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

Groups > comp.lang.c > #91385 > unrolled thread

Re: [OT] Denis Ritchie Education

Back to article view | Back to comp.lang.c

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [OT] Denis Ritchie Education "J. Clarke" <j.clarke.873638@gmail.com> - 2016-07-31 05:30 -0400
    Re: [OT] Denis Ritchie Education Tim Rentsch <txr@alumni.caltech.edu> - 2016-07-31 12:40 -0700
      Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-07-31 16:58 -0400
        Re: [OT] Denis Ritchie Education Tim Rentsch <txr@alumni.caltech.edu> - 2016-07-31 15:13 -0700
          Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-07-31 21:21 -0400
            Re: [OT] Denis Ritchie Education Tim Rentsch <txr@alumni.caltech.edu> - 2016-08-01 08:12 -0700
        Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-07-31 16:41 -0600
          Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-07-31 21:24 -0400
            Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-07-31 22:15 -0600
              Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-08-01 09:38 -0400
                Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-08-01 10:18 -0600
                  Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-08-01 16:22 -0400
                    Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-08-01 15:05 -0600
                      Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-08-04 14:01 -0400
                        Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-08-04 12:22 -0600
                          Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-08-04 16:47 -0400
                        Re: [OT] Denis Ritchie Education "Osmium" <r124c4u102@comcast.net> - 2016-08-04 14:22 -0500
                          Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-08-04 16:52 -0400
                            Re: [OT] Denis Ritchie Education "Osmium" <r124c4u102@comcast.net> - 2016-08-04 16:10 -0500
                              Re: [OT] Denis Ritchie Education Jerry Stuckle <jstucklex@attglobal.net> - 2016-08-04 20:30 -0400
      Re: [OT] Denis Ritchie Education luser droog <luser.droog@gmail.com> - 2016-07-31 18:20 -0700
      Re: [OT] Denis Ritchie Education Gareth Owen <gwowen@gmail.com> - 2016-08-01 19:52 +0100
        Re: [OT] Denis Ritchie Education supercat@casperkitty.com - 2016-08-01 15:25 -0700
          Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-08-01 16:38 -0600
            Re: [OT] Denis Ritchie Education supercat@casperkitty.com - 2016-08-01 15:56 -0700
              Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-08-01 17:41 -0600
              Re: [OT] Denis Ritchie Education Gareth Owen <gwowen@gmail.com> - 2016-08-02 06:22 +0100
        Re: [OT] Denis Ritchie Education Tim Rentsch <txr@alumni.caltech.edu> - 2016-08-01 16:48 -0700
          Re: [OT] Denis Ritchie Education Gareth Owen <gwowen@gmail.com> - 2016-08-02 06:27 +0100
            Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-08-02 08:42 -0600
              Re: [OT] Denis Ritchie Education Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-08-02 16:28 +0100
                Re: [OT] Denis Ritchie Education Joe Pfeiffer <pfeiffer@cs.nmsu.edu> - 2016-08-02 11:15 -0600
                  Re: [OT] Denis Ritchie Education Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-08-02 19:37 +0100
                    Re: [OT] Denis Ritchie Education supercat@casperkitty.com - 2016-08-02 12:08 -0700
                      Re: [OT] Denis Ritchie Education Ben Bacarisse <ben.usenet@bsb.me.uk> - 2016-08-02 22:40 +0100
            Re: [OT] Denis Ritchie Education Tim Rentsch <txr@alumni.caltech.edu> - 2016-08-03 06:57 -0700
            Re: [OT] Denis Ritchie Education "J. Clarke" <j.clarke.873638@gmail.com> - 2016-10-30 09:06 -0400
              Re: [OT] Denis Ritchie Education supercat@casperkitty.com - 2016-10-30 12:42 -0700
              Re: [OT] Denis Ritchie Education Gareth Owen <gwowen@gmail.com> - 2016-10-31 18:12 +0000

Page 2 of 2 — ← Prev page 1 [2]

#91448

On Sunday, July 31, 2016 at 2:40:41 PM UTC-5, Tim Rentsch wrote:
> "J. Clarke" <j.clarke.873638@gmail.com> writes:
> 

> > Remember, though, this was Harvard, where the computer as we know it was
> > _invented_.  They were _doing_ what we now call "computer science" long
> > before there was a name for it.
> 
> I thought the computer as we know it was invented at the University
> of Pennsylvania and the group at the Institute for Advanced Studies.
> There were several precursors, including the Harvard Mark I, but
> they are not computers as we know them today.  In particular, IIUC,
> the Harvard Mark I was not Turing complete.

Indeed. Based on the biography of Howard Aiken the Mark I was explicitly 
controlled by a paper tape feed. There were pressures from many of the 
operators of the machine and its successor. IIRC it wasn't until the 
Mark III that there was a dedicated program store (Mercury-filled
acoustic-delay tubes) and not until the Mark IV that there was anything
resembling control instructions.

The big application for these things was solving differential equations.
And the "beta" to beat was differential analyzers -- large noisy 
machines something like a cross between a fooz-ball table and a
bowling lane. cf. http://www.jsoftware.com/papers/MSLDE.htm
One of the first commissions was naval ballistics tables.

> I don't mean to trivialize the contribution of the group at Harvard,
> which was substantial, but I don't think they invented the computer
> as we know it today.  John von Neumann, who played a role in both
> the UPenn and IAS groups, attributed the key concept to Turing's 1936
> paper.  Note that Turing spent almost two years at Princeton right
> around that time, next to von Neumann at the IAS.

Yes, but the real "feat" of the project appears to be persuading
IBM to pay for it (for which they did not receive the public credit
which they (rightly) anticipated).

-- 
Harvard architecture: code is code and data is data and ne'er the
twain shall they mix.

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

#91512

Tim Rentsch <txr@alumni.caltech.edu> writes:

> "J. Clarke" <j.clarke.873638@gmail.com> writes:
>
>> In article <nl966b$c57$1@dont-email.me>, nospam@nspam.invalid says...
>>
>>>     I was told that DMR' thesis was in "recursive functions".  Now I've
>>> always assumed maybe being wrong that back then this may have been a
>>> mathematics recursive functions and he was a Mathematician.  Does anyone know
>>> if his PhD was mathematics or computer science?
>>>
>>> Bill
>>
>> There's an interview that addresses this a little bit at
>> <https://www.princeton.edu/~hos/mike/transcripts/ritchie.htm>.
>>
>> Remember, though, this was Harvard, where the computer as we know it was
>> _invented_.  They were _doing_ what we now call "computer science" long
>> before there was a name for it.
>
> I thought the computer as we know it was invented at the University
> of Pennsylvania and the group at the Institute for Advanced Studies.
> There were several precursors, including the Harvard Mark I, but
> they are not computers as we know them today.  In particular, IIUC,
> the Harvard Mark I was not Turing complete.

Yeah, well as a holder of a degree from the University of Manchester,
I say YAH BOO SUCKS to all you lot.  

If it ain't executing a stored program, 
then it ain't a "computer as we know it".

(See also "The Myth of the Lone Inventor")

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

#91539

On Monday, August 1, 2016 at 1:52:10 PM UTC-5, gwowen wrote:
> If it ain't executing a stored program, 
> then it ain't a "computer as we know it".

Does that mean the 4004 shouldn't really be called the first microprocessor
since its design is not conducive to a stored-program microcomputer (I
haven't looked into the 8008, but the 8080 was of course used in a huge
number of stored-program microcomputers).

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

#91542

supercat@casperkitty.com writes:

> On Monday, August 1, 2016 at 1:52:10 PM UTC-5, gwowen wrote:
>> If it ain't executing a stored program, 
>> then it ain't a "computer as we know it".
>
> Does that mean the 4004 shouldn't really be called the first microprocessor
> since its design is not conducive to a stored-program microcomputer (I
> haven't looked into the 8008, but the 8080 was of course used in a huge
> number of stored-program microcomputers).

In what sense do you mean it isn't conducive to a stored-program
microcomputer?  It fetched instructions from ROM and executed them.

The 8008 was very similar to the 8080.

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

#91544

On Monday, August 1, 2016 at 5:38:37 PM UTC-5, Joe Pfeiffer wrote:
> In what sense do you mean it isn't conducive to a stored-program
> microcomputer?  It fetched instructions from ROM and executed them.

The things that are typically called "computers" today are capable of storing
programs into memory *themselves*, rather than merely executing instructions
from a read-only medium.  The 4004 used a separate address space for
instructions and data, and had no mechanism via which the processor could
write to the instruction store.

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

#91547

supercat@casperkitty.com writes:

> On Monday, August 1, 2016 at 5:38:37 PM UTC-5, Joe Pfeiffer wrote:
>> In what sense do you mean it isn't conducive to a stored-program
>> microcomputer?  It fetched instructions from ROM and executed them.
>
> The things that are typically called "computers" today are capable of storing
> programs into memory *themselves*, rather than merely executing instructions
> from a read-only medium.  The 4004 used a separate address space for
> instructions and data, and had no mechanism via which the processor could
> write to the instruction store.

That's a quibble I've seen (it's also a characteristic of many current
microcontrollers).  It seems more common, though, to refer to a computer
that uses the same memory for both program and data (and typically can
load its own programs) as a von Neumann or Princeton computer, while the
ones that keep instructions in a separate ROM are Harvard computers.
They're both stored program computers.

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

#91557

supercat@casperkitty.com writes:

> On Monday, August 1, 2016 at 5:38:37 PM UTC-5, Joe Pfeiffer wrote:
>> In what sense do you mean it isn't conducive to a stored-program
>> microcomputer?  It fetched instructions from ROM and executed them.
>
> The things that are typically called "computers" today are capable of storing
> programs into memory *themselves*, rather than merely executing instructions
> from a read-only medium

But that has no bearing on what I said.  An 8008 executes a stored program.

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

#91548

Gareth Owen <gwowen@gmail.com> writes:

> Tim Rentsch <txr@alumni.caltech.edu> writes:
>
>> "J. Clarke" <j.clarke.873638@gmail.com> writes:
>>
>>> In article <nl966b$c57$1@dont-email.me>, nospam@nspam.invalid says...
>>>
>>>>     I was told that DMR' thesis was in "recursive functions".  Now I've
>>>> always assumed maybe being wrong that back then this may have been a
>>>> mathematics recursive functions and he was a Mathematician.  Does anyone know
>>>> if his PhD was mathematics or computer science?
>>>>
>>>> Bill
>>>
>>> There's an interview that addresses this a little bit at
>>> <https://www.princeton.edu/~hos/mike/transcripts/ritchie.htm>.
>>>
>>> Remember, though, this was Harvard, where the computer as we know it was
>>> _invented_.  They were _doing_ what we now call "computer science" long
>>> before there was a name for it.
>>
>> I thought the computer as we know it was invented at the University
>> of Pennsylvania and the group at the Institute for Advanced Studies.
>> There were several precursors, including the Harvard Mark I, but
>> they are not computers as we know them today.  In particular, IIUC,
>> the Harvard Mark I was not Turing complete.
>
> Yeah, well as a holder of a degree from the University of Manchester,
> I say YAH BOO SUCKS to all you lot.
>
> If it ain't executing a stored program,
> then it ain't a "computer as we know it".

Do you think the computers developed at UPenn and IAS did
not include a stored program functionality?  Or do you
think another location has historical precedent?  If so
please say more.  I for one would like to correct any
historical inaccuracies in what I have read and learned
on the subject.

> (See also "The Myth of the Lone Inventor")

Please note my comments are about where the machines were
developed, not about who had the underlying ideas for them.
AFAIK all of the people who were working on early forms of
computers in the 1900's drew on the work of Charles Babbage.
But I don't think of him as inventing the computer as we
know it today.

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

#91558

Tim Rentsch <txr@alumni.caltech.edu> writes:

> Gareth Owen <gwowen@gmail.com> writes: Do you think the computers
> developed at UPenn and IAS did not include a stored program
> functionality?  

Paper tape and punch cards are not a "stored program" as the term is
commonly used in computing history.  You may choose to consider them to
be so. 

https://en.wikipedia.org/wiki/Stored-program_computer 

"A stored-program computer is one that stores program instructions in
electronic memory" ... 

"The University of Manchester's Small-Scale Experimental Machine (SSEM)
is generally recognized as world's first electronic computer that ran a
stored program—an event that occurred on 21 June 1948. However the SSEM
... more a proof of concept predecessor to the Manchester Mark 1
computer, which was first put to research work in April 1949."

You may choose other definitions to fit your preferred answers.  Whatever.

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

#91565

Gareth Owen <gwowen@gmail.com> writes:

> Tim Rentsch <txr@alumni.caltech.edu> writes:
>
>> Gareth Owen <gwowen@gmail.com> writes: Do you think the computers
>> developed at UPenn and IAS did not include a stored program
>> functionality?  
>
> Paper tape and punch cards are not a "stored program" as the term is
> commonly used in computing history.  You may choose to consider them to
> be so. 
>
> https://en.wikipedia.org/wiki/Stored-program_computer 
>
> "A stored-program computer is one that stores program instructions in
> electronic memory" ... 
>
> "The University of Manchester's Small-Scale Experimental Machine (SSEM)
> is generally recognized as world's first electronic computer that ran a
> stored program—an event that occurred on 21 June 1948. However the SSEM
> ... more a proof of concept predecessor to the Manchester Mark 1
> computer, which was first put to research work in April 1949."
>
> You may choose other definitions to fit your preferred answers.  Whatever.

I'd quibble with wikipedia's requirement that a stored program computer
needs to use "electronic" memory -- the Harvard machine executed a
program stored on paper tape, but wasn't a computer in the sense that it
wasn't Turing complete.

ENIAC wasn't a stored program computer; it was programmed by plugboard.

The IAS machine (EDVAC) was a stored program computer, but wasn't
running before SSEM -- or EDSAC for that matter.  Wilkes acknowledged
that EDSAC was inspired by reading von Neumann's paper; is there a known
link from EDVAC to SSEM?

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

#91568

Joe Pfeiffer <pfeiffer@cs.nmsu.edu> writes:

> Gareth Owen <gwowen@gmail.com> writes:
>
>> Tim Rentsch <txr@alumni.caltech.edu> writes:
>>
>>> Gareth Owen <gwowen@gmail.com> writes: Do you think the computers
>>> developed at UPenn and IAS did not include a stored program
>>> functionality?  
>>
>> Paper tape and punch cards are not a "stored program" as the term is
>> commonly used in computing history.  You may choose to consider them to
>> be so. 
>>
>> https://en.wikipedia.org/wiki/Stored-program_computer 
>>
>> "A stored-program computer is one that stores program instructions in
>> electronic memory" ... 
>>
>> "The University of Manchester's Small-Scale Experimental Machine (SSEM)
>> is generally recognized as world's first electronic computer that ran a
>> stored program—an event that occurred on 21 June 1948. However the SSEM
>> ... more a proof of concept predecessor to the Manchester Mark 1
>> computer, which was first put to research work in April 1949."
>>
>> You may choose other definitions to fit your preferred answers.  Whatever.
>
> I'd quibble with wikipedia's requirement that a stored program computer
> needs to use "electronic" memory -- the Harvard machine executed a
> program stored on paper tape, but wasn't a computer in the sense that it
> wasn't Turing complete.

Presumably you mean the Harvard Mark I as there were a series of them up
to the Mark IV.  Can you say why it is wasn't Turing complete?  The
requirements for completeness are modest (putting aside the far from
modest assumption of unbounded resources) so the reason for failing
would be significant I think.

(It's hard to find enough details on the web to settle this by simple
searching -- I did try for a bit.)

<snip>
-- 
Ben.

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

#91571

Ben Bacarisse <ben.usenet@bsb.me.uk> writes:

> Joe Pfeiffer <pfeiffer@cs.nmsu.edu> writes:
>
>> Gareth Owen <gwowen@gmail.com> writes:
>>
>>> Tim Rentsch <txr@alumni.caltech.edu> writes:
>>>
>>>> Gareth Owen <gwowen@gmail.com> writes: Do you think the computers
>>>> developed at UPenn and IAS did not include a stored program
>>>> functionality?  
>>>
>>> Paper tape and punch cards are not a "stored program" as the term is
>>> commonly used in computing history.  You may choose to consider them to
>>> be so. 
>>>
>>> https://en.wikipedia.org/wiki/Stored-program_computer 
>>>
>>> "A stored-program computer is one that stores program instructions in
>>> electronic memory" ... 
>>>
>>> "The University of Manchester's Small-Scale Experimental Machine (SSEM)
>>> is generally recognized as world's first electronic computer that ran a
>>> stored program—an event that occurred on 21 June 1948. However the SSEM
>>> ... more a proof of concept predecessor to the Manchester Mark 1
>>> computer, which was first put to research work in April 1949."
>>>
>>> You may choose other definitions to fit your preferred answers.  Whatever.
>>
>> I'd quibble with wikipedia's requirement that a stored program computer
>> needs to use "electronic" memory -- the Harvard machine executed a
>> program stored on paper tape, but wasn't a computer in the sense that it
>> wasn't Turing complete.
>
> Presumably you mean the Harvard Mark I as there were a series of them up
> to the Mark IV.  Can you say why it is wasn't Turing complete?  The
> requirements for completeness are modest (putting aside the far from
> modest assumption of unbounded resources) so the reason for failing
> would be significant I think.
>
> (It's hard to find enough details on the web to settle this by simple
> searching -- I did try for a bit.)

As originally built, it couldn't do a conditional branch.  Later it was
able to conditionally switch between several paper tapes for its
instructions, but from what I've read that isn't good enough (like you,
I'd have to get more details on just how it did that to say for myself
whether it was really Turing complete, but the claims I've come across
say "no").  Hmmm, looks like Blaauw and Brooks has a chapter about it...

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

#91578

Joe Pfeiffer <pfeiffer@cs.nmsu.edu> writes:

> Ben Bacarisse <ben.usenet@bsb.me.uk> writes:
>
>> Joe Pfeiffer <pfeiffer@cs.nmsu.edu> writes:
>>
>>> Gareth Owen <gwowen@gmail.com> writes:
>>>
>>>> Tim Rentsch <txr@alumni.caltech.edu> writes:
>>>>
>>>>> Gareth Owen <gwowen@gmail.com> writes: Do you think the computers
>>>>> developed at UPenn and IAS did not include a stored program
>>>>> functionality?  
>>>>
>>>> Paper tape and punch cards are not a "stored program" as the term is
>>>> commonly used in computing history.  You may choose to consider them to
>>>> be so. 
>>>>
>>>> https://en.wikipedia.org/wiki/Stored-program_computer 
>>>>
>>>> "A stored-program computer is one that stores program instructions in
>>>> electronic memory" ... 
>>>>
>>>> "The University of Manchester's Small-Scale Experimental Machine (SSEM)
>>>> is generally recognized as world's first electronic computer that ran a
>>>> stored program—an event that occurred on 21 June 1948. However the SSEM
>>>> ... more a proof of concept predecessor to the Manchester Mark 1
>>>> computer, which was first put to research work in April 1949."
>>>>
>>>> You may choose other definitions to fit your preferred answers.  Whatever.
>>>
>>> I'd quibble with wikipedia's requirement that a stored program computer
>>> needs to use "electronic" memory -- the Harvard machine executed a
>>> program stored on paper tape, but wasn't a computer in the sense that it
>>> wasn't Turing complete.
>>
>> Presumably you mean the Harvard Mark I as there were a series of them up
>> to the Mark IV.  Can you say why it is wasn't Turing complete?  The
>> requirements for completeness are modest (putting aside the far from
>> modest assumption of unbounded resources) so the reason for failing
>> would be significant I think.
>>
>> (It's hard to find enough details on the web to settle this by simple
>> searching -- I did try for a bit.)
>
> As originally built, it couldn't do a conditional branch.  Later it was
> able to conditionally switch between several paper tapes for its
> instructions, but from what I've read that isn't good enough (like you,
> I'd have to get more details on just how it did that to say for myself
> whether it was really Turing complete, but the claims I've come across
> say "no").  Hmmm, looks like Blaauw and Brooks has a chapter about
> it...

Ah, thanks.  I did not know that.  I would rule out (from being a
candidate for the first automatic computer) anything that can't do a
conditional branch (which presumably includes what we'd call a while
loop these days).

-- 
Ben.

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

#91580

On Tuesday, August 2, 2016 at 1:37:28 PM UTC-5, Ben Bacarisse wrote:
> Ah, thanks.  I did not know that.  I would rule out (from being a
> candidate for the first automatic computer) anything that can't do a
> conditional branch (which presumably includes what we'd call a while
> loop these days).

Many old computers and control systems didn't need branches to be capable
of looping.  Simply splice the end of the tape back to the beginning, or
include a device to move cards back to the top of the hopper after they're
processed.  Various other kinds of conditional logic could be emulated
using various kinds of arithmetic (if x is known to be 0 or 1, x ? y:z
could be replaced with x*y + (1-x)*z, especially if multiplication by small
values is fast.

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

#91583

supercat@casperkitty.com writes:

> On Tuesday, August 2, 2016 at 1:37:28 PM UTC-5, Ben Bacarisse wrote:
>> Ah, thanks.  I did not know that.  I would rule out (from being a
>> candidate for the first automatic computer) anything that can't do a
>> conditional branch (which presumably includes what we'd call a while
>> loop these days).
>
> Many old computers and control systems didn't need branches to be capable
> of looping.  Simply splice the end of the tape back to the beginning,

Although, without a choice being made, that's not what a while loop
does.

> or
> include a device to move cards back to the top of the hopper after they're
> processed.

and that's not automatic.  And if you automate the move, you again need
a conditional or you don't get a while loop.

> Various other kinds of conditional logic could be emulated
> using various kinds of arithmetic (if x is known to be 0 or 1, x ? y:z
> could be replaced with x*y + (1-x)*z, especially if multiplication by small
> values is fast.

Yes, you can do a little that way but halting is still a problem.
Formally, some machine's arithmetic are probably Turing complete which
is why I did not come done either way on the completeness issue -- I
just don't know enough about the issue.  I'd rather just say that I'd
want a more direct notion of choice when considering candidates.

-- 
Ben.

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

#91603

Gareth Owen <gwowen@gmail.com> writes:

> Tim Rentsch <txr@alumni.caltech.edu> writes:
>
>> Gareth Owen <gwowen@gmail.com> writes:  Do you think the computers
>> developed at UPenn and IAS did not include a stored program
>> functionality?
>
> Paper tape and punch cards are not a "stored program" as the term is
> commonly used in computing history.  You may choose to consider them to
> be so.
>
> https://en.wikipedia.org/wiki/Stored-program_computer
>
> "A stored-program computer is one that stores program instructions in
> electronic memory" ...
>
> "The University of Manchester's Small-Scale Experimental Machine (SSEM)
> is generally recognized as world's first electronic computer that ran a
> stored program -- an event that occurred on 21 June 1948.  However the SSEM
> ... more a proof of concept predecessor to the Manchester Mark 1
> computer, which was first put to research work in April 1949."
>
> You may choose other definitions to fit your preferred answers.  Whatever.

I see.  What I was trying to articulate was something a little
different - not where building a "like now" machine was first
completed, but where such an architecture was first designed.
The idea for putting program instructions in the same memory as
data occurs even earlier, at least as far back as Konrad Zuse in
1936.  AFAIK though the first design of a contemporary-style
architecture was done by the UPenn/IAS effort, starting in 1944.

None of that takes away from the accomplishment of the Manchester
group, which deserves more recognition than it has gotten.  I
didn't know about this before - thank you for the education.

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

#95409

In article <87wpjz68re.fsf@gmail.com>, gwowen@gmail.com says...
> 
> Tim Rentsch <txr@alumni.caltech.edu> writes:
> 
> > Gareth Owen <gwowen@gmail.com> writes: Do you think the computers
> > developed at UPenn and IAS did not include a stored program
> > functionality?  
> 
> Paper tape and punch cards are not a "stored program" as the term is
> commonly used in computing history.  You may choose to consider them to
> be so. 
> 
> https://en.wikipedia.org/wiki/Stored-program_computer 
> 
> "A stored-program computer is one that stores program instructions in
> electronic memory" ... 
> 
> "The University of Manchester's Small-Scale Experimental Machine (SSEM)
> is generally recognized as world's first electronic computer that ran a
> stored program—an event that occurred on 21 June 1948. However the SSEM
> ... more a proof of concept predecessor to the Manchester Mark 1
> computer, which was first put to research work in April 1949."
> 
> You may choose other definitions to fit your preferred answers.  Whatever.

So the Analytical Engine would not have been a "stored-program 
computer" because the memory, like every other part of it, was not 
electronic?

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

#95415

On Sunday, October 30, 2016 at 8:06:30 AM UTC-5, J. Clarke wrote:
> So the Analytical Engine would not have been a "stored-program 
> computer" because the memory, like every other part of it, was not 
> electronic?

The modern stored computer adds two useful abilities which were not present
in the earliest machines: the ability to process instructions in a sequence
controlled by the program, and the ability to have a machine generate code
and then execute it.  What is important is not whether a program is stored
electronically or mechanically, but whether the machine has the indicated
abilities.

If Babbage's engine were constructed to recirculate cards, and had the
ability to skip cards under program control, those abilities would IMHO
have elevated it to the first   If it had included instructions for e.g.

   1. Skip the number of cards specified in register zero.

   2. Skip the number of cards specified in register zero, insert a new
      card punched with the contents of registers 2, 3, 4, etc. up to
      the capacity of the card, and then skip the number of cards
      specified in register one.

   3. Skip the number of cards specified in register zero, eject the next
      card, and then skip the number of cards specified in register one.

then those instructions would have been sufficient to make it a self-
programmable stored program computer [self-programming using only such
instructions would likely be hideously slow, and someone wanting a machine
to be capable of usefully generating code would probably equip it with
two or more sets of recirculating cards and instructions to move cards
between sets, but the ability to execute instructions in code-controlled
sequence as well as punch new ones would be sufficient to make something
a self-programmable stored program computer--even if it's a slow one.

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

#95456

"J. Clarke" <j.clarke.873638@gmail.com> writes:

>> You may choose other definitions to fit your preferred answers.  Whatever.
>
> So the Analytical Engine would not have been a "stored-program 
> computer" because the memory, like every other part of it, was not 
> electronic?

You may choose other definitions to fit your preferred answers.  Whatever.

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | comp.lang.c

csiph-web