I was back in the oversize area the other day looking at old Sallyports and just for fun I started opening up random envelopes. Look what I found:
They’re the plans for the Math Science Building (Herman Brown Hall) that we’ve recently been watching go up as we simultaneously watch the parking lot disappear. The drawings of the basement are by far the most interesting ones–the other floors are pretty standard layouts of offices and lounges. I think I get the gist of what’s what here from previous discussions.
I’m not sure I understand, though, exactly what is meant by “Keypunch.”
The most revealing thing in this booklet, though, and the thing that makes me bother writing about it at all, are some handwritten notations on a map of campus.
That’s President Pitzer’s handwriting, by the way, and I really wish I could ask him what was going through his mind here. I suspect that he was casting about for possible locations other than this odd site. Note also what he’s got written down where George R. Brown Hall is today: “save for Chem. repl. present bldg.”
Bonus: I’ve seen this and it’s good.
So, when computers had punch cards, they had machines that you would type your cards with, called “keypunch” machines. You would sit and type in your code onto the cards, and when you were done, you would take your cards to the window and give them to the operator. He would run your cards through the computer and leave a printout in a cubby whenever he would get around to it. You would then retrieve the printout, replace cards with bad code or add new cards with the keypunch machine, and repeat.
The keypunch room was where those machines were. When I took Fortran at Rice Summer School for High School Students between 7th and 8th grade in summer of 1977, I used those keypunch machines. By the time I went to college in 1983, the punch cards were gone, and the Shepherd School had moved into the basement.
So there were a number of these machines–like typewriters?–sitting on tables or desks?
It didn’t sit on the desk; it was a desk! There was a bin to put in blank cards, a feeder which lined up a new card for you to type with, a bin for good cards, a bin for rejects, and another for the chads which were punched out. There was enough space for you to put a notebook or a printout so you could type in what you had hand-written. When you typed, you could accept a card, or reject a card.
This wikipedia article has much more detail: http://en.wikipedia.org/wiki/Keypunch
Ah! I see! I had this envisioned completely wrong. Thanks so much.
I wonder (idly) what happened to them.
Here’s a funny tidbit for you…. On more than one occasion, someone who had been working long and hard at the often frustrating job of getting a program to run properly would vent by hurling the deck of cards from a height — in my case, I remember someone throwing their deck off the Brown rooftop (which at that time was fully available to students for sunbathing, studying, whatever — accessed by the stairs from the 8th floor.) That meant some 100-200 or so cards, punched and already successfully run with data properly transferred into the system, floated on down to the grass below. A mess to pick up, but she had friends who helped….. It was reportedly cathartic — and no, it wasn’t me, but I remember helping clean up.
They were rentals and returned at the end of lease as they were no longer needed.
When I was there, starting in 1975, the room had IBM 029 keypunch machines. You might want to look carefully at the keyboard layout, your current keyboard probably does not have “DUP” and “LEFT ZERO” keys.
http://www.columbia.edu/cu/computinghistory/029.html
The 029 punched the EBCDIC character set, an extended version of BCD. It replaced the BCD-only 026 punch. My roommate, Bob Puckette, somehow obtained an 026 keypunch and modified the relay logic so that some athletic multi-key combinations could punch the EBCDIC characters required for his PL/I programming course. Later, we figured out that PL/I supported BCD equivalents for the extended characters.
When we moved out of the apartment, we hauled the 026 downstairs and got it into the dumpster. Now that I’ve done it, I do not recommend trying to lift an IBM 026 Card Punch over your head.
Keypunch has to refer to the method of computing that was in place in the late seventies and early eighties, when computer programs were run based on cards which had holes punched into them in order to be “read” by the computer. The cards were horizontal rectangles, and one was “punched” for each line of code, essentially. The holes were created by a kepunch machine which had a keyboard, and what the individual typed was translated into punching holes in a pre-designated space on the card, which was then “read” by the computer. Can you tell I only took one computer class, hence my very rudimentary explanation? I’m sure computer science folks will be way clearer, but this is my interpretation.
I’ve heard the film is great, by the way — still waiting to see it! Thanks for the daily blog that always has interesting tidbits!
Thank you, Gloria. “Tidbits” is my middle name!
Keypunch cards left a significant amount of debris in the form of tiny paper rectangles—comparable to the infamous “chad” from voting machines of years gone by. At a friend’s wedding in 1976, all the EEs threw punchcard chad at the bride and groom rather than rice. I drove the “getaway” car for their wedding and was still vacuuming chad from the back seat of the car a year after the wedding!
Those tiny paper rectangles were called chad, and the chad of voting machine fame would have been nearly identical. The difference is that a keypunch used a vertical array of rectangular metal rods to punch one or more holes into unpunched card stock, while the voting machines used semi-perforated card stock through which voters manually poke handheld metal rods to remove the chad. If done incorrectly, the chad is incompletely removed, with one or more of the connecting paper bridges not severed, resulting in “handing chad” or even “dimpled chad” if the voter applies even less pressure.
These semi-perforated cards were also used for situations when a keypunch machine was unavailable or inappropriate. Some programmers used them for quick patches. Some data collectors used them. In both cases, the card was placed over a metal plate with holes in the appropriate locations (a 12×80 array). Some such plates came with sliding blocks through which metal rods could be inserted to punch out the pre-perfed chads; some were used freehand. One brand was Porta-Punch, as I recall. At least some of the card voting used virtually identical equipment.
The machines to read these cards worked or two different principles. Some passed the cards between a metal plate and metal brushes, detecting holes when the brushes made electrical contact with the plate. Others detected the holes optically. Both methods worked fine for standard cards punched with keypunch machines, but badly manufactured pre-perfed cards tended to shed extra chad when passed through the brush arrays, adding unintended holes. One residential college had a bit of a fiasco when cards intended for computer date matching lost many, many chad (or chads, if you prefer).
By the way, we were warned not to throw chad(s) as confetti. Seems the corners could be very sharp and damaging to corneas.
I agree that it was not a good idea to throw chad around as confetti—we did a lot of stupid things back then… 🙂 Nonetheless, it was a staple at a lot of weddings involving Rice engineers.
One additional comment on “still vacuuming chad from the back seat of the car a year after the wedding”.
I heard that some enterprising and vindictive students in Hanzen in the early 70s pulled a “jack” on someone they did not particularly care for by taking a big bag of chad, using some kind of industrial vacuum on reverse to force the chad all over the room including all sorts of hard to reach places in the victim’s room in the college, and finally using squirt bottles to get the chad damp to make it even harder to sweep or vacuum up.
The use of punched/Hollerith/IBM cards goes way back to the dawn of automation: mechanized textile looms. Wikipedia has a nice rundown of the history(http://en.wikipedia.org/wiki/Punched_card). A colleague of mine at IBM collected the cylindrical plates used to print specially marked 80-column punched cards for customers. Because IBM already had a card-printing plant in San Jose, they located their West Coast research lab there in 1952. (Most scientists graduating from the increasingly influential California universities in the economically prosperous post-WWII years would not relocate back East.) The lab’s first innovationwas the magnetic hard-disk drive, a technology which, unlike punched cards, is still improving.
The sound of a 1960s-vintage card-reader machine shredding and jamming on humidity-swelled punched cards is one that still haunts me.
Yeah, I’m kind of starting to get it. It seems odd that so much should hinge on what’s really a fairly crude mechanical process.
IT was used because it worked. Keypunch large data sets was a very easy way to do entry. The research designed a form to record the data and turned the completed forms over trained data entry operators who would initially turn the sheets into pucnhed cards. The research would then take the data deck and by knowing the format would input it into the statistical package. In the later years many data entry had moved from card to key-to-disk but the concept was the same. When I started a BCM we had a very large data entry department that did both cards and key-to-tape. all of the accounting information, payrol, HR, clinical billing and registration had been reduced to forms our operators could quickly enter. Their work product was then turned over to the requester or turned over to operations so it could be input into the appropriate computer system. This is how it had been done in the unit records day and that carried forward into the first computer wave.
My first few years at BCM were involved in writing administrative programs to allow end users to put data directly in but the data entry operators existed until I think the mid 90s
I just uncovered the card stacks that I have with the membership of S.L.O.B. entered, along with a bubble-sort program written in PL/1 to print out a list. I could show you those.
I donated my last case of fresh punch cards to the chair of Comp-Sci a few years ago, along with the big, red sign that hung in the keypunch room, instructing users on how to store their card stacks.
But… that drawing does not show the location of the keypunch room that I recall.
It was in the WNW corner…
The room labeled “Data Preparation” was where the keypunching was done. The room labeled “Keypunch” was, if I remember correctly, used for student staff, and may have had a staff keypunch in there. I think the configuration had changed by the time I first remember seeing it in 1970, because instead of opening into the machine room, I remember it opening into the “Lobby” space (behind a counter), with a “Don’t Feed the Animals” sign on the door.
As an architect, I am disturbed by the layout of the emergency exits (fire stairs) in the basement floor plan. This building uses a pair of fire stairs that are interlocked; so-called scissor-stairs. This design compromises the intent of the building code to separate the two required means of egress (i.e. the fire stairs) as much as possible. The interlocking stairs were probably in compliance with applicable codes when originally constructed, but this layout would not be allowed today.
Why would they have done it?
I hate those stairs, by the way. One of the grimmest places on campus.
I’m puzzled by the design. Scissor stairs are typically used in square-shaped floors where space is at a premium and the stairs can be located in the center of the floor. I don’t know why this building has such a layout.
There was discussion of adding an exterior exit stair from the west end of the basement when the machine room space was converted to an orchestra practice area after the computer center moved out. I didn’t pay attention to the outcome of that discussion.
As far as I know, the room labeled “Analog Computer” was never used for that, but became the Terminal Room.
There was an analog computer in Abercrombie while I was there (1975-81). Maybe they were planning to move or replace it and changed their mind.
Burrus came to Rice in 1965 and started working on digital signal processing, so the analog computer may have become less interesting to the faculty after these 1966 plans. I guess it could have been used for control systems work for a while.
Just for your info (and I spent quite a bit of time punching cards myself), the cards themselves are about the size of a typical time card.
And who here remembers playing the Star Trek game on the computer, punch cards and all.
I don’t think the size similarity between punch cards and time cards is at all coincidental. Many time clocks once actually punched the timestamp into the cards for computer time accounting. Later models, of course, transmitted the data electronically.
While the shape of the spaces is mainly accurate, the use to which they were put by the end of the spring semester of 1970 suggest that the drawing does not represent the “as built” configuration. Below are some recollections of the space.
By Fall 1970 “Analog Computer” had become home to about 12 ASR-33 Teletype machines which were used every evening M-F and Saturday morning by the introductory computing course, ENGI 240, when it was moved to the Burroughs B-5500 mainframe. (Prior to that, the course used the IBM 1620 in Ryon entirely with punch cards.) To my knowledge, no analog computer installation ever occupied that room.
“Data Preparation” was in use as the user keypunch room when I first went down into the basement during the bomb scares of late April / early May 1970 with another undergraduate who had a job keypunching for Dr. Pat Haug of the Chemistry Dept. The wall between that room and “Machine Room” was lined with “cubbyholes” for storing boxes of punched cards. The cubbyholes appeared to be more than 75% full of boxes indicating that they probably had been accumulating for a while. Users assembled punched cards into decks and then carried them into the room labelled “Lobby 225 SF” where they were placed into holding trays which the computer operators took into “Machine Room” and loaded into the card reader to run the users’ jobs. “Data Preparation” appears in another photo in this blog about a year ago, although I believe that photo was taken after the IBM 370/155 replaced the Burroughs B5500 in 1972. A few months after the IBM was installed, a small card reader was installed from all three doors in “Data Preparation” to permit users to read their card decks into the computer without operators handling the decks.
In the summer of 1982, “Data Preparation” was physically split by a wall erected between the two doorways that did not go into “Machine Room”. The area nearer to “Machine Room” was turned into a small computer room to hold two DEC VAX-11/750 computers running Berkeley UNIX to support undergraduate Computer Science education. The other section of the room housed a reduced number of keypunches, demand for which had decreased due to a greatly increased number of online terminals and dial-in ports.
Keypunches were not designed to be physically accommodating. If your legs were long or your body large you might have difficulty fitting fitting under the tabletop that the keyboard sat on. I still recall an undergraduate-before-he-became-a-Professor with a full cast on one leg trying to find a position to type on the keyboard with his leg sticking out perpendicular to the keypunch. After several painful, unproductive sessions, he finally was able to convince a female friend to punch his program for him. (Her very long fingernails, typing skills, and colorful vocabulary are an amusing anecdote that will have to be lost to history!)
The holding trays for user card decks sat on a dark blue Formica-covered counter that sat along where the wall separating “Lobby” from “Keypunch” is drawn. (I don’t believe that wall ever existed.) If the room labelled “Keypunch” was ever used for that purpose (other than one keypunch used by operators to make simple corrections), it was before May 1970. This area was reserved for the computer operators to handle the collection of card decks, deliver printouts, and talk to users who were on the “Lobby” side of the blue counter. “Keypunch” actually extended all the way to the south exterior wall, cutting off the part of “Stg.” that projected toward “Mech”. There was a doorway into “Stg.” near the exterior wall. The blue counter would later be severally cut-down during a later remodel of the area that installed a doorway between “Lobby” and “Keypunch” and reconfigured the way people, card decks, and printouts moved in and out of the secure area behind the counter.
To my knowledge, the room labelled “Mech” was never used for that purpose. I believe that by May 1970, there were a limited number of bins, cut into the wall between “Mech” and “Lobby”, into which the operators placed printouts generated by the computer and returned card decks. The bins were used by both computer center staff and users at that time. By the mid-1970s, the size of the bins had been narrowed, and the number of rows and columns increased. Computer center staff bins had been constructed on the exterior wall of “Mech” so that staff output was secured from users. (This kept enterprising undergraduates from fixing and resubmitting staff jobs — yes, another story.) To my knowledge, the wall of “Mech” with the doorway never existed.
The room labelled “Stg.” was sometimes used for that purpose. In 1972-73, the part of “Stg.” along the exterior wall housed three undergraduate student programmers. By mid-1970s, the area closest to the door into “Keypunch” became the Operations Manager’s office with parts of the area still housing supplies storage.
The room labelled “Mfr. Rep.” housed the mainframe vendor-du-jour onsite employees and their books and tools. The electrical panels for “Machine Room” were located in the wall between the computer room and “Mfr. Rep.” inside “Mfr. Rep.” (This was a frequent source of friction between the electricians, vendor’s people, and the fire marshal.) The cavity in that wall connected to the plenum below the raised floor in “Machine Room”. The installation outgrew the electrical panels in the late 1970’s and more modern power distribution units were installed on the floor in “Machine Room” to augment the available power.
The doors from “Machine Room” into “Stg.” and “Mfr. Rep.” are seen behind the two workers in the left background of the picture of “Machine Room” raised floor being installed in “Math. Science Construction, Part II Plus Parking Woe”. That photo appears to have been taken from the vicinity of the window near the door between “Machine Room” and “Data Preparation”.
The space was vacated by the computer center in early August 1983 when the operation moved to the newly completed Mudd Lab (which became the box directly above the Math Science building in the site plan bearing President Pitzer’s handwriting). Much (all?) of the basement was converted to use by the Shepherd School.
It was all Shepherd School. That coincided with an expansion of the Shepherd School during my student days.There were Wenger soundproof modules for four or five faculty teaching studios and about fifteen very small practice rooms. The large room with the sunken floor was actually never used as an orchestra rehearsal room; it was too small. It was used for violin orchestra repertoire class and later for the first opera rehearsal and storage area. Melissa and I will, at some point, go over the different locations of the Shepherd School.
Did they create a new exit stair in the space at west end of the basement? The low level of occupancy of the computer room apparently didn’t require one when the building was built. However, the music school’s occupancy level was much higher and could have forced the issue.
Did that “low level of occupancy” include all the students sitting on the floor in the halls waiting for their run to finish or for a turn at the keypunches? 🙂
I’m pretty sure I have photos of that somewhere.
The students in the halls counted toward occupancy of the halls, not the computer room. An exit stair from the computer room wouldn’t have helped them if they (nominally) couldn’t get into the computer room to use it.
And the accessible spaces were within code limits even when Mudd was designed in 1982. The problem areas were the rear of “Machine Room” and the areas from which one could exit only by going through “Machine Room”.
Of course, you can be certain that the fire marshal’s office would have had a whole herd of cows if they had ever seen the mobs of students sitting in the hallways before a major project deadline!
Well, yeah. Or on “Oh Come, Oh Come, Oh Manual” days.
Har! Har! Har!
Not much to add to the keypunch discussion, except that if you had punchcards, you needed a sorting machine, similar to this: http://www.columbia.edu/cu/computinghistory/082.jpg . Some of my earliest memories are running machines like that (and bursters) for my dad when I was less than ten years old.
When I started down their (took Elec 220 in my sophmore year, spring of ’73) the keypunch room had both a card sorter and a gang punch (a survivor of the unit record days). By wiring the control panel correctly you could feed in cards and punch new cards based on the instructions of your patch panel such as punch what is in column 1 in column 72 or even a bit of programming logic such as if column 1 is punched, punch something in column 50). Keypunch data entry was widely used and a good data entry operator using a modern key punch could turn a a phenominal number of cards. You also had verification keypunchs where a 2nd operator would verify what the first operator did to prevent errors. Keypunch/data entry continued well into the 80s
Also, in this era there were no online storage (at least for students) so all of the program were done on cards. The introduction courses might have a deck as large as 1/3 to 1/2 of a box. I think the card decks we did for Ken Kennady’s EE/MATHSCI 320 (an emulator) got up to around a box and the compiler course (ELEC/MATHSCI 423) would have up to 2-3 boxes for a parser. One learned very quickly to punch sequence number in 72-80 for your programs (most of the compilers ignored those columns on input) so that you could sort the deck if it got dropped
“Data Preparation” also had a sorter until the space was divided in 1982. I think that was a donated piece of equipment and was junked when it was removed.
There used to be an interpreter in “Data Preparation” as well, that would print the “interpretation” of a punched card. This was handy for me and my roommates as we had an old IBM keypunch that had no printer, and I had (still do) terrible typing, so I would interpret the cards and check them before submitting them.
When the hand on card reader got installed in that room, there began “unattended batch service” (I think it was nicknamed “Night Owls” or something) and we could submit jobs as long as nothing crashed. I remember being in line behind a few other folks and a freshman at the beginning of the line lifted the weight on the stack of cards being read and dropped a deck in the hopper with the rubber band still around it. Two or three of us jostled with the freshman trying to pull the deck out before the rubber band went into the reader, but we failed, thus ended batch service for the night.
NIGHT OWLS it was. I had forgotten that.
For the obsolete equipment geeks out there, the non printing keypunch that lived in our apartment (mentioned in my post on the interpreter) was an IBM 024 from the late 40’s painted a glorious battleship grey.
“Night Owls” began life as “Night Owls Time Sharing” in 1972 because timesharing terminals were the only way to access the system until the card reader was added in “Data Preparation”.
Night Owls was the 65 hours a week for which no computer operator was scheduled. The system ran “unattended” which meant if something went down it was down until the next operator shift started unless a staff member happened by and fixed the problem. Operators were scheduled 19 hours a day M-F (0700-0200 following day) and 8 hours on Saturday. Operator hours were expanded slightly in the years after the move to Mudd, but operators were not scheduled 7/24 until 1987.
It was Rice policy to charge back mainframe services to departments, grants, and external users well into the 1990’s. The popularity of Night Owls in the early years was due in part to the reduction or elimination of certain charges during Night Owls to steer demand from heavily used times do day into this period of time. Differential pricing remained in placed even after operators were scheduled 7/24.
On a different tangent, re the map, I miss the large lawn in front of Hamman. That was always such a great spot to hang out on a nice day.
I don’t think an additional stair was added when the Shepherd School moved in. If so, it would still be there today. My memory of the occupancy was there was never all that many people down there. Orchestra rep or opera class would have been fifteen or so, plus a handful more practicing or having lessons. It never, ever seemed crowded, certainly not to the level described above.