Explanation needed of binary operators

M

Martin Gregorie

Roedy said:
I presume you were an "operator" at some point in your career and had
a faulty mylar tape loop that controlled the vertical tab stops on the
printer, causing the paper to slew endlessly at full rate. If it
happened when the covers were up you had an great arc in the air. If
closed, it packed the printer cover tight as a mummy case. To stop it
you stomped your foot on the input paper box to break the paper.
We were a small service bureau with a 1903S to keep busy. Among the
systems staff we did everything - analyzed, designed, coded and, when
necessary, operated too. I was never good enough to know what George 3
wanted by listening to the control teletype, but I could tell "LP 3 FIX"
when I was lining up paper from requests to, e.g. load a magnetic tape.
I knew operators who could drive the system entirely off sound for an
hour or so when the teletype's print head failed.

I don't remember our fast printer ever turning into a paper fountain -
or the paper tape loop breaking, but we did tend to use tougher material
than plain paper tape for production loops. I seem to remember that the
1900 printer would only throw about 3 feet of paper (i.e. about two
pages) before timing out and stopping. I know for sure that I never
broke the feed paper to stop the printer.

The 2900 printers were a nice improvement: they used a software
implementation of the paper loop and as well as telling the spooler what
sort of paper the job needed, you also told it what control loop to load.
 
M

Martin Gregorie

Roedy said:
I don't follow. EBCDC '0' is not binary 0. Further , IIRC, the
letters A-Z and a-z are not contiguous blocks of binary assignments.
>
I think the approach is fairly clear: you adjust the binary code values
so that sorting on ascending code value gives you the collation sequence
you want. In the case of EBCDIC that's pretty weird because the gaps
between I and J and between R and S are not empty: they contain a wild
assortment of punctuation and other symbols.

John says that the collation sequence predates EBCDIC. I'll go further
and guess that it predates computers as well. It was most likely defined
by IBM's original card sorters: businesses were running card-based
accounting systems in the '30s if not earlier using a room full of
sorters, collators and other electro-mechanical monsters.

FWIW the Manhattan Project calculations for the plutonium bomb design
were run using IBM card handling kit under the direction of Richard
Feynman. It was a faster replacement for the armies of girls with
hand-cranked Monroe calculators who had been doing the job. IIRC Feynman
thought up the idea of using punched cards.
 
J

John W. Kennedy

Martin said:
I think the approach is fairly clear: you adjust the binary code values
so that sorting on ascending code value gives you the collation sequence
you want. In the case of EBCDIC that's pretty weird because the gaps
between I and J and between R and S are not empty: they contain a wild
assortment of punctuation and other symbols.

They do /now/, in EBCDIC-version-of-ISO-8859-1 and the like. But all
those spaces were empty in 1964. Apart from control characters and
lower-case letters, the original EBCDIC had only about 64 characters. So
the 64 characters of traditional BCD collated in EBCDIC more or less as
they always had, but with a straight binary compare, instead of special
collating-sequence hardware.

(That special hardware is why the basic-model 1401 Compare instruction
could only compare equal/not-equal; numerics could be high/low/equal
compared with a subtraction, but if you wanted to high/low/equal compare
alphameric data, you had to both buy a hardware add-on and accept
slighly reduced CPU performance.)

Mainframes are slowly moving away from EBCDIC, of course. The newest
System Z machines include full support of Unicode, including opcodes to
translate among UTF-8, UTF-16, and UTF-32.
John says that the collation sequence predates EBCDIC. I'll go further
and guess that it predates computers as well. It was most likely defined
by IBM's original card sorters: businesses were running card-based
accounting systems in the '30s if not earlier using a room full of
sorters, collators and other electro-mechanical monsters.

Pretty much, yes.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

Forum statistics

Threads
473,744
Messages
2,569,484
Members
44,903
Latest member
orderPeak8CBDGummies

Latest Threads

Top