WEBVTT 00:00.000 --> 00:10.160 So we are starting now with a small panelist discussion, so as I mentioned at the beginning 00:10.160 --> 00:15.040 of the session, but not everyone was there, so we had normally a talk by Mark Mayer, but 00:15.040 --> 00:21.840 unfortunately he's sick, he could not join, so we just hope that he recovers very soon, 00:21.840 --> 00:26.000 and then we have a small discussion with some panelists, not all of them, it's impossible 00:26.000 --> 00:31.600 to invite 15 person, especially that we have only one microphone, I didn't know that also in advance. 00:32.640 --> 00:40.560 So I just introduced the topic, and then we can pass the mic and people can express, let's say, 00:41.600 --> 00:46.240 the opinion, so one of the first things that comes to my mind, and it's a bit follow up 00:46.240 --> 00:50.160 of what the metry has been presenting, so when I see what's being done today, 00:51.040 --> 00:57.920 in vitro computing, there's a hardware part, software part, I would suggest that we start with the hardware, 00:57.920 --> 01:05.040 so it's a little bit, what I see very much is that either there are new systems or new 01:05.040 --> 01:13.600 new boards being built to adapt existing like old miles or old joystick to make them work with 01:14.240 --> 01:20.720 modern computer, or we see the opposite as well, we see like adapters from USB to have the 01:20.720 --> 01:27.840 the PlayStation controller being able to work on a Commodore 64 or something else, Atari, 01:30.240 --> 01:37.840 these are the kind of things also adapters for the video display, for example, so we see 01:37.920 --> 01:48.400 new boards being created, sometimes it's this little, like ESP or whatever, or Raspberry PICO, 01:48.400 --> 01:56.720 PICO or something like that. Maybe one reflection is that something that you see as well, 01:56.720 --> 02:00.960 secondly, do you see something that that is missing so far that that could have been done, 02:00.960 --> 02:05.120 but that has not been done yet, and that would be like a nice challenge. 02:06.000 --> 02:11.280 So I suggest that I just pass to you and then we go to the full line. 02:13.040 --> 02:18.480 Yeah, I mean, I want to add to that that it's not only about the I only devices, but also about 02:18.480 --> 02:25.360 the computers, so building new extension hardware for old computers is also something that is 02:25.360 --> 02:34.240 very popular, I think, there is this preservation aspect of it, and some people like the 02:34.320 --> 02:41.040 preservation more than they like extending existing systems with modern stuff and think of that as 02:41.040 --> 02:47.200 being not true to the history of things and others do other stuff, so I like, I like these like 02:48.800 --> 02:56.800 cross, cross centuries systems, but others try to find all the original parts and all of that is good, 02:56.800 --> 03:00.960 I think, I mean, it's part of history preservation project if you are. 03:03.360 --> 03:08.160 Yeah, I will say, these two directions, two opposite directions you have mentioned, 03:08.720 --> 03:17.840 probably have, well, when you add a modern IO devices or something similar to an old computer, 03:17.840 --> 03:27.920 it's, you know, a way to get work with it with despite some apps and parts, so it's a sort 03:27.920 --> 03:35.840 of digital prosthetics, maybe. And it's definitely important and very practical task, 03:36.800 --> 03:42.320 because, well, sometimes some devices, I audio devices, were so indruble that 03:43.200 --> 03:52.080 impossible to find them, especially in an operational state. The opposite task is, yes, it's mostly 03:52.080 --> 03:59.920 a preserving of a digital heritage, and while you can run software like operating systems or similar 03:59.920 --> 04:07.120 stuff in emulators on modern computer, because you may have a problem to feed these IBM mainframe 04:07.120 --> 04:19.920 in your home, even if you find it, but the emulator can, but, you know, the user experience, 04:21.520 --> 04:29.280 sometimes demands, sometimes it's different when you use modern pointing device, well, 04:29.920 --> 04:37.040 I'm speaking about mice, as you may guess, so you can preserve running software, but the 04:37.040 --> 04:43.760 user experience will be a little bit different, because, well, you can simulate a display, 04:43.760 --> 04:52.000 but how to simulate input hardware, okay, it presses good keyboard on screen, it's definitely 04:52.000 --> 04:58.960 better than anything, but from the ergonomics point of view, it still leaves place for questions. 04:59.600 --> 05:07.040 So connecting all peripheral devices to modern computer and to simulators, I would say, 05:07.920 --> 05:11.440 play this important role of preserving the user experience. 05:14.480 --> 05:18.320 Yeah, you may not emulate us, I think that it's also a very important 05:18.320 --> 05:23.680 area of which we're computing, because there are many systems that are so rare nowadays, 05:23.680 --> 05:31.760 that you can't have one, so in order to experience the system emulation is your only way, 05:31.760 --> 05:36.880 and there are actually still a lot of systems that can be emulated today. 05:36.880 --> 05:42.080 I will be describing one of them today in my presentation, so I would say it's important, 05:42.960 --> 05:49.520 because otherwise, it's just not possible to use this machine at all and it gets to be forgotten. 05:53.920 --> 06:00.160 Yeah, we are basically dealing with a very old machine from the 60s, and it's a mine frame. 06:00.160 --> 06:08.560 Some components are literally unusable, because they're full of little mud, 06:09.440 --> 06:17.040 little mud, because they were stored for decades in the wrong place, so it's impossible to 06:17.040 --> 06:24.480 recover them. In some cases, we also had the problem that we can restore most of it, but the 06:24.480 --> 06:30.560 mechanical parts are so delicate that we don't want to use that, so we just replace that with a 06:30.560 --> 06:38.160 microcontroller that does the same thing, but we try to keep it as visually close as the original, 06:38.160 --> 06:45.440 and you'll see more about this in our talk, but sometimes it's also me and my old PSX controller, 06:45.520 --> 06:53.200 I love, and I have that for 20 years, and I just want to use it on my Xbox Cloud gaming machine, 06:53.200 --> 06:59.600 and I just made a converter to USB, that's this big, and I have open source, so it's maybe 07:00.000 --> 07:07.520 replaced the PSX one. It's actually very easy to make an embedded device that converts an old bus 07:07.520 --> 07:13.760 to a USB modern device, so you can even restore, I don't know, your old steering wheel, 07:13.840 --> 07:20.160 I added to reproduce the same experience you had as a teenager or something, and you want to add 07:20.160 --> 07:28.080 something. For me, it's extremely important, the discussions about emulators, because for me, 07:28.080 --> 07:33.440 emulators are not a way to reproduce the original experience of using certain machines, 07:34.240 --> 07:37.520 mostly because you don't want the original experience of running machines, you don't want 07:38.480 --> 07:43.600 two rooms full of iron, just to have half a mega of computing power. For me, 07:43.600 --> 07:48.240 what's most important is that emulators are running documentation, which means that you know 07:48.240 --> 07:51.680 that the documentation you have is correct, and you can probe it, you can compare it to the original 07:51.680 --> 07:58.240 hardware, if you're so lucky to have it, so yeah, it's a very important thing, so even for a 07:58.240 --> 08:02.960 younger people, like in the programming scene, that they are this kind of factions between, 08:02.960 --> 08:07.120 you know, or real emulator, or real if GA is, I should have used the main, 08:07.840 --> 08:11.760 doesn't matter, that's not the thing, the thing is that we have that, we have the machine description, 08:11.760 --> 08:17.520 we can recreate it, so we have the FPGA, we have the emulator, we can, and also something very 08:17.520 --> 08:23.840 important is like, don't be ashamed or don't be scared to mix and match the technologies, 08:23.840 --> 08:30.080 centuries, and whatever, my personal, let's say religion, fixed point is that your modifications 08:30.160 --> 08:35.280 should, in my opinion, should be irreversible, some people don't agree to that, but yeah, 08:35.280 --> 08:41.360 like interface, the old machines to the new one, the gods of your computing will not be against you 08:41.360 --> 08:45.680 for that, actually quite the opposite, you will breathe new life into new machines, and that's very 08:45.680 --> 08:51.280 important. Thank you. So the problem with being last is everything's been said, so 08:53.520 --> 08:58.560 so I'm going to have to rely on false personality and jokes, well that's certainly not going to work, 08:58.800 --> 09:06.480 so my thing with the IO thing is not limiting it to pieces of what I would call simple hardware, 09:06.480 --> 09:11.280 mice, I can power to be simple hardware, potentially omiters, yeah, I understand them, 09:11.280 --> 09:15.440 their specs are out there, but there's also IO that talks to things like processes, 09:16.320 --> 09:21.440 back in the good old days, when I was actually young, don't laugh at that one, I'll tell you 09:21.440 --> 09:28.000 when I'm doing the funnies, back when I was young, BBC Micro had a tube port that would allow you 09:28.000 --> 09:33.760 to put other processes to connect to your BBC. That's fine, you can emulate that stuff and that 09:33.760 --> 09:39.760 works, but then fast forward another 20 years, and you can't do it anymore, because you can't do it, 09:39.760 --> 09:44.240 any IO to the processor, because someone's put some gunk in there that says, oh, unless you're 09:44.240 --> 09:49.200 running on the hardware, you can't work out what's going on, Northbridge, Southbridge, Xbox, 09:49.200 --> 09:53.920 chips, they have special things in, we're not going to be able to ever emulate them until we 09:53.920 --> 10:01.520 crack what's going on under the hood. So the IO from the olden days, versus what we're going to have 10:01.520 --> 10:07.680 as a problem coming up in 20 years, are different problems, and it's solvable to say, yeah, 10:07.680 --> 10:13.120 I can see a transistor, I know how that works, but you get a chip from 20 years ago, you can't, 10:13.120 --> 10:18.560 most people can't just take the top off, scan the silicon inside and work out how to emulate that 10:18.640 --> 10:28.560 or provide any sort of modern version of it. Thank you. I don't know if you wanted to 10:28.560 --> 10:35.760 ask something. Okay, like Steve and I said, probably everything has been said already, 10:35.760 --> 10:40.080 being last so that's why I will switch to software, and I think Michal, you're the first 10:40.080 --> 10:45.680 emulated, actually, the, sorry, who mentioned the emulation, emulation part, 10:45.760 --> 10:50.240 think when we look about software, again, I also mentioned at the beginning of the session that 10:50.240 --> 10:56.880 there is mostly game, new games that are still being released, operating system, all these, 10:56.880 --> 11:01.440 most probably would not happen if we would not have, like, modern emulation today, because people 11:01.440 --> 11:06.400 are not working anymore or not that much, and I don't think that many people are still working, 11:06.400 --> 11:13.280 we don't the real old hardware, hardware, they're working on emulated, emulated systems. 11:14.240 --> 11:20.000 So there may be same kind of reflection, we can go quickly, we still have, like, seven minutes, 11:20.720 --> 11:29.200 around the room, you want to sound off? Ah, yes, okay, okay, yeah, just about emulation, 11:29.200 --> 11:34.640 our things, still missing, or yeah, a bit to your perception on this topic. 11:36.480 --> 11:38.160 So the point will be, be going first, 11:38.880 --> 11:44.480 if you're going to find something really quick to say, when everyone's looking at you, 11:45.200 --> 11:50.240 I've written a few emulators, and they're really quite good fun to do up to a point. 11:50.960 --> 11:54.880 The problem is after you get to the point where the first thing works, like, the basic pump 11:54.880 --> 12:02.640 comes up from the ROM, it's like, how do I prove it works? How do I know that this ROM is emulated 12:02.720 --> 12:08.320 correctly for everything that will ever run on it? And it's one of those problems of, 12:09.120 --> 12:14.880 it's very easy to get bored with it, it's very easy to get frustrated with it, and unless you're doing 12:14.880 --> 12:21.600 a full spoiler simulation rather than emulation of a platform with all the timing, 12:21.600 --> 12:26.080 idios and crassities that the original had that your software emulator might not have, 12:26.800 --> 12:31.360 how are we ever going to know that we have actually preserved the thing we were trying to preserve 12:32.080 --> 12:34.720 in our software emulator simulation? 12:36.880 --> 12:42.320 Yeah, like, I have as well only a small thing to add to this is that you can also not care, in the sense 12:42.320 --> 12:46.960 that adding an emulator, that only emulates five instructions out of 10, it's good than having 12:46.960 --> 12:53.040 no emulators, and adding an emulator that like, implement 35 instructions, but in three instructions, 12:53.040 --> 13:01.120 like the fifth cycle is slightly wrong, yeah, we can find it later, like as soon as we have 13:01.120 --> 13:09.680 something to start to can jump off of everything should be fine, and so it's a complicated task, 13:09.680 --> 13:15.920 it's a very involved task, it's very time consuming and a brain power consuming, but also 13:16.480 --> 13:21.280 nobody's forcing you to go to completion, this is a hobby for some foremost, at least for now, 13:21.280 --> 13:27.360 so enjoy it, because it's great to be part of history and to be in connection with this story, 13:27.360 --> 13:30.880 also because it can not have new a lot more of what's going on right now. 13:33.040 --> 13:38.000 So yeah, it's great that you're ready to question about the difference between simulators and emulators, 13:38.000 --> 13:43.200 it's also perfectly okay to read the simulator if your purpose is not using what the machine 13:43.200 --> 13:48.480 was doing, but for instance, verify that the software you've dumped from a very old media 13:48.560 --> 13:54.160 is doing what's supposed to do, so and you know, you can double check that, but of course an emulator 13:54.160 --> 13:59.360 gets closer to what the machine was doing and that's a right approach for instance to verify 13:59.360 --> 14:06.320 that the machine that you restoring is doing what it's doing, and the gap between there is documentation, 14:06.320 --> 14:11.040 how much the documentation you can find, we were so lucky to have a machine from the 14:11.040 --> 14:15.440 60s as you will see, but with a lot of documentation with it, so we could start attacking the 14:15.440 --> 14:20.000 problem from both sides, and one thing is restoring the hardware, as you said, the beginning, 14:20.000 --> 14:25.200 and the thing is restoring the software, and those things go in parallel, and of course the software 14:25.200 --> 14:30.640 until the machine is done is probably taking us a few years, so in the meanwhile, we want to be able 14:30.640 --> 14:35.120 to verify the software that we dumped, it's corresponding to what it was supposed to do. 14:37.520 --> 14:44.480 I think it's also worth mentioning hardware emulation, so today you have several FPGA platforms 14:44.560 --> 14:52.160 that can reprogram to recreate a machine very close to the original hardware, so this is maybe a 14:52.160 --> 14:58.000 closer to what you said, trying to get the original hardware XPS and not just a window in the computer, 14:58.000 --> 15:00.800 and I think that's also an interesting area here. 15:03.680 --> 15:13.680 Well, if speaking about emulators and simulators, I would mention that starting from some years, 15:14.320 --> 15:22.160 we have got official SDKs with emulator included for some old platform or something like this, 15:22.160 --> 15:30.240 and it's not obviously true for 60s, but somewhere in 80s, when these software development 15:30.240 --> 15:36.560 kit with emulators do appear, it became a little bit easier to preserve. You just use nested 15:36.640 --> 15:45.920 virtualization to run this simulator in some virtual machine, which itself may sit inside of 15:45.920 --> 15:55.040 some virtual machine. Unfortunately, the difference in all resources, computational resources 15:55.040 --> 16:06.240 helps to drive this without much pain. Obviously, that's not true for, well, prep 80s systems, 16:07.200 --> 16:14.000 maybe a present is for sure, so I would say that the biggest challenge here is 16:15.920 --> 16:26.880 yeah, to run something which was invented before virtualization exists in heads of some engineers, 16:28.400 --> 16:31.920 late days just in all, and it's a task sometimes. 16:32.000 --> 16:40.400 To me, it is quite fascinating that for a single person, it is now possible to 16:41.200 --> 16:46.960 with the modern tooling to recreate software simulators or emulators of systems that have been 16:46.960 --> 16:53.920 created by dozens or hundreds of engineers, and that really is also speaking to the progress that we've 16:53.920 --> 17:00.320 made, so of course, fundamentally there is not a big difference between these old systems 17:00.400 --> 17:05.040 and what we have today. There's quite some architectural similarities in how they work, 17:05.680 --> 17:12.480 but from the perspective of how much of that we can now move with the tooling that we have, 17:12.480 --> 17:17.760 that's really fascinating to me, so my recreation of a system that 17:17.760 --> 17:22.320 had dozens and hundreds of people working on that. I mean, it's not complete, but still, 17:22.320 --> 17:26.960 it's just the tools that made me do it, not something else. 17:27.840 --> 17:34.320 It thinks that would be the final words because we are done reaching the time, so many things 17:34.320 --> 17:42.000 I have Hans Dimitri, Michal, Daniele, Antonio, and Steven. Steven, you're staying? 17:45.040 --> 17:48.640 I mean, it feels what now, so many things again to all.