Re: Oh why did Apple dump IBM....

Richard Tobin <richard@xxxxxxxxxxxxxxx> wrote:

In article <1i8xpti.tybiu0141e5xnN%real-address-in-sig@xxxxxxxxxxxxxxx>,
Rowland McDonnell <real-address-in-sig@xxxxxxxxxxxxxxx> wrote:

I thought it was obvious I was talking from that point of view.

But the point is that `to program in machine code' is not why we use
computers.

Actually, when I used things like 6502s that *was* the point for
people like me. We didn't use development boards because we really
wanted to control our central heating more efficiently. It was for
the fun of programming them.

Oh, okay. But I thought we weren't supposed to admit to that sort of
fetish? I came into PCs when we had Basic interpreters with everything
pretty much, so I never did a lot of machine code. It's a pain when you
have to assemble on paper in any case.

I programmed the 8080, the 6502, the PDP-11, and the "SC/MP" at about
the same time. I hadn't done any assembly language programming before
that, so I don't think I can have had that expectation.

<puzzled> Why not? You programmed all of those, got the hang of the
different ways of doing things, got used to the niceness of having
complex instructions and big register sets ('cos it does make life more
comfy for the programmer which is why they did it in the first place) -
and when those things were missing, you missed them.

My point was that it wasn't some pre-existing bias, but a view resulting
from using them all. Your comment:

by which you mean
`tedious to program for someone who expects a large general purpose
register set'.

suggested (to me) that it's just a matter of what you're used to. But
I wasn't used to anything. I conclude that a large general purpose
register set really is better for programming.

OH yes - not that you had such a set available on any microprocessor
until some time in the development of the 68000 series, though.

Not that that matters: the point about the 6502 was that you could
afford to buy one to program when you might not have been able to
afford to buy an 8080 to program,

I don't remember how much they cost - not that I ever purchased any
with my own money.

I'd have to look up the details, but ISTR that the 6502 was something
like 1/8th the price of the 8080 when it was launched - cheaper than the
6800, too, so I gather.

It also did more per clock tick.

I'm not sure why that's an advantage. For years many RISC processors
wre faster than x86s, despite doing less per clock tick. If doing
more per clock means you can't run the clock so fast, it's not an
advantage.

A 2MHz 6502 box beat a 4MHz Z80 box any time I used 'em - even with the
usual slower clock rate, the 6502 won. On top of that, the 6502 was the
reduced instruction set non-microcoded pipelined part, while the faster
clock speed Z80 was the microcoded CISC slug. Yes, that's how they
were.

The Z80 /needed/ to have a higher clock speed because of its `four clock
ticks per machine cycle' thingy. A 6502 could do things in two clock
ticks.

Things didn't divide up in the obvious ways back then. It always looked
to me as if a 4MHz Z80 ought to be up with a 2MHz 6502, but I never met
a 4MHz Z80 box that felt fast to me, while I met a lot of 2MHz 6502
boxes that did.

Which way it turns out seems to vary over time: current
x86s have gained much of their speed by doing far more per clock than
they did 15 years ago. As I remember it, the 6502 generally came at
lower clock speeds than the 8080 and Z80.

OH yes - and it was the CPU of choice for the fast PCs way back when.
Yer BBC Micro blows all other 8 bit home micros into the weeds in terms
of sheer speed.

The ARM is, from one point of view, a 6502 `done properly' - Acorn used
the 6502 as a conceptual starting point and `just did it right, with
full on RISC, 32 bits, a nice interrupt architecture, and so on'.

Anyway, point is, there's more to microprocessor elegance than the
suitability of the instruction set for programming ease. There's X per
Y figures to consider too. Price per CPU, price per processing power,
power consumption per unit, power consumption per processing power, and
so on.

Yes of course... but my comments were entirely prompted by your claim
that the 6502 was "pretty damned elegant" in its ISA.

<cough> I think you'll find that I said the 6502 was pretty damned
elegant - the idea is that when considered as a complete part in full
context, it's elegant: the ISA works out elegant only because it's the
consequence of the 6502 being a dirt cheap sawn-off speed demon.

I don't think
I disagree with any of your other comments.

You mistake me for someone who is making an argument about which
processor was more cost-effective. I'm just recalling how pleasant
I found programming them.

ARGH! Yeah, but how does that fit in to what I was saying?

See above!

Still don't get it - never mind.

Rowland.

--
Remove the animal for email address: rowland.mcdonnell@xxxxxxxxxxxxxxx
Sorry - the spam got to me
http://www.mag-uk.org http://www.bmf.co.uk
UK biker? Join MAG and the BMF and stop the Eurocrats banning biking
.

Relevant Pages

  • Re: single daemon startup
    ... simple guy programming a daemon that runs just one copy. ... So you may create a virtual clock owned by a process group. ...
    (comp.unix.programmer)
  • Re: Any way to upgrade CMOS clock??
    ... modify whatever software grabs the time of day from the regular system clock. ... Any way to upgrade CMOS clock?? ... Timecode is a good start, but it needs a sequence number added to it, at ... Bad programming is based on bad assumptions. ...
    (alt.sys.pc-clone.dell)
  • Re: computer-driven fans
    ... The ISP programming clock should be set to no more ... the chip can appear to be faulty. ...
    (comp.arch.embedded)
  • Re: How Overclock 50g Hp calculator?
    ... i thinking about programming in C, but i see that the result file size ... if the answer is yes, this be a great, because you can adapt the clock ... the clock increment the battery usage, and if you can disable the LCD ... Consider this;) (please sorry the english, ...
    (comp.sys.hp48)
  • Re: Any way to upgrade CMOS clock??
    ... Any way to upgrade CMOS clock?? ... Timecode is a good start, but it needs a sequence number added to it, at ... Bad programming is based on bad assumptions. ...
    (alt.sys.pc-clone.dell)
Loading