Retrotechtacular: A Desktop Computer From 1965

About eight years before the Xerox built the Alto at PARC and over a decade before the Apple ][ premiered, Italian business equipment manufacturer Olivetti produced a bona fide desktop personal computer. When Olivetti debuted this typewriter-sized marvel in 1965 at a business convention in New York City, people were in absolute awe that this tiny, self-contained unit could perform the same types of functions as the hulking room-sized mainframes of the time. Some were sure that it was simply a small input device for a much bigger machine hiding behind the curtain.

But the revolutionary Olivetti Programma 101 was no joke. It performed standard four-banger operations and could handle square root and absolute value calculations. The Olivetti had 16 jump instructions as well as 16 conditional jump instructions, which put it firmly in state machine territory. Programs could be printed on a roll of paper or stored long-term on long magnetic cards.

Magnetostrictive acoustic delay line memory. Image souce: The Old Computer Hut
Magnetostrictive acoustic delay line memory. Image source: The Old Computer Hut

In order to store data internally, the Programma used magnetostrictive acoustic delay line memory wherein a thin wire is twisted by a transducer inside of an electromagnet. Each zero coming in twists the wire one way, and each one twists it the other. These torsion waves move down the wire to a piezoelectric transducer on the other end. The signal is amplified, reshaped, and sent back into circulation until called upon by a program.

At at an introductory price of $3200, people and businesses all over the world clamored to get their hands on these machines. Olivetti could not keep up with the ridiculous demand, although they did manage to make about 10,000 of them per year.


There are several videos below. The first is a demonstration of the Programma using two programs created by the presenter. The first program is fairly simple and calculates the Fibonacci sequence between two limits entered at runtime. The second is much more taxing. For any number entered, the program will return a 1 if the number is prime. If not, the Programma calculates and returns the smallest divisor.

Wish you could get your hands on one of these magnificent beasts? We can’t help you with that, but here is a simulator. If you do find one, here’s the manual (PDF). Oh, and you can make your own delay line memory with one of those 101-in-one circuit trainers.


How it works (hit CC for English subtitles):


Old promotional footage interspersed with discussion by the Programma’s engineers (and Bruce Sterling!):


Firing up a Programma after 45 years:


Thanks for the tip, [Fabio]!

Retrotechtacular is a column featuring hacks, technology, and kitsch from ages of yore. Help keep it fresh by sending in your ideas for future installments.

44 thoughts on “Retrotechtacular: A Desktop Computer From 1965

  1. I got to see one of these back in, oh, maybe 1972 or so. I was in high school and one was loaned to the math department for a few days. I was fascinated. Even the teachers didn’t know much about it. This sure takes me back, I knew as soon as I saw the picture it was the same thing we messed with back then. Guess my age is showing, eh? sigh.

    1. My high school had one as well. The math teachers had it but didn’t know how to use it and didn’t have or didn’t share the manual. It wasn’t that bad to use, but was limited to number crunching. The magnetic paper cards weren’t that reliable and after entering in the same program over and again I lost interest.

    2. Doug is the first name here. remember the same time period. The big desk top cacacularrs where for imdtructor only tool. By the time we where seniors kids whose dada worked for the major oil companies had pocket calculators. By the time I went to tech school for Electricity/Electronics a year or so later I had a four function calculator and a slide rule for square root functions.. With a year of oil field wages a calculator with the square root function was a luxury when I had the slide rule left over from High School. and it would still be more time when a calculator that could replace the slide rule entirely was affordable.

  2. I used one of these in the design office of a shipyard in the UK in the early ’70s, I programmed quite a number of ship calculations on this machine for stability and other calculations. It was a most useful computer; for more complex computing we used a time share computer on the other side of the country we accessed using a teletype machine and punch hole tape – this was a very slow process.

  3. This is amazing. I was always under the mistaken assumption that nothing like this existed prior to about 1971. I happily stand corrected. I would love to get a chance to play with one, provided there was an expert nearby to teach me how to do it. I’m very curious about how the programming is done (both how the paper is encoded, and the encoding scheme itself), how the memory works (beyond the overview given by the author), and the nature of the circuitry doing the calculations.
    I realize that it’s not much more than a programmable calculator by today’s standards (My TI 36X is no doubt far more capable), but for something like this to come out in 1965 was nothing short of revolutionary. I’m honestly surprised that I don’t remember hearing about this before now.

  4. We have one in our museum. Unfortunately 952 contacts need to be replaced on the backplane to get it working again. :(
    I really like the way the boards stack together
    The keyboard is mechanically encoded in 7 bits and the printer has one hammer and a drum with all characters behind the paper, timed by a electromagnetic encoder on the drum shaft. Very neat.

      1. The contacts look like the contacts in these: (page 2, single row contacts). The problem is that they tend to break at the right-angle bends, apparently because they have been gold-coated, THEN bent, causing cracks in the gold layer, and allowing oxidation of the contact spring material.

        The connectors shown can’t be used directly because the spacing is different; 3D-printing a body and populating it with the contacts from them might be an option, but it’s one we’re not really looking forward to.

          1. Is the *body* as you call it still intact? The beryllium copper mentioned would have the malleability needed but I doubt you would get the required spring tension.

            You can buy spring wire that is used for hobby modelling and bend it with a bender (a stepper motor plate and two pins) slowly so you have long lengths of many pins. Then you can plate with electrolysis. Silver is an ideal conductor (better than gold in fact) but it is also much harder than gold. It’s usable but don’t disturb the cards much once in situ. You can get gold vapour plating done without too much expense. Or even ENIG.

            If you can tell me the spacing between contacts, the card (PCB) thickness and number of contacts then I can explore other possible options. The previous link sent had pin spacing of about 5/64th’s and was and odd dimension.

          2. If you think you can find a commercial connector with the right pin spacing AND the staggered pin layout AND the right number of contacts, you are very much mistaken. I have worked through all kinds of selections from Digikey, Mouser, RS, Farnell as well as directly from connector manufacturers, read the datasheets (several dozen at the very minimum) of any of the connectors that looked vaguely promising, and found NONE. The one I linked has contacts that are about the right size and shape, but would need 3D-printed bodies because of the spacing and pin staggering.

            Apart from getting the right connectors crafted (by whatever way) for this project, getting the old connectors out and the new ones fitted, then trying to figure out what other problems are lurking in this particular pile of technology from half a century ago, we’ve got at least 20 systems that require SERIOUS effort to fix, more than double that with lesser problems, over 30 boxes of documentation and schematics to scan, ten or so boxes of tape and floppy media to copy before all the bits fall of (most of the papertape is done, though), and some loose odds and ends such as documenting it all.

            Oh, and we have to move in a couple of months.

          3. Quote [Stoneshop]: “If you think you can find a commercial connector with the right pin spacing AND the staggered pin layout AND the right number of contacts”

            Pin cont doesn’t matter. Cut a longer connector. Join smaller connectors together.

            Pin spacing can be a multiple of the existing spacing and given that in those days measurements were *fractions* of imperial measurements then a multiple of the pins spacing is more likely than you think.

            Staggered doesn’t matter either. Just about all double sided connectors have pins that be reversed. Reverse every second one – no biggie.

            AND the original design engineers had a plan ‘B’ with the staggered connector holes on the cards, I wonder what that was?

            Anyway I am getting a less than warm reception to these questions lol so I’m outa here.

          4. When some wiseass comes in suggesting things we’ve already investigated, tried out and found wanting, or take too much time and effort taking into account all the other things we have on our plate as well, then yes, you can bet I won’t be overly enthusiastic hearing those suggestions yet again.

            Cutting up a few connectors to create a longer one won’t help if the contact spacing DOES NOT MATCH anything available commercially today.

            Snipping pins on alternate sides won’t help if the contact spacing and pin offset DO NOT MATCH anything available commercially today.

            People who come in and actually help to fix things are very welcome. People who just tell us what we should do, when we’ve already tried that, from somewhere across the internet? Tends to get tiresome.

          5. @[Stoneshop] I think you’re in the wrong place here.

            This place is mostly engineers.

            We generally help each other out here. It’s not uncommon that when we have a ‘sounding board’ question that someone comes up with something a little off beat. We are able to deal with that without calling people a wizeass because we generally appreciate the assistance we get here.

            This is also an open forum – no moderations so when people like you decide to flame then you can expect back what you give out.

            So first off you have publicly displayed your complete lack of intelligence by choosing to slander a veteran in this field. I am over 50 and I expect I have far more experience and success at restoring old equipment.

            So to translate that to something that even someone like you can understand – fu(k off to your little hole of unwarranted ego you fu(king wanker!

          6. You are the one who interpreted “replacing 952 contacts” as a call for “how do we do that”. It wasn’t. It’s an explanation of why the project was stopped, because, as you may have read, we have other things to do, things that too require time, and might well be more worthwhile.

            Oh, and assuming someone else is a) younger and b) less experienced? Just assume as you like. However, you could turn out to have mis-assumed.

          7. Wow [Stoneshop]

            A completely dumb ass fu(kwit would have worked out that it’s better to drop the thread. Did I mention that this is an open forum – didi I?

            We can go on for as long as you like Fu(kwit. Just don’t think for a moment that I am going to fold and kiss your ass.

            And NO I am not going to return to a rational conversation with you because your’e just not worth it. As far as I am concerned you’re most probably borderline personality disorder and the real problem with the restorations is that no one (in their right mind) wants to work with you.

            What did you expect a different response – did you?

            Well perhaps what is happening here is what reasonable people EXPECT to happen when you go calling people wise ass.

            Hey, seriously people have a real problem with associating repercussions with initial actions when they’re teenagers!!! What’s your excuse?

            Posts comments to forum, complains when others comment. How fu(ken dumb is that. Oh don’t tell me – you were just thinking out loud!!!

            So back to you – oh and by the way, I’ll give you a bit of a heads up … others are watching this thread to and their probably judjing you by it.

            If you drop it now then they will thing – well about time!

            And if you keep going then there going to think dumb ass. Just dumb.

            So yeah, please confirm your absolute stupidity for everyone.

  5. Wow. A flash from the past. Used to give courses on this machine at my high school in the mid 70s :).

    We had this little digital computer and in addition an analog computer which you could wire up to an osci to play around at our school. The Programma 101 had its own little room, as it was quite expensive equipment at that time.

    The Programma featured only a few registers, but you could split them in half, reducing precision by half as well. I think you had at most access to 14 registers. The only real function available was the square root, so usually one of the first programs one wrote were some calculations yielding other mathematical functions. Registers were actually stored in a delay line (mentioned and displayed in the above article). Programs would be stored on sheets of magnetic paper (!), about the size of a letter envelope. The magnetic card reader is actually the metalic part right in the center of the machine, above the keyboard. There were also programming sheets printed on paper available which you could use to design and write down a program before keying it into the machine.

    Another, similar beast from the past is surely the PET 2001 (released a few years later) which was one of the first >complete< BASIC-computers private people and small businesses could afford buying at this time (end of the 70s). Other computers at that time were available, but either lacked proper housing or needed to be connected to additional peripherals and data storage devices.

  6. Here’s a hack to try. Make an acoustic delay line memory using car horns to relay the bits and keep them moving. It would need microphones and some way to actuate the horns on the cars. Might be best to use many different makes and models of cars with different horn notes so that whatever is listening for the bit-notes can discriminate to only hear and repeat the horn next to it.

    Also, not a project to do near anywhere near any village, town or city.

    Why do something like this? Science! And to very loudly demonstrate the principle of the acoustic delay line memory.

  7. If some one walked up to me and handed me a piece of wire and asked me to store data on it, I would say fu(k off idiot, it a god damned piece of wire, you can’t store data on wire – fwit.

    lol – and there it is, from the 1960’s – wow

    1. Well, the Germans (maybe others) had wire recorders (pre-tape recorder) during WWII.
      An iron wire spooled off of a reel through a coil which magnetized it at IIRC, audio frequency, before being taken up by another spool.

      1. I think that these two techniques are a little different. The wire recording was magnetic. This one here is acoustic. They used sound through a solid as electrons travel too fast. The sound is picked up at the far end and any ‘data modifications’ are made before it is sent through again making the wire a media for a constantly moving sound wave patter. Well if I understand this correctly that is. It’s kind of like bubble memory but works at the speed of sound.


  8. We live in a world where there’s an arduino shield for almost any need, so we don’t have to think too hard about how to solve problems. Back then engineering required more creativity and out of the box thinking.

    1. I think the entirely mechanical, 5-bit teletype terminals featured here a few weeks ago were an even better example of that. A UART that used 5 cams, lifting up levers that contacted bars that were either pushed forward, or not, by the keyboard, to send the key code. Well it wasn’t entirely a UART but it was a serial transmitter. Receiver was something similarly complicated but ingenious, can’t remember right now.

    1. LOL. I went to a local museum in my home town and saw a piece of equipment that had written on the side (in my own hand writing) my old office number. I had written it there so they knew who to call for service.

      A sobering moment indeed.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.