1. 33
    1. 4

      You can also order a Unicomp keyboard with APL key caps. It’s awesome.

      I wonder if they could do APL and Mac. That’d be rad.

      1. 5

        I remember my first job at Esso Petroleum in London in 1987; while I had a 3278 terminal, some of the others had 3279’s with APL symbols on the keycaps too, rather like the image on this page about such a keyboard converted to USB with the caption “Exhibit #1 - The OG of OG IBM APL keys”. Both terminal types had buckling spring keyboards. Happy times (and I guess I didn’t know at the time how lucky I was!).

    2. 3

      The Unicomp boards are really nice, and they sell more than the Model M. I picked up a Sun Unix layout model about a month ago and it’s great. The buckling springs are so nice to type on - heavy but crisp.

      They will also customize any of their boards for a reasonable fee, so if you have something specific that you want they can probably accommodate you.

    3. 3

      For someone who’s never used buckling spring, how is the feel different from, let’s say, cherry mx browns or blues?

      Always been curious about the Model M. It seems to have a dedicated loyal following.

      1. 9

        They’re a lot heavier.

        1. 9

          Also you spouse, cat, neighbours, and your neighbour’s neighbours will know when you’re coding at night.

      2. 6

        I love the sound and feel of buckling spring switches. They have a satisfying ‘snap’ to them when pressed – I can’t think of a switch I’ve typed on that had better feedback. That said, the Model M itself is not a very ergonomic keyboard, which is why I don’t use one. I’ve become quite dependent on having 12 thumb keys, nice hand separation, concave key wells, and an ortholinear layout..

        1. 1

          What keyboards do you recommend that meet your criteria for being ergonomic?

          1. 1

            Based on the description, I think spudlyo is thinking of the Kinesis Advantage.

          2. 1

            Yup, as avh-on1 guessed, I’m a Kinesis Advantage user, and have been for the past 10 years or so. Nowadays there are a few other keyboards that have some of these features, like the ErgoDox EZ which are gaining in popularity. Since switching to the Kinesis, I’ve not had any issues with pain in my hands or wrists, and now that Control and Meta are on my thumbs no reoccurrence of “Emacs pinkie” either.

        2. 1

          I’ve always wanted to try a nice ergonomic spring-buckling keyboard. Sadly, no one seems to make them (and the physical layout seems different enough that adapting an existing one would be too much effort), so I’ve stuck with my trusty kinesis.

      3. 4

        Cherry MX blues feel linear but with a high-pitched bump in the middle (which makes sense because they use a metal click leaf). The Model M’s buckling springs feel heavier and “springier” (the force curve isn’t linear), and make a lower-pitched sound.

      4. 3

        MX switches are essentially friction based, using plastic on metal to create fric…feedback. Buckling springs operate in a entirely different manner and are hard to compare. You press down on a spring, which eventually buckles and creates the sound, the tactile feedback and triggers the key press.

        I kinda liked my Model M (which I found in an electronics dumpster at work), but decided to throw 400 USD at a new Model F keyboard when drunk for the lulz.

        The Model F is such a nice keyboard. It’s hard to go back to anything else! But it’s also loud and ping-y so I can barely use it at home or at work (which these days is the same thing haha).

      5. 2

        It’s like the difference between hitting a wiffle ball versus a baseball. Buckling springs are heavier and more gratifying.

        Full disclosure, I prefer cherry reds. to extend the metaphor, they feel like swinging a wiffle ball bat and missing, which sounds bad, but I like the effortlessness. I have been through both unicomp and IBM model Ms and loved them but at some point I joined the light (actuation force) side. I think the reason the unicomp ones are lighter is because they have an aluminum backplate (rather than steel)

      6. 1

        You have to press harder. Laying your hands across the keyboard will not type a key by accident. The spring has a bit of a ringing sound after the keypress. “Ka-chunggg”

    4. 2

      Keypress to event on usb latency measurements would be nice to have.

      1. 1

        If you worry about latency over USB, consider going back to PS/2.

        1. 2

          Going back to old hardware isn’t the solution. Making current/future hardware good is.

          Objective metrics are a necessity if we are to make any progress.

          1. 3

            USB vs PS/2 is pooling vs interrupts . Different designs for different things but USB will always have some measure of latency compared with PS/2.

            1. 3

              USB 3 switched off polling on to full duplex. Unicomp keyboards aren’t USB 3, though.

            2. 1

              Yes, I’m aware how even input devices are binned into time slots.

              This doesn’t mean a keyboard or mouse can’t be faster than another. Even with polling at 1000Hz, it does usually take several slots from keypress to sending the event on USB. Because everything is shit nowadays.

          2. 2

            I honestly don’t think anybody will ever care to redesign a generic interface like USB to account for latency like this.

            And as it turns out, the Unicomp Model M seems to be pretty decent in at least one latency measurement:


            …however, that article is filled with (quoting):

            …then throws in a bunch more scientific mumbo jumbo to say that no one could really notice latencies below 100ms. This is a little unusual in that the commenter claims some kind of special authority…

            …so I’m not so sure about the validity in that measurement anymore. :)

            Oh this was interesting as well, but not concerning keyboards but latency measurements: https://thume.ca/2020/05/20/making-a-latency-tester/

            1. 2

              …however, that article is filled with (quoting)

              He’s got a point though. Human reaction time has nothing to do with human ability to notice latencies under 100ms. According to the literature, musicians definitely can, and drummers will identify jitter even under 3ms. He’s absolutely right in pointing that out.

              …so I’m not so sure about the validity in that measurement anymore. :)

              Measurements are fortunately objective, regardless of the guy’s opinions on human sensitivity to latency.

              The worst is that keyboard latency is just a small part of the whole latency pipeline, from keypress to reaction. Every step adds latency, and it’s important that nothing neglects to minimize it.

              If the input device does do so at the very start of the pipeline, it is already hopeless for the whole.

              1. 2

                Measurements are fortunately objective, regardless of the guy’s opinions on human sensitivity to latency.

                Yes! My comment was about his use of language, especially “scientific mumbo jumbo”. It implies ignorance and/or hubris regarding his own method & choices. Science is not perfect by any means, but dismissing an article/someone else for using terminology within their field feels wierd. Dismissing others findings feels alarming and makes me suspicious.

                The worst is that keyboard latency is just a small part of the whole latency pipeline, from keypress to reaction. Every step adds latency, and it’s important that nothing neglects to minimize it.


                I remember reading Kirsch, “The Intelligent Use of Space” at uni, and one of the parts applies here (I think). He studied Tetris players (very good Tetris players) and how they made use of the ‘space’ in the game to figure out which parts they were playing before they could even see them. Given that competitive Tetris players still use old television sets to practice (due to old TVs very low inherent latency?) I guess they care a lot about this, and try to minimize the entire latency by any means necessary.

        2. 1

          I made sure my most recent desktop includes PS/2 ports because I was hoping to plug in my vintage Model M, then noticed that I’ve come to depend on modifier keys it’s lacking… Maybe it’s time to see if I could hack those in.

    5. 1

      I’ve owned a couple of Unicomp boards now, and my daily driver for much of the last couple years has been an EnduraPro - basically a Model M with a trackpoint.

      I haven’t been quite as satisfied with this most recent board as with the first one I owned. It had an e key that would get stuck for a bit there, and the trackpoint has never worked remotely as well as the real-deal ones on a ThinkPad keyboard, though it seemed to get more usable as it wore in a bit. I have a sense that maybe build quality was a little shakier with this one than the first time I ordered from them.

      Still and all, I’m glad they’re in business and hope it stays that way. It’d be great if I had the option to keep a Model M-alike in the rotation for the rest of my working life.