Card Columns (was Why did they make ... ?) 
Author Message
 Card Columns (was Why did they make ... ?)

Quote:



>>Yes, and it printed 120 characters.  Many 407's ended their lives simply
>>pre-printing fortran source to check for accuracy before compiling.

> <snip>

> How did they check?  This is a badly worded question but I
> don't know enough to write a good one.

> I suppose that there could alphanumeric checks and print
> stars around questionable cards.  Or that there were continuation
> cards in places that didn't make much sense.

All the 407 was used for was to print an 80-80 listing of the source deck.

The programmer then performed desk checking.  Nobody in his right mind submitted
a program for compilation without thoroughly desk checking it.  Especially if
you could only get one or two turnarounds a day.

Needless to say there were always some people around who weren't in their right
minds.



Sun, 10 Jul 2005 03:27:21 GMT  
 Card Columns (was Why did they make ... ?)


Quote:
> [1] Tommy Flowers, who developed much of the Collosus hardware actually
>     worked for the Post Office so would have been more familiar with punched
>     tape than cards.   Tape was also used on the FISH bombe which preceded
>     Collosus.

Errm, Tommy Flowers was STILL working for the Post Office when Colossus
was designed and built.  It was only post-war (I think, BICBW) that he
left their employ when Baby was being built.

(Oh, re-reading that, you DIDN'T say that he had *previously* worked for
the Post Office, which was my initial interpretation.  Still, I think
I'll post this anyway.)

--

    "We can no longer stand apart from Europe if we would.  Yet we are
    untrained to mix with our neighbours, or even talk to them".
                                              George Macaulay Trevelyan, 1919



Sun, 10 Jul 2005 03:45:24 GMT  
 Card Columns (was Why did they make ... ?)

Quote:



>>Many 407's ended their lives simply
>>pre-printing FORTRAN source to check for accuracy before compiling.

> <snip>

> How did they check?  This is a badly worded question but I
> don't know enough to write a good one.

Human eyeballs.

--
John W. Kennedy
"The poor have sometimes objected to being governed badly;
the rich have always objected to being governed at all."
   -- G. K. Chesterton, "The Man Who Was Thursday"



Sun, 10 Jul 2005 04:17:11 GMT  
 Card Columns (was Why did they make ... ?)

Quote:


>>Columns 73-80 were ignored because the reader literally couldn't
>>pick them up, unless you dropped in
>>another control panel that was jumpered differently.
> So you couldn't read them if you wired the board wrong? That's not
> even close to "literally couldn't".

You could read only 72 of the 80 columns.  A different board could
choose a different 72 columns, but the norm was 1-72.

--
John W. Kennedy
"The poor have sometimes objected to being governed badly;
the rich have always objected to being governed at all."
   -- G. K. Chesterton, "The Man Who Was Thursday"



Sun, 10 Jul 2005 04:17:13 GMT  
 Card Columns (was Why did they make ... ?)

Quote:


>>The 1401 mode on the 30, I believe, was actually "compatibilty"
>>rather than
>>emulation as such, as it could run entirely without a
>>360-architecture program.
> No.

As originally shipped, you needed to boot a "Compatibility
Initialization Deck", which was shipped as part of the hardware feature.
  So the appearance, at least, was that it ran with no 360 program.

Later, someone came up with a Type III program called Extra-Large
Compatibility Initialization Deck (EL-CID).

--
John W. Kennedy
"The poor have sometimes objected to being governed badly;
the rich have always objected to being governed at all."
   -- G. K. Chesterton, "The Man Who Was Thursday"



Sun, 10 Jul 2005 04:17:14 GMT  
 Card Columns (was Why did they make ... ?)

Quote:


>>>You bring back fond memories. My very first experience as a programmer
>>>was on a 16K 1401 at North American Aviation in 1961. We used Autocoder
>>>but later developed our own assembler because Autocoder took minutes to
>>>spin through the program tape looking for macros even when our source
>>>code had none.  For business data processing (decimal arithmetic and
>>>string processing) the 1401 had (and still has) the best instruction set
>>>in existence. In addition to the luxury of 16K, we also had 5 tape
>>>drives, a card-reader/punch and a 300LPM chain printer.

>>??  The normal rating of the chain-based 1403's was 600LPM.

> Only when printing blank lines.
> For printing alphanumeric stuff, it slowed down a lot.
> It couldn't keep up with the 600 card-per-minute reader.

What 600 card-per-minute reader?  The 1402 card reader was 800 cards per
minute.  (However, the basic model had a single clutchpoint, and if the
machine didn't drive it at 800, it slowed to 400.)

--
John W. Kennedy
"The poor have sometimes objected to being governed badly;
the rich have always objected to being governed at all."
   -- G. K. Chesterton, "The Man Who Was Thursday"



Sun, 10 Jul 2005 04:17:16 GMT  
 Card Columns (was Why did they make ... ?)

Quote:
> What 600 card-per-minute reader?  The 1402 card reader was 800
cards per
> minute.  (However, the basic model had a single clutchpoint,
and if the
> machine didn't drive it at 800, it slowed to 400.)

ISTR the 1442 reader was rated at 600 cpm, and there was a "reed"
printer, the 1443, that print at about 300 lpm.  The CPU in
question was called a 1440, as you might expect, and it was
related to the 1401, although I'm pretty sure it was cheaper.
Perhaps that's the machine we're talking about here.

There was a 1442-N1 reader/punch for the 360's, too. It could
read and punch cards on a single card path, first reading, and
then optionally punching. The card path through the machine had
two right-angle turns in it: one right after the hopper, then the
card fed column-by-column throuh the reader and punch units, and
then took another turn before being dropped into one of those
angled stackers like you have on the 2540.  Can you spell
"kludge?".

On the 360/20, you could get a similar machine that could read,
punch, and print cards all in one operation. It was termed a
Multi-Function Card Machine, or MFCM, affectionately known to
CE's (out of customers' earshot) as a Mother-F***ing
Card-Mulcher.



Sun, 10 Jul 2005 04:45:07 GMT  
 Card Columns (was Why did they make ... ?)

Quote:

>Probably false advertising, as didn't IBM have a 1401 emulator for the
>/360 ?

Why do you say "false advertising"?  Several of the S/360 models
had an (extra-cost) option of 1401 emulation.

Joe Morris



Sun, 10 Jul 2005 08:05:16 GMT  
 Card Columns (was Why did they make ... ?)

Quote:
>I think, but I do not know, that the IBM 407 could read all 80 columns;
>at least, the one I used could.  

...and could read the same card multiple times if necessary.  Assuming
that one could figure out the documentation for the MLR hubs <grin>

Every 407 I worked with had 120 print wheels.

Quote:
>                                The machine was programmed by a plug
>board.  In the one I had access to, the plugs for columns 73-80 were
>wired to a dead socket.  

Why?  If you have a listing board and don't want to print certain
columns, you just don't wire the second-read brushes for those
columns to the printwheels.

Quote:
>                          Column 1 was wired, in a way I never
>understood, to throw a page if it saw a '1', to overprint if it saw a
>'+', and do something unusually magic for other Fortran characters.

That's what the 407 documentation called a "program".  Detecting
a '1' in cc1 would start an overflow program (just as if the carriage
tape hole in channel 12 had been detected).  Recognizing "+" would call
for suppressing the signal to the space hub for that line.  (Slightly
different programming/wiring would be required depending on whether
it was to interpret "+" as "overprint previous line" or "overprint
next line.")

Remember that the name of the 407 is "accounting machine".  It was designed
to allow the board to be programmed to do things like recognize when a
billing document had reached the bottom of a page (typically calling for
subtotals to be printed, the paper ejected to the top of a new page,
and a continuation header printed), or when a new customer's records
show up in the input stream (print final totals, eject the page, and
start a new page header).  It even had a special final-card program
that could be used to print grand and/or hash totals.

Quote:
>Unfortunately, some vandal unplugged everything one day.

Repeat after me: "We WILL keep backups and documentation."

Joe Morris



Sun, 10 Jul 2005 08:35:36 GMT  
 Card Columns (was Why did they make ... ?)

Quote:


>> BTW, I do agree with the poster that the 36 bit machine had something
>> to do with the 72 column card limit but I seem to recall the 407 (??)
>> accounting machine only had 72 columns of print out as well and maybe
>> it impacted this limit as well.
>I agree that the 72-column limit was because of the 704's hardware
>limitations.

The 704/9/9x data channels could transfer 24, 36-bit words to or from
a unit record device in one operation; in other words, 72 columns' worth
of data.

Quote:
>The 407 manual says that it has 120, not 72, print wheels, so I don't
>think it can be blamed.  I wonder what they did with all those columns
>when the cards only held 80 characters apiece.

There wasn't a hardwired correspondence between card columns and
printwheels; everything was done on the plugboard.  See my earlier
posting from tonight in this thread about the name "accounting machine".

Joe Morris



Sun, 10 Jul 2005 08:38:31 GMT  
 Card Columns (was Why did they make ... ?)

|> I think, but I do not know, that the IBM 407 could read all 80 columns;
|> at least, the one I used could.
|
| ...and could read the same card multiple times if necessary.  Assuming
| that one could figure out the documentation for the MLR hubs <grin>

| Every 407 I worked with had 120 print wheels.

|>                                The machine was programmed by a plug
|>board.  In the one I had access to, the plugs for columns 73-80 were
|>wired to a dead socket.

| Why?  If you have a listing board and don't want to print certain
| columns, you just don't wire the second-read brushes for those
| columns to the printwheels.

|>                          Column 1 was wired, in a way I never
|>understood, to throw a page if it saw a '1', to overprint if it saw a
|>'+', and do something unusually magic for other Fortran characters.

| That's what the 407 documentation called a "program".  Detecting
| a '1' in cc1 would start an overflow program (just as if the carriage
| tape hole in channel 12 had been detected).  Recognizing "+" would call
| for suppressing the signal to the space hub for that line.  (Slightly
| different programming/wiring would be required depending on whether
| it was to interpret "+" as "overprint previous line" or "overprint
| next line.")

At the University I attended, we had a striped-down IBM 407.  Every once
in a blue moon, it would be worked on (down), so we had to go to the
administration office to use theirs.   Wow:   it had every feature except
floating point divide.  There were cables everywhere --- going to the
IBM 088 (sorter?, # could be wrong), the keypunch for punching cards,
another IBM 407 (could have been an IBM 40x printer), some other machine,
etc.   Apparently, the 407 could keep running totals in decimal and also
floating point formats.   That particular IBM 407 was almost like a
computer, with all it's bells and whistles and connected I/O gear.
_____________________________________________________________Gerard S.

| Remember that the name of the 407 is "accounting machine".  It was designed
| to allow the board to be programmed to do things like recognize when a
| billing document had reached the bottom of a page (typically calling for
| subtotals to be printed, the paper ejected to the top of a new page,
| and a continuation header printed), or when a new customer's records
| show up in the input stream (print final totals, eject the page, and
| start a new page header).  It even had a special final-card program
| that could be used to print grand and/or hash totals.

|>Unfortunately, some vandal unplugged everything one day.

| Repeat after me: "We WILL keep backups and documentation."



Sun, 10 Jul 2005 08:58:45 GMT  
 Card Columns (was Why did they make ... ?)

Quote:


>> assembly source code via another paper tape. The "T" at the
>> end of the FLIT de{*filter*} name, and the "T" in TECO both
>> originally meant tape, as in paper tape. And "DDT" was the
>> "dynamic debugging tape". IIRC.
>IIRC, the abbreviation DDT stood for "dynamic debugging technique".
>DDT *has* to have been a backronym.

Not to anyone who knows what "Flit" is.  It's more likely a "necessary"
acronym to which a set of English words had to be found.

Sadly, to most people today the phrase "Quick, Henry, the Flit!"
has absolutely no meaning.

And those words were "DEC Debugging Tape".  At this instant I'm
looking at the introductory memo dated 15 February 1962, which
begins:

   DDT (Dec [capitalization sic] Debugging Tape) is a symbolic debugging
   program for the PDP-1.  It occupies upper memory starting at register
   6000 [...] down to register 5543.

For "register" read "memory address" for 18-bit words.

Unlike most of the memos from the MIT PDP-1 group, it has nobody's
name as either author or approver.

Joe Morris



Sun, 10 Jul 2005 09:08:36 GMT  
 
 [ 373 post ]  Go to page: [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]

 Relevant Pages 
 

 
Powered by phpBB® Forum Software