14 Aug
2017
14 Aug
'17
6:02 a.m.
On Sun, 13 Aug 2017, Dave Horsfall wrote:
On Sat, 12 Aug 2017, Steve Johnson wrote:
It turns out the original FORTRAN (manual published in October 1956; code first shipped
around April 1957) included the arithmetic IF as well as the assigned and computed GOTO
statements — see chapter 4 of this manual:
J.W. Backus, R.J. Beeber, S. Best, R. Goldberg, H.L. Herrick, R.A. Hughes, L.B. Mitchell,
R.A. Nelson, R. Nutt, D. Sayre, P.B. Sheridan, H. Stern, I. Ziller. The FORTRAN Automatic
Coding System for the IBM 704 EDPM : Programmer's Reference Manual. Applied Science
Division and Programming Research Department, International Business Machines Corporation,
October 15, 1956, 51 pages.
http://www.bitsavers.org/pdf/ibm/704/704_FortranProgRefMan_Oct56.pdf
(For more on the original FORTRAN compiler, see
http://www.softwarepreservation.org/projects/FORTRAN/.)
A little Googling shows that the IF I mentioned
was called the
"arithmetic IF".
Ah yes. It was in FORTRAN II, as I recall. 
14 Aug
14 Aug
10:09 p.m.
New subject: origin of string.h and ctype.h
On 8/13/17, Paul McJones <paul(a)mcjones.org> wrote:
On Sun, 13 Aug 2017, Dave Horsfall wrote:
> On Sat, 12 Aug 2017, Steve Johnson wrote:
>
>> A little Googling shows that the IF I mentioned was called the
>> "arithmetic IF".
>
> Ah yes. It was in FORTRAN II, as I recall.
The FORTRAN II arithmetic IF, with its three-way branch, was probably
introduced to facilitate use of the conditional branch instructions of
the IBM 704, which test the value of the accumulator and take a branch
accordingly:
TMI (transfer on minus)
TMP (transfer on plus)
TZE (transfer on zero)
TNZ (transfer on not zero)
It takes at most two of these instructions to implement the arithmetic IF.
-Paul W.
It turns out the original FORTRAN (manual published in October 1956; code
first shipped around April 1957) included the arithmetic IF as well as the
assigned and computed GOTO statements — see chapter 4 of this manual:
J.W. Backus, R.J. Beeber, S. Best, R. Goldberg, H.L. Herrick, R.A. Hughes,
L.B. Mitchell, R.A. Nelson, R. Nutt, D. Sayre, P.B. Sheridan, H. Stern, I.
Ziller. The FORTRAN Automatic Coding System for the IBM 704 EDPM :
Programmer's Reference Manual. Applied Science Division and Programming
Research Department, International Business Machines Corporation, October
15, 1956, 51 pages.
http://www.bitsavers.org/pdf/ibm/704/704_FortranProgRefMan_Oct56.pdf
(For more on the original FORTRAN compiler, see
http://www.softwarepreservation.org/projects/FORTRAN/.)
11:16 p.m.
New subject: origin of string.h and ctype.h
On Aug 14, 2017, at 9:09 AM, Paul Winalski
<paul.winalski(a)gmail.com> wrote:
And the 709 added CAS (compare accumulator with storage): it skips 0, 1, or 2 instructions
depending on whether the contents of the accumulator are algebraically greater, equal, or
less than the contents of the referenced memory location. (See page 49 of
http://www.bitsavers.org/pdf/ibm/709/709_RefMan_Fortran.pdf.)
> ...
The FORTRAN II arithmetic IF, with its three-way branch, was probably
introduced to facilitate use of the conditional branch instructions of
the IBM 704, which test the value of the accumulator and take a branch
accordingly:
TMI (transfer on minus)
TMP (transfer on plus)
TZE (transfer on zero)
TNZ (transfer on not zero)
It takes at most two of these instructions to implement the arithmetic IF. 
15 Aug
15 Aug
9:57 a.m.
New subject: origin of string.h and ctype.h
On Monday, 14 August 2017 at 10:16:57 -0700, Paul McJones wrote:
It was my understanding that the 704 had this instruction too, and
that it was almost certainly the background for arithmetic IF.
Unfortunately, no reference I can find can confirm or deny this
supposition. Another one is http://www.quadibloc.com/comp/cp0309.htm,
which compares the instruction sets, but it's not categorical enough
for my liking. But if I interpret it correctly, CAS was implemented
on the 704.
Do you see a proof of the contrary?
In passing, my document also mentions the Honeywell 516, better known
to Unix people as IMP. It, too, had a CAS-like instruction.
Greg
--
Sent from my desktop computer.
Finger grog(a)lemis.com for PGP public key.
See complete headers for address and phone numbers.
This message is digitally signed. If your Microsoft mail program
reports problems, please read http://lemis.com/broken-MUA
> On Aug 14, 2017, at 9:09 AM, Paul Winalski
<paul.winalski(a)gmail.com> wrote:
>
>>> ...
>
> The FORTRAN II arithmetic IF, with its three-way branch, was probably
> introduced to facilitate use of the conditional branch instructions of
> the IBM 704, which test the value of the accumulator and take a branch
> accordingly:
>
> TMI (transfer on minus)
> TMP (transfer on plus)
My reference says TPL.
TZE
(transfer on zero)
TNZ (transfer on not zero)
It takes at most two of these instructions to implement the arithmetic IF.
And the 709 added CAS (compare accumulator with storage): it skips
0, 1, or 2 instructions depending on whether the contents of the
accumulator are algebraically greater, equal, or less than the
contents of the referenced memory location.
10:13 a.m.
New subject: origin of string.h and ctype.h
On Aug 14, 2017, at 8:57 PM, Greg 'groggy' Lehey <grog(a)lemis.com> wrote:
It was my understanding that the 704 had this
instruction [CAS] too, and
that it was almost certainly the background for arithmetic IF.
Unfortunately, no reference I can find can confirm or deny this
supposition. Another one is http://www.quadibloc.com/comp/cp0309.htm,
which compares the instruction sets, but it's not categorical enough
for my liking. But if I interpret it correctly, CAS was implemented
on the 704.
Do you see a proof of the contrary?
It turns out CAS is used in many places in the source code of the 4K and 8K drum versions
of the final IBM 704 FORTRAN II compiler. See
http://www.softwarepreservation.org/projects/FORTRAN/index.html#Source_code; in
particular:
Transcribed source:
http://www.softwarepreservation.org/projects/FORTRAN/source/fortran-ii/fort…
Assembly listing:
http://www.softwarepreservation.org/projects/FORTRAN/source/fortran-ii/fort…
I didn’t try to figure out if this code generates CAS, but it certainly uses it.
2686
days inactive
2687
days old
5 comments
4 participants
participants (4)
-
Dave Horsfall -
Greg 'groggy' Lehey -
Paul McJones -
Paul Winalski