Prev: Strange change of behaviour, executing daughter programs in Windows
Next: Pause in program execution
From: Richard Maine on 12 Apr 2008 12:04
Louis Krupp <lkrupp(a)pssw.nospam.com.invalid> wrote:

> Richard Maine wrote:
> <snip>
> > 2. In the strictest sense, pretty much nothing in I/O is guaranteed.
> <snip>
> > Classic examples include things like reading from a printer (that
> > actually makes sense for many of today's printers at some level, but
> > we'll ignore that) or writing to a card punch.
>
> I think you mean "card reader."

Oops. Yes. That just came out wrong.

--
Richard Maine | Good judgement comes from experience;
email: last name at domain . net | experience comes from bad judgement.
domain: summertriangle | -- Mark Twain
From: glen herrmannsfeldt on 13 Apr 2008 04:40
Louis Krupp wrote:
> Richard Maine wrote:
> <snip>

>> 2. In the strictest sense, pretty much nothing in I/O is guaranteed.

> <snip>

>> Classic examples include things like reading from a printer (that
>> actually makes sense for many of today's printers at some level, but
>> we'll ignore that) or writing to a card punch.

> I think you mean "card reader." There wasn't much one could say to them.

Then there was the IBM 3525 which could, in one pass, read a card, punch
new data on it, and print on it.

I believe if done in Fortran it would require three units.

Some DEC systems would assign both unit 5 and unit 6 to the terminal.
It was allowed to write unit 5 or 6, or read unit 5 or 6 for
terminal I/O.

-- glen

From: Kevin G. Rhoads on 15 Apr 2008 18:32
>> > we'll ignore that) or writing to a card punch.
>>
>> I think you mean "card reader."
>
>Oops. Yes. That just came out wrong.

However, for certain models there was a punch station option
for card readers. Typically, it would go
Read station
<null> or Punch Station
Read station
Similarly many punches had read stations after the punch station
for verification.

This allowed for read-after-write verification of cards and
for modification of previously punched cards. This was sometimes
used for such things as copy prevention -- I remember my 6.251
professor telling of a specialty compiler he wrote that people
bought. One company wanted just a single copy. He provided a
binary which checked the machine serial number and, in the case
of failed match, overpunched the input deck (literally the
sometimes proverbial "read and shred card" ...)

So writing to the card reader OR reading from the card punch
was not necessarily an oxymoron even 40 years ago.

Still, modern keyboards, even the clicky kind are much quieter ...
From: Kevin G. Rhoads on 15 Apr 2008 18:33
> One company wanted just a single copy.
They later bought several more copies ...
From: glen herrmannsfeldt on 15 Apr 2008 19:10
Kevin G. Rhoads wrote:
(snip)

> So writing to the card reader OR reading from the card punch
> was not necessarily an oxymoron even 40 years ago.

For the 3525 you would need two or three Fortran units
and corresponding DD cards. (read, punch, print)
Though spooled I/O makes such things harder to do.

-- glen

 | 
Pages: 1
Prev: Strange change of behaviour, executing daughter programs in Windows
Next: Pause in program execution