that both processing power and memory storage of the average computer doubles every two years, it won't be very long at all before this is a typical configuration for a home computer.
It's tempting to speculate where the TT might be today had Atari persevered with the development of System V. A low-cost UNIX workstation would be attractive to a large number of professionals. It would also be popular among universities which previously had to purchase expensive Sun or Hewlett-Packard machines. If Atari had been able to market the TT as a sub one thousand pound UNIX workstation then they would have attracted a lot of attention in the market.
Using the single-tasking TOS on a TT is like driving a Ferrari only in first gear. The only way to show the TT's true talents would have been with an industrial-strength operating system such as UNIX. Someone, someday, will make an enormous amount of money from marketing a low-cost UNIX computer. It's a shame that it probably won't be Atari.
William Hern