Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #176200 > unrolled thread

Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?]

Back to article view | Back to linux.debian.user

Contents

  Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?] Joel Rees <joel.rees@gmail.com> - 2017-01-01 16:00 +0100
    Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?] Lisi Reisz <lisi.reisz@gmail.com> - 2017-01-01 17:10 +0100
      Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?] deloptes <deloptes@gmail.com> - 2017-01-01 20:00 +0100
      Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?] rhkramer@gmail.com - 2017-01-01 20:20 +0100
      Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?] Joel Rees <joel.rees@gmail.com> - 2017-01-02 00:50 +0100
        Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?] Lisi Reisz <lisi.reisz@gmail.com> - 2017-01-02 02:10 +0100

#176200 — Re: [OT] Antikythera mechanism [was Re: Do have programs have poor documentation?]

On Sat, Dec 31, 2016 at 11:11 AM, Miles Fidelman
<mfidelman@meetinghouse.net> wrote:
> Talk about a thread going South!  (Perhaps we can get back to bashing
> systemd?)

There's a chuckle.

> On 12/30/16 7:07 PM, deloptes wrote:
>>>
>>> In what way is the Antikythera mechanism not a computer?  And where did
>>> your 400 years come from?

Without a functioning Antikythera mechanism, we really can't answer
that question in a useful manner. However, I could guess that I could
not program that machine with anything that looks like a full C
compiler.

(Guess. For all we know, there were nanotech mechanical CPUs in the
thing before the seawater made it non-functional.)

Subset C, maybe. The difference is important.

>> I understand what you mean, but it was in the last 400y that this machine
>> took shape. In fact it was Turing that defined it. But he would not be
>> able
>> to define it if it was not the mathematicians before him. I agree with you
>> as well, we could go to the roots of mathematics, however even if the
>> definition of such a machine was so old, it wouldn't be possible to build
>> it without the technical advantage, so ... I still think my statement is
>> true. You can argue as long as you will.
>
>
> Well, you kind of forget:
> Joseph Jacquard (and maybe Basile Bouchon)
> not to mention Charles Babbage and Ada Lovelance
> Alonzo Church.
> And of course,  John von Neumann (if you want to talk actual hardware
> architecture)

Interesting thing about the siggie and the above.

> --
> In theory, there is no difference between theory and practice.
> In practice, there is.  .... Yogi Berra

Intel really still wants us to believe that the 8086, because it was
Turing complete -- other than the memory limitation (cough) -- was
equivalent to the 68000. Nobody in their right mind used an 8086 to
control an engine, however. (But we do now use subsets of the Power PC
architecture and variations of the SH architecture.)

You can program PLAs in something that looks like a subset of C, but
it's not the same.

You can construct a CPU with a PLA, but you can get much more energy
efficiency and much better CPU speed by laying out the various CPU
parts as dedicated blocks of logic.

On the converse, simulating a switch grid with a CPU introduces
serious inefficiencies, as well.

Different classes of complexity. Programming, but not the same kind of programs.

The info system is another example. Very powerful, but I didn't want
to have to learn the info system just to wade through the info info
pages. It was very intuitive for someone who already had certain
keyboard habits, but not for the rest of us. Keyboard macros are not
the same thing as Forth or LISP primitives or M4 or cpp macrose.

html is a bit less obtuse than info, less concise, and a shallower
learning curve.

And plain text coupled with the apropos command (man -k), with the
in-page search function, still get me a lot farther into something
new, quicker, than info files. Much less keyboard dancing.

My personal vote for the original topic is man 7, as someone else
mentioned. (Yes, the man pages did, from back in the system 6 days,
even, include a _little_ bit of tutorial.)

-- 
Joel Rees

I'm imagining I'm a novelist:
http://reiisi.blogspot.jp/p/novels-i-am-writing.html

[toc] | [next] | [standalone]

#176204

On Sunday 01 January 2017 14:54:09 Joel Rees wrote:
> On Sat, Dec 31, 2016 at 11:11 AM, Miles Fidelman
>
> <mfidelman@meetinghouse.net> wrote:

> > On 12/30/16 7:07 PM, deloptes wrote:
No, I wrote that.

> >>> In what way is the Antikythera mechanism not a computer?  And where did
> >>> your 400 years come from?
>
> Without a functioning Antikythera mechanism, we really can't answer
> that question in a useful manner. However, I could guess that I could
> not program that machine with anything that looks like a full C
> compiler.

So something that can't be programmed with anything that looks like a full C 
compiler is not a computer???  So Colossus was not a computer?? :-)

C itself, of course, is MUCH later than Colossus, 

<quote> C was originally developed by Dennis Ritchie between 1969 and 1973 at 
Bell Labs, </quote>
https://en.wikipedia.org/wiki/C_(programming_language)

<quote>
Colossus was a set of computers developed by British codebreakers in 
1943-1945 .... The prototype, Colossus Mark 1, was shown to be working in 
December 1943 
</quote>
https://en.wikipedia.org/wiki/Colossus_computer

but I don't think that Colossus could compile with anything.  It had to be 
directly programmed.

Lisi
> (Guess. For all we know, there were nanotech mechanical CPUs in the
> thing before the seawater made it non-functional.)
>
> Subset C, maybe. The difference is important.
>
> >> I understand what you mean, but it was in the last 400y that this
> >> machine took shape. In fact it was Turing that defined it. But he would
> >> not be able
> >> to define it if it was not the mathematicians before him. I agree with
> >> you as well, we could go to the roots of mathematics, however even if
> >> the definition of such a machine was so old, it wouldn't be possible to
> >> build it without the technical advantage, so ... I still think my
> >> statement is true. You can argue as long as you will.
> >
> > Well, you kind of forget:
> > Joseph Jacquard (and maybe Basile Bouchon)
> > not to mention Charles Babbage and Ada Lovelance
> > Alonzo Church.
> > And of course,  John von Neumann (if you want to talk actual hardware
> > architecture)
>
> Interesting thing about the siggie and the above.
>
> > --
> > In theory, there is no difference between theory and practice.
> > In practice, there is.  .... Yogi Berra
>
> Intel really still wants us to believe that the 8086, because it was
> Turing complete -- other than the memory limitation (cough) -- was
> equivalent to the 68000. Nobody in their right mind used an 8086 to
> control an engine, however. (But we do now use subsets of the Power PC
> architecture and variations of the SH architecture.)
>
> You can program PLAs in something that looks like a subset of C, but
> it's not the same.
>
> You can construct a CPU with a PLA, but you can get much more energy
> efficiency and much better CPU speed by laying out the various CPU
> parts as dedicated blocks of logic.
>
> On the converse, simulating a switch grid with a CPU introduces
> serious inefficiencies, as well.
>
> Different classes of complexity. Programming, but not the same kind of
> programs.
>
> The info system is another example. Very powerful, but I didn't want
> to have to learn the info system just to wade through the info info
> pages. It was very intuitive for someone who already had certain
> keyboard habits, but not for the rest of us. Keyboard macros are not
> the same thing as Forth or LISP primitives or M4 or cpp macrose.
>
> html is a bit less obtuse than info, less concise, and a shallower
> learning curve.
>
> And plain text coupled with the apropos command (man -k), with the
> in-page search function, still get me a lot farther into something
> new, quicker, than info files. Much less keyboard dancing.
>
> My personal vote for the original topic is man 7, as someone else
> mentioned. (Yes, the man pages did, from back in the system 6 days,
> even, include a _little_ bit of tutorial.)

[toc] | [prev] | [next] | [standalone]

#176215

Lisi Reisz wrote:

> So something that can't be programmed with anything that looks like a full
> C compiler is not a computer???  So Colossus was not a computer?? :-)
> 
> C itself, of course, is MUCH later than Colossus,
> 
> <quote> C was originally developed by Dennis Ritchie between 1969 and 1973
> at Bell Labs, </quote>
> https://en.wikipedia.org/wiki/C_(programming_language)
> 
> <quote>
> Colossus was a set of computers developed by British codebreakers in
> 1943-1945 .... The prototype, Colossus Mark 1, was shown to be working in
> December 1943
> </quote>
> https://en.wikipedia.org/wiki/Colossus_computer
> 
> but I don't think that Colossus could compile with anything.  It had to be
> directly programmed.
> 
> Lisi

I agree with you Lisi. Originally my point was about this tradition in
(christian) philosophy and mathematics that made this possible. Of course
in the beginning it was not like today (semiconductors), but still it is
about the idea. The idea (vision) was developed here and later came the
implementations. There are still nostalgic discussions about analog
computers. To me it is all part of a historic paradigm. 

regards

[toc] | [prev] | [next] | [standalone]

#176219

On Sunday, January 01, 2017 11:10:20 AM Lisi Reisz wrote:
> C itself, of course, is MUCH later than Colossus,

And, there were a lot of languages before C.

[toc] | [prev] | [next] | [standalone]

#176232

On Mon, Jan 2, 2017 at 1:10 AM, Lisi Reisz <lisi.reisz@gmail.com> wrote:
> On Sunday 01 January 2017 14:54:09 Joel Rees wrote:
>> On Sat, Dec 31, 2016 at 11:11 AM, Miles Fidelman
>>
>> <mfidelman@meetinghouse.net> wrote:
>
>> > On 12/30/16 7:07 PM, deloptes wrote:
> No, I wrote that.

Somebody besides me trimmed your name out of the quotes, and I didn't
check. Sorry.

>> >>> In what way is the Antikythera mechanism not a computer?  And where did
>> >>> your 400 years come from?
>>
>> Without a functioning Antikythera mechanism, we really can't answer
>> that question in a useful manner. However, I could guess that I could
>> not program that machine with anything that looks like a full C
>> compiler.
>
> So something that can't be programmed with anything that looks like a full C
> compiler is not a computer???  So Colossus was not a computer?? :-)

Shoot, a thermostat is a computer.

> C itself, of course, is MUCH later than Colossus,

Well, yeah.

Formal descriptions of procedures do, in fact, date back before the
Antikythera mech.

Most human languages do allow description of algorithms in a Turing
complete fashion, although the selection of the symbols and grammar
constructs is not clear in the old records. And, without a clear
delineation between the language being used for the algorithm and the
general language, things can get confusing.

I think said mechanism is thought to predate modern algebraic
notation, but algebraic notation is not Turing complete without some
parts that we usually don't deal with. The languages of the Calculus
and of formal logic almost get us there, but not quite.

> <quote> C was originally developed by Dennis Ritchie between 1969 and 1973 at
> Bell Labs, </quote>
> https://en.wikipedia.org/wiki/C_(programming_language)

... based on ideas and constructs evident in bcpl, Algol, and other
languages not too many people have heard of.

> <quote>
> Colossus was a set of computers developed by British codebreakers in
> 1943-1945 .... The prototype, Colossus Mark 1, was shown to be working in
> December 1943
> </quote>
> https://en.wikipedia.org/wiki/Colossus_computer
>
> but I don't think that Colossus could compile with anything.  It had to be
> directly programmed.

I'm sure that, if we wanted to, we could define a subset of C or some
other Turing complete language that could be compiled to a wiring
description for the Colossus.

Subset.

And it would be really easy to write programs in such a subset that
could not be run on the Colossus -- without a lot of hardware
augmentation. And the problem is not just lack of storage area.

> Lisi
>> (Guess. For all we know, there were nanotech mechanical CPUs in the
>> thing before the seawater made it non-functional.)
>>
>> Subset C, maybe. The difference is important.

So, do you think the differences are not important?

>> >> I understand what you mean, but it was in the last 400y that this
>> >> machine took shape. In fact it was Turing that defined it. But he would
>> >> not be able
>> >> to define it if it was not the mathematicians before him. I agree with
>> >> you as well, we could go to the roots of mathematics, however even if
>> >> the definition of such a machine was so old, it wouldn't be possible to
>> >> build it without the technical advantage, so ... I still think my
>> >> statement is true. You can argue as long as you will.
>> >
>> > Well, you kind of forget:
>> > Joseph Jacquard (and maybe Basile Bouchon)
>> > not to mention Charles Babbage and Ada Lovelance
>> > Alonzo Church.
>> > And of course,  John von Neumann (if you want to talk actual hardware
>> > architecture)
>>
>> Interesting thing about the siggie and the above.
>>
>> > --
>> > In theory, there is no difference between theory and practice.
>> > In practice, there is.  .... Yogi Berra
>[...]
>> My personal vote for the original topic is man 7, as someone else
>> mentioned. (Yes, the man pages did, from back in the system 6 days,
>> even, include a _little_ bit of tutorial.)

I'm not sure who's arguing what in these threads, but, for example,
documentation being inaccessible is a direct cause of a lot of code
churn in the community. (Not that code churn is inherently evil, but,
...)

And I've seen a lot of odd things asserted in these threads, like the
idea that man pages are not the place for tutorial content.

-- 
Joel Rees

I'm imagining I'm a novelist:
http://reiisi.blogspot.jp/p/novels-i-am-writing.html

[toc] | [prev] | [next] | [standalone]

#176238

On Sunday 01 January 2017 23:48:47 Joel Rees wrote:
> On Mon, Jan 2, 2017 at 1:10 AM, Lisi Reisz <lisi.reisz@gmail.com> wrote:
> > On Sunday 01 January 2017 14:54:09 Joel Rees wrote:
> >> On Sat, Dec 31, 2016 at 11:11 AM, Miles Fidelman
> >>
> >> <mfidelman@meetinghouse.net> wrote:
> >> > On 12/30/16 7:07 PM, deloptes wrote:
> >
> > No, I wrote that.
>
> Somebody besides me trimmed your name out of the quotes, and I didn't
> check. Sorry.
>
> >> >>> In what way is the Antikythera mechanism not a computer?  And where
> >> >>> did your 400 years come from?
> >>
> >> Without a functioning Antikythera mechanism, we really can't answer
> >> that question in a useful manner. However, I could guess that I could
> >> not program that machine with anything that looks like a full C
> >> compiler.
> >
> > So something that can't be programmed with anything that looks like a
> > full C compiler is not a computer???  So Colossus was not a computer??
> > :-)
>
> Shoot, a thermostat is a computer.
>
> > C itself, of course, is MUCH later than Colossus,
>
> Well, yeah.
>
> Formal descriptions of procedures do, in fact, date back before the
> Antikythera mech.
>
> Most human languages do allow description of algorithms in a Turing
> complete fashion, although the selection of the symbols and grammar
> constructs is not clear in the old records. And, without a clear
> delineation between the language being used for the algorithm and the
> general language, things can get confusing.
>
> I think said mechanism is thought to predate modern algebraic
> notation, but algebraic notation is not Turing complete without some
> parts that we usually don't deal with. The languages of the Calculus
> and of formal logic almost get us there, but not quite.
>
> > <quote> C was originally developed by Dennis Ritchie between 1969 and
> > 1973 at Bell Labs, </quote>
> > https://en.wikipedia.org/wiki/C_(programming_language)
>
> ... based on ideas and constructs evident in bcpl, Algol, and other
> languages not too many people have heard of.
>
> > <quote>
> > Colossus was a set of computers developed by British codebreakers in
> > 1943-1945 .... The prototype, Colossus Mark 1, was shown to be working in
> > December 1943
> > </quote>
> > https://en.wikipedia.org/wiki/Colossus_computer
> >
> > but I don't think that Colossus could compile with anything.  It had to
> > be directly programmed.
>
> I'm sure that, if we wanted to, we could define a subset of C or some
> other Turing complete language that could be compiled to a wiring
> description for the Colossus.
>
> Subset.
>
> And it would be really easy to write programs in such a subset that
> could not be run on the Colossus -- without a lot of hardware
> augmentation. And the problem is not just lack of storage area.
>
> > Lisi

I signed off here.  Nothing below this refers to me so I have not attempted to 
answer your comments on things said by other people.

Lisi
> >
> >> (Guess. For all we know, there were nanotech mechanical CPUs in the
> >> thing before the seawater made it non-functional.)
> >>
> >> Subset C, maybe. The difference is important.
>
> So, do you think the differences are not important?
>
> >> >> I understand what you mean, but it was in the last 400y that this
> >> >> machine took shape. In fact it was Turing that defined it. But he
> >> >> would not be able
> >> >> to define it if it was not the mathematicians before him. I agree
> >> >> with you as well, we could go to the roots of mathematics, however
> >> >> even if the definition of such a machine was so old, it wouldn't be
> >> >> possible to build it without the technical advantage, so ... I still
> >> >> think my statement is true. You can argue as long as you will.
> >> >
> >> > Well, you kind of forget:
> >> > Joseph Jacquard (and maybe Basile Bouchon)
> >> > not to mention Charles Babbage and Ada Lovelance
> >> > Alonzo Church.
> >> > And of course,  John von Neumann (if you want to talk actual hardware
> >> > architecture)
> >>
> >> Interesting thing about the siggie and the above.
> >>
> >> > --
> >> > In theory, there is no difference between theory and practice.
> >> > In practice, there is.  .... Yogi Berra
> >
> >[...]
> >
> >> My personal vote for the original topic is man 7, as someone else
> >> mentioned. (Yes, the man pages did, from back in the system 6 days,
> >> even, include a _little_ bit of tutorial.)
>
> I'm not sure who's arguing what in these threads, but, for example,
> documentation being inaccessible is a direct cause of a lot of code
> churn in the community. (Not that code churn is inherently evil, but,
> ...)
>
> And I've seen a lot of odd things asserted in these threads, like the
> idea that man pages are not the place for tutorial content.

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web