On 2022-11-10 13:08, Charlie Gibbs wrote:
> On 2022-11-10, lar3ryca <larry@invalid.ca> wrote:
>
>> On 2022-11-09 15:32, Charlie Gibbs wrote:
>>
>>> On 2022-11-09, lar3ryca <larry@invalid.ca> wrote:
>>>
>>>> Biggest mistake I ever made was to throw all my Amigas out.
>>>
>>> Would you like one or two?
>>
>> Gasp! Two, please.
>> What would you want for them?
>> And how should I contact you?
>
> E-mail me. See my .sig.

Sent email yesterday.

--
I don't have an accent.
This is just how things sound when they're promounced properly.
- Jimmy Carr

On 12/11/22 12:16, Peter Flass wrote:
> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>> On 11/11/2022 03:45, Rich Alderson wrote:

>>> Then that's a function of your failing memory.
>>
>> Agreed. Thank you, and also Bob for the doc link.
>>
>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>
> I thought it was the acme of instruction sets. Everything was
> downhill from there.

I still give top score to the PDP-11. OK, it was not as powerful as the
PDP-10, but the design of its addressing modes was brilliant.

Once microprocessors started to appear, it was obvious (from the
instruction sets) which designers had an IBM computer background and
which ones had a DEC computer background. And, just as with mainframes,
it soon became clear that there was no correlation between elegance of
design and market dominance.

--
Peter Moylan Newcastle, NSW http://www.pmoylan.org

On 11/11/22 14:47, Rich Alderson wrote:
> "Anders D. Nygaard" <news2012adn@gmail.com> writes:
>
>> Den 10-11-2022 kl. 18:04 skrev Tak To:
>>> On 11/10/2022 8:40 AM, Peter Flass wrote:
>>>> [... I ...] think systems with 36-bit words that stored
>>>> characters in 9 bits may have called them bytes.
>>>> “Characters” was the term of art earlier.
>>>
>>> The only system I know of that fits that description is Multics,
>>> and I don't remember how those 9-bit entities were called.
>>
>> My memory is *very* hazy, but the system I used in my first year
>> at university answers to that description. I'm fairly sure it was
>> a UNIVAC; probably a model 1100.
>>
>> Since then, every byte I've come across has been 8 bits.
>
> Because the 400kg gorilla from Armonk changed the definition.

I don't think that was the main reason. There were, in my opinion, two
more important factors.

1. In the early days of computing, BCD was very important. That set a
precedent that suggested that the best choice of word size is a multiple
of 4 bits.

2. The choice of character width was all over the place, but gradually
the design of serial I/O interfaces - and serial was more pervasive than
parallel back in those days - was converging to the idea that a
character code should be 7 bits plus one parity bits. (Plus some start
and stop bits, but those didn't have to be stored once the character
reached the computer.) Thus, there was heavy off-she-shelf hardware
support for 8-bit characters.

These days we accept that a parity bit (or other form of
error-detecting/correcting) is needed only while passing the information
over a communications channel, and can be ignored once the character is
in computer memory, but that wasn't the thinking back then.

--
Peter Moylan Newcastle, NSW http://www.pmoylan.org

On Fri, 11 Nov 2022 19:52:07 -0500, Tak To wrote:

> On 11/11/2022 5:47 PM, Bob Eager wrote:
>> On Fri, 11 Nov 2022 21:30:33 +0000, Vir Campestris wrote:
>>
>>> On 11/11/2022 03:45, Rich Alderson wrote:
>>>> Then that's a function of your failing memory.
>>>
>>> Agreed. Thank you, and also Bob for the doc link.
>>>
>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>
>> It was about my fourth. My first was the Elliott/ICL 4100 series, which
>> is rarely mentioned.
>>
>> The PDP-10's instruction set was so orthogonal that there was a
>> complete set of conditional jumps and skips. For example, JUMP was a
>> no-op. JUMPA was an unconditional jump. (but JRST was reputedly faster)
>
> Yes. And the fastest no-op is JFCL.

I'd forgotten that one!

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

On Fri, 11 Nov 2022 18:57:48 -0800, Snidely wrote:

> Just this Friday, Vir Campestris explained that ...
>> On 11/11/2022 03:45, Rich Alderson wrote:
>>> Then that's a function of your failing memory.
>>
>> Agreed. Thank you, and also Bob for the doc link.
>>
>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>
>> Andy
>
> My first was the PDP-8. A straight 8 in a cabinet, with DECtape. Later
> upgraded to 8K words (8192 x 12 bits). You had to bank-switch the
> memory to use it all.

I've built a couple of SBC-6120s, and a PiDP-8. I have wanted an 8 since
1973.

NO real ones. But I do have four 11s and three VAXes!

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

On Fri, 11 Nov 2022 18:16:36 -0700, Peter Flass wrote:

> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>> On 11/11/2022 03:45, Rich Alderson wrote:
>>> Then that's a function of your failing memory.
>>
>> Agreed. Thank you, and also Bob for the doc link.
>>
>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>
> I thought it was the acme of instruction sets. Everything was downhill
> from there.

I rather like the ICL 2900 (based loosely on MU5). I plan to do a web
page about it soon.

Single accumulator, single index register, single descriptor register.
Stack frame and top registers and two off-stack pointers. 128
instructions and 32 addressing modes.

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

On 12 Nov 2022 10:19:48 GMT
Bob Eager <news0009@eager.cx> wrote:

> On Fri, 11 Nov 2022 18:16:36 -0700, Peter Flass wrote:
>
> > Vir Campestris <vir.campestris@invalid.invalid> wrote:
> >> On 11/11/2022 03:45, Rich Alderson wrote:
> >>> Then that's a function of your failing memory.
> >>
> >> Agreed. Thank you, and also Bob for the doc link.
> >>
> >> I really liked the PDP-10 instruction set. OTOH, it was my first!
> >
> > I thought it was the acme of instruction sets. Everything was downhill
> > from there.
>
> I rather like the ICL 2900 (based loosely on MU5). I plan to do a web
> page about it soon.
>
> Single accumulator, single index register, single descriptor register.
> Stack frame and top registers and two off-stack pointers. 128
> instructions and 32 addressing modes.
>
>
Sounds ideal for a Forth!

--
Bah, and indeed Humbug.

Peter Moylan <peter@pmoylan.org.invalid> wrote:
> On 11/11/22 14:47, Rich Alderson wrote:
>> "Anders D. Nygaard" <news2012adn@gmail.com> writes:
>>
>>> Den 10-11-2022 kl. 18:04 skrev Tak To:
>>>> On 11/10/2022 8:40 AM, Peter Flass wrote:
>>>>> [... I ...] think systems with 36-bit words that stored
>>>>> characters in 9 bits may have called them bytes.
>>>>> “Characters” was the term of art earlier.
>>>>
>>>> The only system I know of that fits that description is Multics,
>>>> and I don't remember how those 9-bit entities were called.
>>>
>>> My memory is *very* hazy, but the system I used in my first year
>>> at university answers to that description. I'm fairly sure it was
>>> a UNIVAC; probably a model 1100.
>>>
>>> Since then, every byte I've come across has been 8 bits.
>>
>> Because the 400kg gorilla from Armonk changed the definition.
>
> I don't think that was the main reason. There were, in my opinion, two
> more important factors.
>
> 1. In the early days of computing, BCD was very important. That set a
> precedent that suggested that the best choice of word size is a multiple
> of 4 bits.
>
> 2. The choice of character width was all over the place, but gradually
> the design of serial I/O interfaces - and serial was more pervasive than
> parallel back in those days - was converging to the idea that a
> character code should be 7 bits plus one parity bits. (Plus some start
> and stop bits, but those didn't have to be stored once the character
> reached the computer.) Thus, there was heavy off-she-shelf hardware
> support for 8-bit characters.
>
> These days we accept that a parity bit (or other form of
> error-detecting/correcting) is needed only while passing the information
> over a communications channel, and can be ignored once the character is
> in computer memory, but that wasn't the thinking back then.
>

Back then memories were a lot flakier, although now I think data is stored
with lots of check bits that only the hardware sees.

--
Pete

On 2022-11-12 03:50, Peter Moylan wrote:
> On 12/11/22 12:16, Peter Flass wrote:
>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>> On 11/11/2022 03:45, Rich Alderson wrote:
>
>>>> Then that's a function of your failing memory.
>>>
>>> Agreed. Thank you, and also Bob for the doc link.
>>>
>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>
>> I thought it was the acme of instruction sets. Everything was
>> downhill from there.
>
> I still give top score to the PDP-11. OK, it was not as powerful as the
> PDP-10, but the design of its addressing modes was brilliant.
>
> Once microprocessors started to appear, it was obvious (from the
> instruction sets) which designers had an IBM computer background and
> which ones had a DEC computer background. And, just as with mainframes,
> it soon became clear that there was no correlation between elegance of
> design and market dominance.

Boy howdy!

I never did like the 8080 or any of its successors, and the RCA COSMAC,
Z80, and PIC were in the same category. Actually the PIC was in a class
of its own, the worst microprocessor I ever used.

The ones I used and liked were the Signetics 2650, the 6502 and the
ones that followed it, and the Motorola chips.

The ones I didn't like would have been perfectly acceptable, of course,
had there been a high level language available when I was programming then.

--
If swimming is so good for your figure, how do you explain whales?

On 12/11/2022 09:50, Peter Moylan wrote:
> I still give top score to the PDP-11. OK, it was not as powerful as the
> PDP-10, but the design of its addressing modes was brilliant.
>
> Once microprocessors started to appear, it was obvious (from the
> instruction sets) which designers had an IBM computer background and
> which ones had a DEC computer background. And, just as with mainframes,
> it soon became clear that there was no correlation between elegance of
> design and market dominance.
>

Indeed. From the point of view of elegance of instruction set and the
joy of writing assember therein, it's a pity that the Motorola
68000 series ws not the choice for the IBM PC

On 12/11/2022 11:22, Kerr-Mudd, John wrote:
> On 12 Nov 2022 10:19:48 GMT
> Bob Eager <news0009@eager.cx> wrote:
>>
>> I rather like the ICL 2900 (based loosely on MU5). I plan to do a web
>> page about it soon.
>>
>> Single accumulator, single index register, single descriptor register.
>> Stack frame and top registers and two off-stack pointers. 128
>> instructions and 32 addressing modes.
>>
>>
> Sounds ideal for a Forth!

I wonder whether perhaps exposure to the architecture of the KDF9
was the inspiration for Charles Moore to invent FORTH

On 11/12/2022 4:50 AM, Peter Moylan wrote:
> On 12/11/22 12:16, Peter Flass wrote:
>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>> On 11/11/2022 03:45, Rich Alderson wrote:
>
>>>> Then that's a function of your failing memory.
>>>
>>> Agreed. Thank you, and also Bob for the doc link.
>>>
>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>
>> I thought it was the acme of instruction sets. Everything was
>> downhill from there.
>
> I still give top score to the PDP-11. OK, it was not as powerful as the
> PDP-10, but the design of its addressing modes was brilliant.

Agree. The PDP-10 instruction set was the acme in orthogonality,
but at the cost of having too many no-ops. All in all, it was
doomed by the limitation of fixed length instructions
architecture.

Byte-addressable and variable length instructions was the
future and IBM got it right, if not being the first one to
do so.

> Once microprocessors started to appear, it was obvious (from the
> instruction sets) which designers had an IBM computer background and
> which ones had a DEC computer background.
> And, just as with mainframes,
> it soon became clear that there was no correlation between elegance of
> design and market dominance.

Elegance is hard to compare without a common set of desiderata.

--
Tak
----------------------------------------------------------------+-----
Tak To takto@alum.mit.eduxx
--------------------------------------------------------------------^^
[taode takto ~{LU5B~}] NB: trim the xx to get my real email addr

On 11/12/2022 10:32 AM, gareth evans wrote:
> On 12/11/2022 09:50, Peter Moylan wrote:
>> I still give top score to the PDP-11. OK, it was not as powerful as the
>> PDP-10, but the design of its addressing modes was brilliant.
>>
>> Once microprocessors started to appear, it was obvious (from the
>> instruction sets) which designers had an IBM computer background and
>> which ones had a DEC computer background. And, just as with mainframes,
>> it soon became clear that there was no correlation between elegance of
>> design and market dominance.
>
> Indeed. From the point of view of elegance of instruction set and the
> joy of writing assember therein, it's a pity that the Motorola
> 68000 series ws not the choice for the IBM PC

The Motorola 68000 was not ready at the time of the first PC,
and the future of the personal computer market was read
horribly incorrectly by IBM (and DEC, and almost every key
player at that time).

Yes, there were many regrettable moments in the history of
computing. IBM did not partner with MIT to develop Multics
-- to name just one.

--
Tak
----------------------------------------------------------------+-----
Tak To takto@alum.mit.eduxx
--------------------------------------------------------------------^^
[taode takto ~{LU5B~}] NB: trim the xx to get my real email addr

[F'up]

Tak To <takto@alum.mit.eduxx> schrieb:
> On 11/12/2022 4:50 AM, Peter Moylan wrote:
>> On 12/11/22 12:16, Peter Flass wrote:
>>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>>> On 11/11/2022 03:45, Rich Alderson wrote:
>>
>>>>> Then that's a function of your failing memory.
>>>>
>>>> Agreed. Thank you, and also Bob for the doc link.
>>>>
>>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>>
>>> I thought it was the acme of instruction sets. Everything was
>>> downhill from there.
>>
>> I still give top score to the PDP-11. OK, it was not as powerful as the
>> PDP-10, but the design of its addressing modes was brilliant.
>
> Agree. The PDP-10 instruction set was the acme in orthogonality,
> but at the cost of having too many no-ops. All in all, it was
> doomed by the limitation of fixed length instructions
> architecture.

Fixed length instruction is not a problem per se, as can
be seen from numerous RISC chips.

But it would have been interesting to see what a 72-bit
version (who needs a tiny 64-bit address space, anyway :-)
would have looked like.

> Byte-addressable and variable length instructions was the
> future and IBM got it right, if not being the first one to
> do so.

The /360 was indeed groundbreaking. The eight-bit byte (even if
it was motivated by BCD) changed computers in a fundamental way.
It also had its faults, lots of them, found in hindsight and later
corrected in subsequent revisions and in other instruction set
architectures.

>> Once microprocessors started to appear, it was obvious (from the
>> instruction sets) which designers had an IBM computer background and
>> which ones had a DEC computer background.
>> And, just as with mainframes,
>> it soon became clear that there was no correlation between elegance of
>> design and market dominance.
>
> Elegance is hard to compare without a common set of desiderata.

There is a kind of elegance to more or less pure RISC designs like
the MIPS and the Alpha, but they have their warts, too.

lar3ryca <larry@invalid.ca> wrote:
> On 2022-11-12 03:50, Peter Moylan wrote:
>> On 12/11/22 12:16, Peter Flass wrote:
>>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>>> On 11/11/2022 03:45, Rich Alderson wrote:
>>
>>>>> Then that's a function of your failing memory.
>>>>
>>>> Agreed. Thank you, and also Bob for the doc link.
>>>>
>>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>>
>>> I thought it was the acme of instruction sets. Everything was
>>> downhill from there.
>>
>> I still give top score to the PDP-11. OK, it was not as powerful as the
>> PDP-10, but the design of its addressing modes was brilliant.
>>
>> Once microprocessors started to appear, it was obvious (from the
>> instruction sets) which designers had an IBM computer background and
>> which ones had a DEC computer background. And, just as with mainframes,
>> it soon became clear that there was no correlation between elegance of
>> design and market dominance.
>
> Boy howdy!
>
> I never did like the 8080 or any of its successors, and the RCA COSMAC,
> Z80, and PIC were in the same category. Actually the PIC was in a class
> of its own, the worst microprocessor I ever used.
>
> The ones I used and liked were the Signetics 2650, the 6502 and the
> ones that followed it, and the Motorola chips.
>
> The ones I didn't like would have been perfectly acceptable, of course,
> had there been a high level language available when I was programming then.
>

When microprocessors first started to get popular I read a book that
compared the then-popular chips: 6502, Cosmac, 8080, 6800, (and maybe
etc.). Looking at the instruction sets, interfacing considerations, etc., I
decided that hands-down the best was the 6800, and the 6502 was well down
the list. Well, it was VHS vs. Beta all over again.

--
Pete

Thomas Koenig <tkoenig@netcologne.de> wrote:
> [F'up]
>
> Tak To <takto@alum.mit.eduxx> schrieb:
>> On 11/12/2022 4:50 AM, Peter Moylan wrote:
>>> On 12/11/22 12:16, Peter Flass wrote:
>>>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>>>> On 11/11/2022 03:45, Rich Alderson wrote:
>>>
>>>>>> Then that's a function of your failing memory.
>>>>>
>>>>> Agreed. Thank you, and also Bob for the doc link.
>>>>>
>>>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>>>
>>>> I thought it was the acme of instruction sets. Everything was
>>>> downhill from there.
>>>
>>> I still give top score to the PDP-11. OK, it was not as powerful as the
>>> PDP-10, but the design of its addressing modes was brilliant.
>>
>> Agree. The PDP-10 instruction set was the acme in orthogonality,
>> but at the cost of having too many no-ops. All in all, it was
>> doomed by the limitation of fixed length instructions
>> architecture.
>
> Fixed length instruction is not a problem per se, as can
> be seen from numerous RISC chips.
>
> But it would have been interesting to see what a 72-bit
> version (who needs a tiny 64-bit address space, anyway :-)
> would have looked like.

Heck, these days you could write a simulator and have it perform reasonably
well, but, of course, lack of software would be a huge problem. You could
have it transparently run 36-bit stuff.

--
Pete

On Sat, 12 Nov 2022 12:55:24 -0500
Tak To <takto@alum.mit.eduxx> wrote:

> On 11/12/2022 10:32 AM, gareth evans wrote:

> > Indeed. From the point of view of elegance of instruction set and the
> > joy of writing assember therein, it's a pity that the Motorola
> > 68000 series ws not the choice for the IBM PC
>
> The Motorola 68000 was not ready at the time of the first PC,

This is true - but even the 6809 would have been nicer than the
8088.

> and the future of the personal computer market was read
> horribly incorrectly by IBM (and DEC, and almost every key
> player at that time).

Yep. I could have wished that IBM had left it a couple of years
longer before getting involved. A 68000 and XENIX with IBM leaning on the
license fee like Godzilla with a hangover would have been nice in 1985.
Instead we had to wait a while and get BSD or Linux on an 80386 (shudder)
in 1992.

--
Steve O'Hara-Smith
Odds and Ends at http://www.sohara.org/

On 11/12/2022 1:08 PM, Thomas Koenig wrote:
> [F'up]
>
> Tak To <takto@alum.mit.eduxx> schrieb:
>> On 11/12/2022 4:50 AM, Peter Moylan wrote:
>>> On 12/11/22 12:16, Peter Flass wrote:
>>>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>>>> On 11/11/2022 03:45, Rich Alderson wrote:
>>>
>>>>>> Then that's a function of your failing memory.
>>>>>
>>>>> Agreed. Thank you, and also Bob for the doc link.
>>>>>
>>>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>>>
>>>> I thought it was the acme of instruction sets. Everything was
>>>> downhill from there.
>>>
>>> I still give top score to the PDP-11. OK, it was not as powerful as the
>>> PDP-10, but the design of its addressing modes was brilliant.
>>
>> Agree. The PDP-10 instruction set was the acme in orthogonality,
>> but at the cost of having too many no-ops. All in all, it was
>> doomed by the limitation of fixed length instructions
>> architecture.
>
> Fixed length instruction is not a problem per se, as can
> be seen from numerous RISC chips.

.... only when suitably pipe-lined.

> But it would have been interesting to see what a 72-bit
> version (who needs a tiny 64-bit address space, anyway :-)
> would have looked like.

No doubt advantageous to some kind of programs and dis-
advantageous to others.

>> Byte-addressable and variable length instructions was the
>> future and IBM got it right, if not being the first one to
>> do so.
>
> The /360 was indeed groundbreaking. The eight-bit byte (even if
> it was motivated by BCD) changed computers in a fundamental way.
> It also had its faults, lots of them, found in hindsight and later
> corrected in subsequent revisions and in other instruction set
> architectures.
>
>>> Once microprocessors started to appear, it was obvious (from the
>>> instruction sets) which designers had an IBM computer background and
>>> which ones had a DEC computer background.
>>> And, just as with mainframes,
>>> it soon became clear that there was no correlation between elegance of
>>> design and market dominance.
>>
>> Elegance is hard to compare without a common set of desiderata.
>
> There is a kind of elegance to more or less pure RISC designs like
> the MIPS and the Alpha, but they have their warts, too.

FWIW, the later PDP-10's were already micro-coded.

--
Tak
----------------------------------------------------------------+-----
Tak To takto@alum.mit.eduxx
--------------------------------------------------------------------^^
[taode takto ~{LU5B~}] NB: trim the xx to get my real email addr

Ahem A Rivet's Shot <steveo@eircom.net> schrieb:
> On Sat, 12 Nov 2022 12:55:24 -0500
> Tak To <takto@alum.mit.eduxx> wrote:
>
>> On 11/12/2022 10:32 AM, gareth evans wrote:
>
>> > Indeed. From the point of view of elegance of instruction set and the
>> > joy of writing assember therein, it's a pity that the Motorola
>> > 68000 series ws not the choice for the IBM PC
>>
>> The Motorola 68000 was not ready at the time of the first PC,
>
> This is true - but even the 6809 would have been nicer than the
> 8088.

The 6809 only has a 16-bit address bus, and even the severely
crippeled 8088 can address a megabyte (insert fear and loathing
about just _how_).

Now, if Motorola had built an ARM v2 instead of the 68000 (entirely
possible at that timeframe, it had far fewer transistors, and much
higher performance) history might have been different.

>
>> and the future of the personal computer market was read
>> horribly incorrectly by IBM (and DEC, and almost every key
>> player at that time).
>
> Yep. I could have wished that IBM had left it a couple of years
> longer before getting involved. A 68000 and XENIX with IBM leaning on the
> license fee like Godzilla with a hangover would have been nice in 1985.
> Instead we had to wait a while and get BSD or Linux on an 80386 (shudder)
> in 1992.

I well remember downloading Slackware on a stack of floppy discs.
At University, I had already worked with HP workstations, which
were a revelation compared to mainframes.

Peter Flass <peter_flass@yahoo.com> schrieb:

> When microprocessors first started to get popular I read a book that
> compared the then-popular chips: 6502, Cosmac, 8080, 6800, (and maybe
> etc.). Looking at the instruction sets, interfacing considerations, etc., I
> decided that hands-down the best was the 6800, and the 6502 was well down
> the list. Well, it was VHS vs. Beta all over again.

Of course, the 6502 was designed by many of the original 6800
design team. You can see the relationship in the opcodes, and
you can even see it in the die shots which look quite similar.

Where it beat the 6800 hands down was price. A 6800 cost $360 at
the time of the 6502's introduction, the 6502 $25 (if I read
the soruces right).

On 11/12/2022 2:25 PM, Thomas Koenig wrote:
> Ahem A Rivet's Shot <steveo@eircom.net> schrieb:
>> On Sat, 12 Nov 2022 12:55:24 -0500
>> Tak To <takto@alum.mit.eduxx> wrote:
>>
>>> On 11/12/2022 10:32 AM, gareth evans wrote:
>>
>>>> Indeed. From the point of view of elegance of instruction set and the
>>>> joy of writing assember therein, it's a pity that the Motorola
>>>> 68000 series ws not the choice for the IBM PC
>>>
>>> The Motorola 68000 was not ready at the time of the first PC,
>>
>> This is true - but even the 6809 would have been nicer than the
>> 8088.
>
> The 6809 only has a 16-bit address bus, and even the severely
> crippeled 8088 can address a megabyte (insert fear and loathing
> about just _how_).
>
> Now, if Motorola had built an ARM v2 instead of the 68000 (entirely
> possible at that timeframe, it had far fewer transistors, and much
> higher performance) history might have been different.

Or if DEC has come up with a single chip LSI-11...

>>> and the future of the personal computer market was read
>>> horribly incorrectly by IBM (and DEC, and almost every key
>>> player at that time).
>>
>> Yep. I could have wished that IBM had left it a couple of years
>> longer before getting involved. A 68000 and XENIX with IBM leaning on the
>> license fee like Godzilla with a hangover would have been nice in 1985.
>> Instead we had to wait a while and get BSD or Linux on an 80386 (shudder)
>> in 1992.
>
> I well remember downloading Slackware on a stack of floppy discs.
> At University, I had already worked with HP workstations, which
> were a revelation compared to mainframes.

--
Tak
----------------------------------------------------------------+-----
Tak To takto@alum.mit.eduxx
--------------------------------------------------------------------^^
[taode takto ~{LU5B~}] NB: trim the xx to get my real email addr

On 2022-11-12, Peter Moylan <peter@pmoylan.org.invalid> wrote:

> On 12/11/22 12:16, Peter Flass wrote:
>
>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>
>>> On 11/11/2022 03:45, Rich Alderson wrote:
>>>
>>>> Then that's a function of your failing memory.
>>>
>>> Agreed. Thank you, and also Bob for the doc link.
>>>
>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>
>> I thought it was the acme of instruction sets. Everything was
>> downhill from there.
>
> I still give top score to the PDP-11. OK, it was not as powerful as the
> PDP-10, but the design of its addressing modes was brilliant.

I never got into the PDP-10, but I remember being dazzled by the elegance
of the PDP-11 when I started analyzing some of its machine code. Making
the program counter and stack pointer just another couple of registers
allowed some nifty tricks.

> Once microprocessors started to appear, it was obvious (from the
> instruction sets) which designers had an IBM computer background and
> which ones had a DEC computer background. And, just as with mainframes,
> it soon became clear that there was no correlation between elegance of
> design and market dominance.

Yes, the dominance of the 80x86 over the 680x0 is a tragic reminder
of the difference between doing it right and doing it right now.

--
/~\ Charlie Gibbs | Microsoft is a dictatorship.
\ / <cgibbs@kltpzyxm.invalid> | Apple is a cult.
X I'm really at ac.dekanfrus | Linux is anarchy.
/ \ if you read it the right way. | Pick your poison.

On 11/12/2022 2:42 PM, Charlie Gibbs wrote:
> On 2022-11-12, Peter Moylan <peter@pmoylan.org.invalid> wrote:
>
>> On 12/11/22 12:16, Peter Flass wrote:
>>
>>> Vir Campestris <vir.campestris@invalid.invalid> wrote:
>>>
>>>> On 11/11/2022 03:45, Rich Alderson wrote:
>>>>
>>>>> Then that's a function of your failing memory.
>>>>
>>>> Agreed. Thank you, and also Bob for the doc link.
>>>>
>>>> I really liked the PDP-10 instruction set. OTOH, it was my first!
>>>
>>> I thought it was the acme of instruction sets. Everything was
>>> downhill from there.
>>
>> I still give top score to the PDP-11. OK, it was not as powerful as the
>> PDP-10, but the design of its addressing modes was brilliant.
>
> I never got into the PDP-10, but I remember being dazzled by the elegance
> of the PDP-11 when I started analyzing some of its machine code. Making
> the program counter and stack pointer just another couple of registers
> allowed some nifty tricks.
>
>> Once microprocessors started to appear, it was obvious (from the
>> instruction sets) which designers had an IBM computer background and
>> which ones had a DEC computer background. And, just as with mainframes,
>> it soon became clear that there was no correlation between elegance of
>> design and market dominance.
>
> Yes, the dominance of the 80x86 over the 680x0 is a tragic reminder
> of the difference between doing it right and doing it right now.

The 80386 wasn't behind the 68020 by that much. The real
tragedy was that IBM-Microsoft essentially sat on the
technology for close to 10 years(!!!) and did not upgrade
the OS accordingly.

--
Tak
----------------------------------------------------------------+-----
Tak To takto@alum.mit.eduxx
--------------------------------------------------------------------^^
[taode takto ~{LU5B~}] NB: trim the xx to get my real email addr

On 12/11/2022 5:55 pm, Tak To wrote:
> On 11/12/2022 10:32 AM, gareth evans wrote:

<snip>

>> Indeed. From the point of view of elegance of instruction set and the
>> joy of writing assember therein, it's a pity that the Motorola
>> 68000 series ws not the choice for the IBM PC
>
> The Motorola 68000 was not ready at the time of the first PC,
> and the future of the personal computer market was read
> horribly incorrectly by IBM (and DEC, and almost every key
> player at that time).

And they also famously turned down the Beatles.

--
Richard Heathfield
Email: rjh at cpax dot org dot uk
"Usenet is a strange place" - dmr 29 July 1999
Sig line 4 vacant - apply within

Ahem A Rivet's Shot <steveo@eircom.net> wrote:
> On Sat, 12 Nov 2022 12:55:24 -0500
> Tak To <takto@alum.mit.eduxx> wrote:
>
>> On 11/12/2022 10:32 AM, gareth evans wrote:
>
>>> Indeed. From the point of view of elegance of instruction set and the
>>> joy of writing assember therein, it's a pity that the Motorola
>>> 68000 series ws not the choice for the IBM PC
>>
>> The Motorola 68000 was not ready at the time of the first PC,
>
> This is true - but even the 6809 would have been nicer than the
> 8088.
>
>> and the future of the personal computer market was read
>> horribly incorrectly by IBM (and DEC, and almost every key
>> player at that time).
>
> Yep. I could have wished that IBM had left it a couple of years
> longer before getting involved. A 68000 and XENIX with IBM leaning on the
> license fee like Godzilla with a hangover would have been nice in 1985.
> Instead we had to wait a while and get BSD or Linux on an 80386 (shudder)
> in 1992.
>

The IBM System 9000 with a 68000 CPU came out in 1982 and ran XENIX.

--
Pete