Timeline for Why were programs entered on punch cards instead of paper tapes?
Current License: CC BY-SA 4.0
23 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 31, 2022 at 14:54 | comment | added | supercat | @davidbak: Many of the early micros I've seen at the Vintage Computer Festival were hooked up to paper-tape equipment, and I would think that would have been a somewhat common means of I/O for many of them. As for performance, I would think a machine powerful enough to run Space War in real time would be table to accept thousands of bytes per second continuously from a tape while loading a tape, even if putting the drive into "maximum speed" mode would mean that it couldn't be stopped quickly. | |
| Oct 29, 2022 at 21:06 | comment | added | rcgldr | @supercat - the 600 cps photo reader I saw was used on an HP 2100 mini computer. There was no "smooth" tape motion. A pinch roller ran at a near constant 60 inches per second, and was probably slowed down only somewhat when first engaging due to the small amount of angular inertia or a roll of paper tape. The brake was nearly instant, stopping within 1/10 to 1/5 of an inch. The software to transfer data from paper tape to a file on a hard disk could keep the tape streaming. Some user made apps did not keep the tape streaming so you could hear the tape stopping intermittently. | |
| Oct 29, 2022 at 18:42 | comment | added | davidbak | @supercat - curious why you keep bringing up 6800/8080 devices as controllers for these things. I was using paper tape drives on PDP-8s and surely those were slower - much slower? (Can't find an on-topic web page right now.) And of course paper tape devices predated the -8 ... These electro-mechanical things were using discrete transistors at best. | |
| Oct 29, 2022 at 18:17 | comment | added | supercat |
...an essentially arbitrary number of zero bytes, followed by opcodes for LD H,80h / JP (HL) / LD SP,16 / OUT xx,xx / LD BC,____ / PUSH BC / LD BC,____ / PUSH BC / ... / JMP 0 could be fed directly into system without requring any kind of bootloader in RAM or ROM. What am I missing in thinking that would be a really cheap and simple, but fast, way of loading code into a system?
|
|
| Oct 29, 2022 at 18:08 | comment | added | supercat | @rcgldr: Am I amiss in thinking that the amount of circuitry needed to stall a 6800 or 8080 until each byte is received from the tape would be less than the amount of circuitry necessary to have a reader convert a tape into a sequence of bytes sent via "serial port", while simultaneously being more convenient to work with for many purposes? If an 8080 with 32K or less of RAM were wired so that all memory reads where A15 is set, or that are performed before the first OUT instruction triggers a flip flop, will wait for and use a byte from tape, I would think a tape that started with... | |
| Oct 29, 2022 at 17:55 | comment | added | supercat | ...had to be able to start and stop smoothly enough, with generous "coasting" limits, not to break the tape. If one is e.g. trying to use an assembler or compiler that is processing code off the tape as it is being read, the tape speed would need to be kept much slower. If, however, one is simply loading pre-assembled programs off tape--a task which would in many cases probably account for the majority of bytes read--a "turbo mode" which could simply feed data as fast as the operator turned the crank would seem very useful. | |
| Oct 29, 2022 at 17:46 | comment | added | supercat | ...as forcing a stall until a debounced falling edge was received on the indexing sensor, and then placing the eight data bits onto the bus, I would think such a system could boot directly off a suitably-punched tape which was either fed by hand, or loaded by hand, attached to a winder, and cranked any any speed within a substantial and adjustable range (adding a second indexing sensor could allow for timing-independent debouncing). Further, as a more general note, a tape reader that requires handshaking at the per-byte level would need to be much more complicated than one which merely... | |
| Oct 29, 2022 at 17:42 | comment | added | supercat | @rcgldr: Many tasks would have required the ability to start and stop the tape smoothly. For many other tasks, however, even a relatively slow computer would have been able to easily keep up up with an "unregulated" data rate of 40,000 bytes/second or more, especially if the tape included padding at some places. If one had a "paper tape reader" that consisted of nothing but some guides, a linear-filament lamp, and nine photocells with suitable apertures over them, and hardware that would treat any read within a certain address range... | |
| Oct 29, 2022 at 8:52 | comment | added | rcgldr | @supercat - stopping the paper tape if the host didn't keep up was an issue. The tape reader had a 1 or 2 character buffer. A pinch roller was used to move the tape, and a round disc used as a brake to stop the tape. Since the sprocket holes provided timing actual tape speed wasn't an issue. Going from 5 feet / sec to zero in 1 or 2 tenths of an inch was probably a limit of what the paper tape could handle. The brake could be turned on and off so fast that with folded tape at the brake but not at the pinch roller, music could be played with the brake (probably wasn't good for it). | |
| Oct 28, 2022 at 19:43 | comment | added | davidbak | @supercat - wikipedia says you got 10 char/inch. So 600 char/sec is already 60 inch/sec. Maybe the limit was paper handling speed? Contra alphazero's comment (first, above) most paper tape wasn't any kind of special durable type. It was actually pretty flimsy, remember? | |
| Oct 28, 2022 at 18:19 | comment | added | supercat | @rcgldr: I wonder what factors limited paper-tape read speed? Was it difficult to get phototransistor response down to the millisecond range? I would think a set of nine phototransistors hooked up to input pins wired to an 8080 or 6800 could handle data coming in at an arbitrary unregulated speed with far less electronics than I see in typical high-speed readers with serial-port outputs. Any idea what difficulty there would have been with connecting a reader "directly" to a processor, and having the processor resolve the timing? | |
| Oct 28, 2022 at 15:20 | history | edited | davidbak | CC BY-SA 4.0 |
added 7 characters in body
|
| Apr 15, 2021 at 8:46 | comment | added | Stefan Skoglund | @alephzero you make me feeling old --- I had some time in 1982-83 (i remember that it was winter - so maybe early 1984) practice while being in school (last year) in an mechanical manufacturer. They had NC machines which used fairly wide black punch tape as as input. The foreman = shop boss (verkmästaren) had an early portable CPM machine to input the program and punch it. | |
| Apr 14, 2021 at 7:37 | comment | added | Simon Wright | Being on the 6th floor meant we could (carefully) dangle the tape whose centre had dropped out down the less-used stairwell | |
| Apr 12, 2021 at 18:39 | comment | added | rcgldr | At my fist job in 1973, we used a photo reader that read paper tape at 600 characters per second, that just spit out the paper tape to create a large pyramid shaped pile of tape on the flow. We used electric winders to quickly wind the paper tape from the pile on the floor, and the tape rarely snagged. | |
| Apr 11, 2021 at 18:31 | comment | added | DrSheldon | @davidbak: Just edit out the "predated" paragraph, and your answer will be correct. | |
| Apr 11, 2021 at 18:29 | comment | added | DrSheldon | @tofro: You are describing the Jacquard loom. The Bouchon loom predated that, and used a paper roll, not punch cards. See picture commons.wikimedia.org/wiki/File:Basile_Bouchon_1725_loom.jpg | |
| Apr 11, 2021 at 17:55 | comment | added | davidbak | @DrSheldon - if you're being pedantic, fine. But though that's awfully nice to know for historical/academic reasons, it doesn't IMO have much to do with the OP's question which is clearly about 20th century machines and programmer's and their workday. (See his linked PDF.) | |
| Apr 11, 2021 at 16:31 | comment | added | tofro | @DrSheldon The early paper tapes for mechanical looms - weren't actually tapes - But rather clipped-together from single cards. So, tape and card aren't that much apart than the OP seems to think. | |
| Apr 11, 2021 at 14:31 | comment | added | DrSheldon | "Punch cards long long long predated paper tape." Really? The first paper-tape loom was invented in 1725 by Bouchon. The first punch-card loom was invented in 1804 by Jacquard. Piano rolls and Hollerith punch cards appeared at about the same time in the 1880s. | |
| Apr 11, 2021 at 0:21 | comment | added | alephzero | Even in the magnetic-tape era, card decks were sometimes used to transfer software between different computer installations because of incompatible mag tape standards. And I remember an occasion where we had shipped off about 15 boxes of cards to a customer, and after the courier had left somebody noticed there was a card lying on the office floor. We sent the customer a telex saying "oops, sorry, when you receive the delivery, please punch a card containing the following code and insert it in the correct sorted position in box 12..." Try doing that with a corrupted download from the web!!! | |
| Apr 11, 2021 at 0:10 | comment | added | alephzero | Durable and non-tangling "paper" tape did exist. It was used for controlling early computer-controlled machine tools, for example. The stuff was actually almost as thick and stiff as punched cards, and with a waxed surface that was oil, grease, and machine-tool-coolant resistant. But it was completely impractical for hand-editing work. | |
| Apr 11, 2021 at 0:01 | history | answered | davidbak | CC BY-SA 4.0 |
