 So that's, we're almost at a wrap. That's 20 years of scale, 21, so I don't know, you know, we were all, we had a break in 2021 with COVID, which I know we're all excited to be out of the house again. 2022, we had to take a little bit of step back, go back to a summer event at the Hilton LAX, which is a great venue, just doesn't fit all you all in the room like we can here. So we're excited to be back to our home in Pasadena. But, you know, we're, we've spent a lot of time this week in our keynotes and in our sessions talking a little bit about the history of our community, of the software that we use, of the relationships we built, how we got to where we are today. Ken, I think we'll continue a little bit of that, Ken Thompson will continue a little bit of that story for us. But wanted to go back a little bit, just a second, look back at scale history. This is about 21 years ago. I had much less hair, I guess, although I'm going to get back to that amount of hair at some point, I imagine starting to go. But yeah, almost everybody that's in this picture is still here in scale in some way or another and you probably see them in the halls, might even be in this room, I can't see you among the crowd, but that's scale one over at USC. So it says November 2nd, 2002. So a lot of the, it's been interesting watching the industry change over the years, logos come and go, speakers come and go, but loved welcoming the community back every year and making new friends, so. With that, I also wanted to take a moment, how many of you folks have noticed the new artwork around scale this year, some of the new designs? So this is, what's interesting, the interesting thing about all these new graphics that we've got this year we've been done 100% with free and open source software using projects like Inkscape. The program that you've got, that you have in your hands was developed using Scribis for the page layout and the printing, everything end to end, and that's the very hard work of two gentlemen that are not here, they did all of this work for the last eight months on all these designs and then they, even though they couldn't make the scale, so wanted to do a big shout out for Chris Rogers and Josh Annler, and I think they're watching the live stream, but they just unfortunately couldn't make it out to LA this week. So, and those are just two of the probably 100 some odd volunteers, maybe even more that make scale happen year round. So if this is an event that you're enjoying, that you think that you'd like to see continue to happen, please do reach out after the show. We're always looking for more hands for next year and it could be something as simple as, you know, ushering people into the room and helping them find seats to helping us with the graphics, to helping us with the network, the program selection, all of that. As we know from our open source communities and our free software communities, many hands make for light work and, you know, this is not, this event wouldn't be possible without those volunteers and without the community that helps make scale happen. So thank you. So yeah, please do. It's just info at SoCal Linux Expo or click the volunteer link on our website. You too can help make scale 21 even better next year scale can drink, I guess. So before we get on to the rest of this, I know there's quite a few of you went through and did that Expo Hall passport program. Not that many of you got all the answers right. Although I did find, as I'm looking at these, I realized that some of the questions have the answers in them. So if you missed those, I'm a little disappointed. It's like not getting your name right, not getting the points for your name at the top of the SATs, I think, but yeah. So we've had a couple of really cool prizes here. The first is I've got a couple of backpacks here from really nice backpacks here from the Bitwarden team. How many folks know are using Bitwarden for keeping their passwords secure? So these are some really amazing Belray bags, entirely leatherless, but seem like they're amazing quality, lots of pockets and such. And so you do have to be present to win these, but is there a Shanglin Chen in the room? Come on up for a second. There's yours. So thank you to the Bitwarden team for that. Hang around, I think we're going to do a picture of all the winners in just a second, if you don't mind. Yeah, just stay for a second. Sam Coleman, man, Sam is the luckiest, either the hardest working Expo passport person alive or the hardest working or both. Sam, where are you? Sam, last year, won a laptop. Sorry, whoever wants, if you want to win this prize next year, you just got to follow Sam around and write down his answers, cheat off of him. That's, he must have gotten, is there a PhD in conference swag or education or maybe? Okay. And then I got one more for, there's no last name on this, you can't, I don't, I'm sure we're gonna have to, sorry. Eric Lawrence. So thanks, Eric. Awesome, enjoy. And he was working hard. I think he came back like three or four times, like I know they're not on the Expo floor, they're not there. I was like, can I get some hints? And so, yeah, no worries. No worries, we're very appreciative of our sponsors, but they also had talks and other things. So yeah, that's, anyways, congrats. So for our final prize, I want to call up Kyle just for a second. And if those of you that don't know Kyle, he gives, he's been giving talks at scale for many years. He's also the president at Purism. Some might say his talks are explosive. But before we call out the winners of the prize, would you mind just telling us really quick, maybe just in a couple seconds what they're winning and why it's awesome? Sure, yeah. So no explosions this year compared to last year, which is great, just a giant Sean head in my talk, but other than that, no unusual things. Yeah, what they're winning is a Purism Libram 5 USA phone, which is made in USA Electronics. It's, we made a design to phone from the ground up to run free software. It runs 100% free software Linux distribution, Pure OS and a laptop kit. And so you can dock your phone to this laptop dock and it turns it into a laptop. It's what I've been using for my personal laptop for like two years now. Instead of a personal laptop, I have my phone and then when I wanna type a long email, I dock it and it's a laptop. That's what they're getting. This is what, like a $2,000 phone, if I recall correctly, plus the dock. So this is the grand prize. And so I hope this person is here. Leif Davison, come on up. So Peter, you wanna just get a picture from us up front here? Cool, so congratulations. And I did promise the Purism folks that there would be a picture of you with the prize winners. So why don't we just do a quick shot here for Peter and the team? So thank you. Awesome. Thanks everybody. Hopefully your seats are still there. I think you might have won prizes, but you might have lost your seats. And that's, I mean, honestly, that's one of the things I love about our community. We have folks coming back year after year and even as their role changed or their interests changed or their focus has changed, they're back here to keep sharing with us. I think Kyle, what was it, skill 11? You were our keynote speaker about 3D printing and spoken about everything from security to electronics to privacy to, again, 3D printing. So it's great to have folks back year after year. Awesome, so let's transition here. We've got, as I mentioned, we've got Ken Thompson joining us and I feel like I don't have to make an introduction because you've packed the room already and you're already fighting over seats to get to know him. But this is one of my favorite parts. I mean, my favorite part of skills, I get to see all of you every year and have this reunion with our, you know, 3,000 or so closest friends and family and do that. And so we missed that when we had to take a break. But the second favorite piece is I get to meet every year as I get to meet somebody that's really just helped enable something foundational in our industry that people that are, whether it's new legends or old legends and Ken Thompson's definitely one of those. He's, like I said, got a long resume that he'll talk with us about today but everything from creating Unix to, or co-creating Unix to being part of UTF-8, Plan 9, Go Lang, how many folks are developing in Go Lang these days? I think I saw a headline this morning that it's up in the top 10 languages being for new software development right now. So that's obviously lots of impactful work. But Ken, why don't you come join us on stage and share the story in your own words? Thank you. Apologies. Can I take half the stage here? Of course. Half of the stage. And then you have your clicker right here. Okay. If you'd like, if not, just forward back. Awesome. So without further ado, I give you Ken Thompson. Thanks again for, he's just close to the speaker. Thanks again to everybody. Well thank you everybody. I don't do this often or much and it scares the hell out of me when I do. So bear with me on stage fright. This is my first slide. Now, it's not a normal first slide. Normally there's a title and a person, things like that, where I live, where I work, whatever. So I had a running title of my 75-year project. This is a project I worked on for 75 years. Now, if you subtract 75 from 20, 23, you get 1940 something. And that's thus the date of the slide. Now, every project starts with financing. And so this is the horse fund. The horse fund is the financing for this project. So anyway, I went on and built the rest of the slide, the rest of the project. And I came back to the front slide, and I was going to fill it in. And right now, this is a draft. And I realized that this is the first draft of the first Google slide I have ever made. And I didn't want to destroy it. So this is the slide. And now, I guess I'll start the talk and tell you what the horse fund is. The horse fund is, of course, the financing of my project. Now, it starts, I'm not that old. But it starts when I was in elementary school or even preschool. And my family, and I guess many families are even most families, when there was a family gathering, they'd give the little kids a present or something like that. And in our family, it was always a dollar. You got a dollar. And I guess it was payola for going to the family reunions or whatever they were. So I have an older brother and a yet older sister who would get this money. And now, my sister was hand to mouth. And the money barely made it to the store. How many of you know the Stanford Marshmallow experiment? OK, well, she was a failure. She was by far the number one poster child. All right, so, and my brother saved for a horse. He was a horse fiend and wanted to buy a horse and ride a horse and everything. So it got the name of a horse fund. And I got money, too, and I called mine a horse fund, too, even though I didn't care about horses at all. Now, these are us. And this is my brother at 14 when he cashed his horse fund in for this car right behind there. That's his horse. And this is Texas. So you could get a license and a car at 14. And he did. And that's my sister, and that's me. I'm trying to figure out which way to go. It's not working. Put it down. Aha, wrong way. OK, now this is my horse. I bought one of the first commercially available tape recorders ever produced. It cost me $127. And that is times about 11 and 1 half cost of living adjustment, whatever you call those things. And I loved it. I took it everywhere. And I was 14 at the time. Complete with spare wheels and tape splicing kits and it comes in around 30 pounds. And so it's technically portable. It does have a handle, but not that portable. Now, I took it to everything, absolutely everything. I'm going to go the right direction again. Yes, I took it to everything. I recorded everything, mostly I recorded music. But I was in a small band, and there was three of us. Calling that a band is probably a huge exaggeration. And it was the voice of a robot I built. Here's my robot if I can get it. There, that's me and my robot. And it was the voice of my robot. And it was also the remote control. It had remote control, but it was real time. You had to push a button to get it to do something. And it was the sequencing of the remote control so you could follow a track. And stuff like that. So that's that part. Now, time goes on. We're in 1982. I used that tape recorder probably 10 years, 15 years. I mean, it was used. In 1982, CDs came out. And in particular, there was a Time Life Best of Year series that came out. And I was just utterly impressed with CDs. I learned how they work and the lasers for them. And it was just fascinating. But mainly, I was well into computers by this time. And I did the calculation. And if you multiply two bytes per sample times stereo mono 2 times 44.1 k, which is the number of samples per second, times 360, 3,600, which is the number of seconds in an hour, you come up with 600, which is about the size of a CD in an hour, maybe more. You come up with around 600 megabytes on a CD. And I was just drooling for that because that was orders of magnitude over whatever you could use, whatever you do. So I was particularly interested in that. One of these days, I'll get the right one. All right, so anyway, this is an example of the particular CD that I was very interested in at the time. It was 1957, all the best music of 1957. I thought that would be a great thing to do to collect all the music of 57 rather than the 10 or 12 that were on this CD. Now, I know I'm on the right one. I give. I'll go back first. It's my fault, which is the button. OK, right, I got you. OK, now, I don't know, space up here. All right, in 1985, this is an adjacent department called the Acoustic Department at Bell Laboratories. It was a sister department to ours, and they were into audio for various reasons, all sorts of audio. But mainly, a guy named JJ Johnson did an encoding called PAC for Perceptual Audio Encoding. And it was based on a model of the ear. And it had one knob on it that's how close the compressed code would be to the model of the ear. And so the better you made the model, the more accurate it was to not making mistakes and things like that. It was actually very good. It predated MP3 by half a dozen years there. And it was actually very, very good. The speedometer on it for how many bits it had to take to do whatever it was expected to do was rated in dB. So it was an exponential scale away from the ear. Sorry, I'm just trying to decide whether I'm going to push down the thing for a second. OK, now I wanted to use it. And so I got Buddy Buddy with the designers of this. And I found an old design for an old modified DSP. Anyway, I had it redesigned by a guy, Skyler Quackenbush, and built a board. No computer in those days could decode in real time, or MP3, but it wasn't there, so it's OK. And they especially couldn't encode in real time. But you could encode in the background, take the data, MP3 pack, and then come out. And so this DSP was fast enough to do the job. And I had an audio section put on it so it can make noise. And then a general purpose bus, a SCSI bus to connect. OK, this is a picture of it. It's in, and most of the incarnations of this were in what you would call today a external box for a CD or a disc or something that you'd put next to your computer. And it used the power supply. This is an external CD drive replaced with a pack card. It stole the power for the CD drive. And then there was a SCSI bus that I've taken off. But it's that back panel card. There's the card up in the upper, upper, right home, the cable. And that cable goes from there to whatever you're connecting. And this is a SCSI disc that's being emulated by a 1G, a monster 1G RAM. So anyway, that's just one incarnation. But there were many, many incarnations of this thing. There was one that had a 3 and 1 half inch floppy and a drive. And you'd walk up, and a floppy was just about the size of a song. It was 1 and 1 half megabytes, which is just about any coded song at good quality. You'd plug it in, it would play the song, and then spit it out. And you'd just go through a pack of 3 and 1 half discs and plug it in play. So it was like a digital record, in fact, is what it was. So this was my digital tape recorder. And it brought me back memories of my old tape recorder. And I got Bell Labs to buy a CD burner. It was one of the very, very, very, very first. It cost $20,000. Those are different dollars than you think now. Because MSSP is a mistake. It's supposed to be SMP, sorry. It's a multiprocessor. Bell Labs had ordered a multiprocessor for evaluation. And it sat around for months and months, and nobody really liked it. So it was going to be evaluated no. And we tried to find the manufacturer, and they went out of business, and they didn't have a phone number, they didn't have a person. So it just sat there for years, actually. And the multiprocessor was fast enough to do the decoding. So I bought two really, really, really fast CD rippers, and put them on this thing and ask everybody in all of Bell Labs to come in and just stick their CD collection into the hole. And I sucked it up, put it down, and then throw away the dupes and things like that, then tried to find out what they were, tried to label them. And then everybody started getting worried about copyright, little things like that. So I took it to the legal department, and they passed on it. I was very surprised, but I wasn't going to question them. There are several exception clauses in the copyright, and one of them is research. And this is a research instituting, doing research on audio so I can steal CDs, right? Next step, next step, OK. This is, remember, 75 years. There was a magazine called Billboard, and it was the audio record music rag. I don't know, it was everything. And in it, every week, it was weekly, every week was a top 100 most popular songs of the day, right? And anyway, I got looking at it, and I said, well, I have a bunch of 57 songs from that first CD I showed you. Let's see what we can do. I'll take all of 1957, see if I can get it, right? So I went off, and this is a page out of a 1957 week of Billboard magazine. And it goes on, it's got top 100, but it's multiple pages, things like that. So anyway, I sat down and I typed in, I got a hold of copies of 50 Billboard magazines, and I'm just up here to prove that I'm crazy. And typed them all in and made beautiful metadata of the records. All the artists were spelled correctly, all the titles were spelled correctly, all of that. I shouldn't say this, but I had good secretary help with that day. And so I got the year, and I called it the list, right? Teal, the list. I want to just abbreviate here. So this is the start of the list. It was all of the 1957 CDs, but it was no more. I didn't expand it, right? I mean, it was a finite thing. And then I went out to try to find those finite records songs. And at that time, all these CDs were very popular, so you could pull them in and you'd get, there'd be like 15 songs on it, and there'd be seven of them on the list. So that was a hit. But then you'd buy 45 records, which were still very popular, and digitize them and put them on. But then you'd have to buy one CD to get one song. So it went on. It was kind of, OK, I would end up with songs in the list and then other songs I got because they came that weren't in the list. And that was kind of a trash list that I kept. So I had two things of songs, one with very good metadata and one just a pile, just a heap, which went with the songs that I gripped from the. Anyway, after I got to where it was hard to get us another song, I counted them up and I got about 80% of the 1957 songs. And I called that a success. And I was done. I had 1957 all in pack, all implemented various implementations of this box. And I was done, right? But it needed input. Just turning it on and running it was kind of stupid. So anyway, I was in Tom's River, New Jersey at some street fair and there were two AMI rebotes. I won't quite tell you what they are yet. For sale, no guarantees, buy them, walk away, that's it. I bought one and, of course, I should have bought two. And this is it. It's the thing that goes on 50 diners, jute box, tables, and you push the buttons and it goes in. And it's got 10 pages. Those are the, I'm supposed to have a laser here. Right there, there's 10 pages. And you sweep those from left to right and it changes the titles. And then it's got 20 buttons. So it's 200 songs. And to be honest, AMI killed Whirlitzer, which was the one button one song jute box because they could put 200, you know, anyway, that's a side. So I had this on my pack, you know, DSP machine to audio with, you know, and I had 200 songs, which of course were all 1957 songs. And it was cool, you know, it was a real, you know, it was cool, it was good, you know. All right, then a very, very bad thing happened. I named Joel Whitburn in a company called Audio Research, or something like that, made a book of billboards, same book I used, same magazine I use, of all of them from 1955, and that's a typo, that's 59. All right, and so I just couldn't pass it up. That's his thing, right? So I sat down and typed in 55, 56, 58, 59, okay, right? And put them in, added them to the list, started looking for CDs again. And anyway, when that list was done, I got about 75%. I didn't work as hard. I didn't work as hard as I could have. So I called it a success, but then the son of a bitch published four more full decade books. Now, to be true, this is still in the 80s, and the last decade book didn't come out until 1990, you know, 1991, or 2000, right, after that decade. So I actually had to wait for the last one, but he published his first three instantly, and he wiped me out for weeks and weeks and weeks. So far, everything is in PAC, and I'd been testing music and songs and things. PAC was being sold or attempted to be licensed by AT&T at this point. And it was going into these audio testing suites, right, where they have these guys called golden ears, as their name. And what they are are people who can hear everything about a song, all the nuances about a song. And they come in, they say, that's good, that's bad, that's good, that's bad. And the winner wins something, right? And PAC was selected for three projects, right? One was FM Digital Audio, which never came about. They never passed the FAA thing. It was just never made a thing, made into a thing, although we were ready for it. The second was Cirrus XM, a download of, you know, probably 90% of you know it. Anyway, it's still using PAC, probably the only installation of PAC. I think it has. There's no way you can look and ask them what your protocol is. But it started off as PAC. And anyway, so I had four songs that I considered to be an expert on these four songs, that I heard all the, I knew these songs. I mean, I knew them because I played them probably 50 times a day in debugging everything, all of this, everything we've done now. And this is how many years? We started out about 85. So this is 15, you know, this is a lot of years. And I knew these four songs. And I could tell when I couldn't differentiate the song from the original. And when that happened, I would call it transparent. That's a technical term, which is my subjective version of I can't tell the difference. And for PAC to go transparent on my four songs, it came in at about a 10 to 1 compression ratio versus, you know, full stereo PCM. Now, MP3 came along earlier than this, maybe 10 years earlier. And I'd been testing it for transparency, way up to there. And I was giving it all the bandwidth that wanted or could use or even gave it more than that. And it never went transparent. And in fact, at some point, if you give it more bandwidth, it would get worse. And until 2000, when the layman coder came out, and I could make it go transparent at 6 to 1, much worse than PAC, but much more portable than PAC, right? You could get MP3 decoders everywhere. And I'd been trying for 10 years to try to get an MP3 encoder good enough to re-encode my list. And then lame came along, and lo and behold, it happened. All right. Anyway, so I don't know why. I really don't know why. I had 1955 to 1999. And I decided I was going to expand it from 1900 to 1999. And you start getting into some troubles there. Before 1925, sound reproduction was acoustical. You scrape things, and you heard. And after that, it was electronic. 25 was, you know. And so I started collecting records, not record. That's a bad word, because it has so many synonyms, listings of things that you could relate to popularity, because there was no billboard back there. And I collected sheet music and piano rolls and early record sales. And then I started, and I made the list. I actually made the list. It was a bunch of dubious splice fitting to get this thing into one whole list that kind of looked good. I won't defend it. But it weighed heavily on a whole bunch of fishy application of Ziff's Law, if you know Ziff's Law. So now my TL is 1990, 1900. It's 20th century, and my TL is the 20th century. Didn't go down. OK, there we are. So I happened to be in another walking on the street. I was in a pinball arcade in Dixon, California. Again, no guarantees. I found, for sale, take it, no guarantees, a 1952 Willard's Jurt Box. It's a beautiful one. It's about this tall. It's just magnificent. And there was an internet at that time, so I could get schematics. And I knew what was inside. I could find out what's inside. The previous one, the AMI remote, I got nothing. Absolutely nothing. I don't think I need them anymore. So this is it. The control panel, which is the lower half of the transparent part, is gone there. And the reason is because after a very careful look, I didn't think I could reproduce it with its original control panel. So I took the control panel out. This is the control panel up on stilts. And I put in all the rough equivalent of iPad displays. The left one has a display. The right one's a vacant. And it's got all the things. Now, my first attempt at this Jute Box is to push a button, play a record, and have the record mechanism come up. First off, you have to put a quarter in. I wasn't going to let them do for free, right, Jed? And the record would come up. And it had a stack of 78 records that came with it. And it'd pull one of the records out at random, bring it up, connect it to the thing, ignore all of that, and then play an MP3 record that corresponded to it. And so anyway, now I had 24 buttons and a computer for this thing. And I decided to expand the list. And so I put in the list as best as I could. I put in Rhythm & Blues, 42 to 99, and Country & Western, 42 to 99. And at this point, the list was 60,000 songs. And these were, I mean, this is very different than saying, I got 60,000 songs on my iPad. This is a list of specific 60,000 songs that I attempted to get one at a time. And with the help of the internet now, again, the internet saved me. I got 99% of them. They're there. All right, now, next chapter. In 2020, a friend of mine's mother died. Now, anyway, when the friend was going through what you do when somebody dies, they found a garage. And in the garage, they found a player piano. And I said, I'll take it. Then nobody wanted it. It didn't work. I said, I'd take it to see how it is. It's another input output device, right? And it was just in too bad a shape. It was horrible. It was just impossible to restore it or do anything with it. So I abandoned that. But in my pushed by this episode, I found a company in Tennessee that you go on their website and they have a bunch of old player pianos. I mean, oh, well, they have a bunch of old pianos, player or not. They have grand pianos and uprights and all sorts of pianos. And you can buy them restored, except they're not restored. You buy them and then they will restore them. And at the end, it's very nice. It's like a no-lose game. You put in like a quarter of the money, of the full money, and then they'll run off and restore it. And then you'll examine it or like it or what not like it, whatever it is. And then you can do one of three things after the restoration is done. You can say, no, get your money back and walk away. You can say, no, but I want you to do another one. I want to do a different one. And then you could take your money and move it to the different one. And then they would start restoring that one, take another year doing that. Or you can buy your piano with three quarters of the money that was missing. Anyway, so that's this company. It's great. And so I bought a player piano. This is the original 19, it's a 1908 Berrywood. I'm sure that means nothing to everybody. But it has, it's backlit. These are windows like church windows. And they're backlit. It's very nice. It's a pretty piano. And in those days, no plastic, no, you know, it's wood, it's beautiful. Anyway, it was restored. And we accepted it. And it had a player automation in it, right, modern. It would pound on the keys and play notes. But it also came with an absolutely horrid, terrible app on your iPhone where you had to pull out your iPhone and go scribbling through it. I want to play that one, you know. And it was like on your screen, you'd get like a selection of two. And then you'd roll it up and you'd get two more. And you'd roll it up and you'd get two more. It was just, it was beyond horrid. So I cut it out and went behind it with a MIDI interface run by a Raspberry Pi that took UDP. You know, UDP is just to send a message and hope somebody gets that kind of thing. But it's in my house at one gig around my house. And someone who wanted to play the player piano would take the MIDI, which is a file system format for playing music and decode it, most of its timing, you know, to get all the timing. And then it would send one of two messages to the player piano, put a key down, put a key up. So the player piano was just a dummy to bang on keys. And then the actual playing was done in some peripheral somewhere like in a laptop or something like that. And then I needed some MIDI files to play, right? So I searched the web for all MIDI files because there's no such thing as a piano-only MIDI file. And got maybe 60 to 100,000 files. I never counted them, so I don't know. And these are downloadable MIDI, but MIDI files from the web that you just get if you do a search. And then I went through and I wrote a MIDI interpreter, which I had already because that to play the MIDI, and found all of those that are mostly, that are all piano because you don't want orchestration and drums and stuff like that in there. And that turned out to be 13,000 files. Most of them are horrible. And a lot of them are duplicates because the whole MIDI community just steals from each other. You know what I'm saying? So each one of them wants the biggest pile of MIDI files they can. And as soon as they find somebody with a bigger pile, they bring it over. And so most of it's not most, it's duplicates. The popular things had like 20 duplicates, but the unique things are unique. So anyway, I selected a few and put them in the list. Okay, the list. And okay, this is a player of the result. And I'm not sure how to do this. This is my jukebox. There it is with the two iPads in it, with the six, you know, 24 buttons. This is my wife. That's voice search it says. Tiger Lilies. And she said Tiger Lilies. And over here comes up Tiger Lilies. She plays Tiger Lilies. You know this one? So that's part of the list. That's called stop. Maple rag. The blue ones are pianos. Okay, that's my talk on my 70 year project, 75 year project. I have no idea what time it is. And if I have time, I'll take questions. Anybody wants to talk about anything? We do have some time for a couple of questions. One did come over the internet, and Hannah is somewhere around here with a mic. But the question is, given your love of music, or at least of technology around music, what music do you listen to when you are working on this technology? When you're coding or designing or building player pianos? I don't know, at times when I was building the older parts of the list, I loved 30s. I mean, there was some strange, strange stuff in the 30s. There was no sensors in the 30s. I don't know if you knew that. And they talked about what they wanted to talk about. Awesome. Somewhere in the back, I see a, okay. During the UNIX was the BSD people, even though they, I think, used some BSD components in the research UNIX. I'm not sure I understood that, but I guess. So what was the UNIX, what was Bell Labs doing with the DSP? Well, there was some conflict with AT&T and Berkeley. No, AT&T and Bell Labs were the same people. Yeah, I mean, like at BSD and Bell Labs. BSD, oh, they had nothing. This is way before BSD ever existed. And the DSP was digital single processor. It's a chip that is hard to, you know, I think it's only features, it's hard to program. I'm sorry if my D came out as a B. Just keep your hands raised ahead of the camera. So basically, considering that you have this library that is so large, and you're still trying to get more for your player piano, what do you actually make it an open bounty for those that want to make a D for it? There's no hope of that. It's so copyrighted, it's pathetic. I'm sorry, that's so far over my head. I can't even talk about it. Hello, how much code do you write today? And if so, what are you working on? I write a lot of code all the time. I've worked on a more usable RAID controller for a disk. And I've worked on, I collect video, I work on, I can't, you know, you got me at a wrong time. But I work on chess, computer chess, and chess in games. And I have a farm of Raspberry Pi's that I use to do parallel computation of chess in games, things like that. Right now, I have two houses, and I'm getting rid of one. And in it, I have all of the CDs that I bought for what we just talked about. And it is five double cabinets, about this tall, with about seven or eight drawers. And the drawers are packed without CD cases, CDs. And on my way back to where I live, I have a project that I've got 28 fast CD readers. And some of the old Raspberry Pi's that I have from chess stuff. And I'm going to just put them in, probably 10,000 CDs, put them in, and get them in a digital form, and then just take the whole pack of CDs down to some music shop and say, here. I think one question was asked there, and I saw on the Twitter stream was, what language were you using when you developed all this? I'll go. I'll go, so there you go. Cool. Ken, I've done a lot of research with early GCC developers, and nobody remembers what compiler the first versions of GCC were bootstrapped on. It could well have been one of yours. Can you tell us right now whether you have a backdoor into every copy of GCC and Linux still today? I assume you're talking about some paper I wrote a long time ago. No, I have no backdoor. That was very carefully controlled because there was some spectacular fumbles before that. I got it released, or I got somebody to steal it from me in a very controlled sense and then tracked whether they found it or not, and they didn't. But they broke it because of some technical effect, but they didn't find out what it was and then track it. Now, so it never got out, if that's what you're talking about. I hate to say this in front of a big audience, but we're all friends here, and they all signed NDAs on the way in, so it's fine. The one question I've been waiting since I wrote that paper is, you got the code. Never been asked. I still have the code. Any questions? So it doesn't come. You might want to keep those CDs for, I don't know, copyright? You sell CDs. That's all gone under copyright. They've lost that. I think this is awesome. Well, I collected a research. It's OK. I'm wondering if you have a velocity information on the keys when you push down on the MIDI information. Is that included in the piano scrolls that you have? There's no scrolls. It's played by MIDI. It's played by UDP. And I don't understand. Maybe I don't either. But is there a velocity information in the MIDI? Yes, yes. The down key has a velocity impact volume. It's called, I think. It has a which key. And it has a down or an up opcode. That's MIDI. It's called an element, a MIDI element. And there are only if there's like only two MIDI elements, but then there's 1,000 things like, you know, there's pedals playing with the pedals. And MIDI gets very, very contorted down towards the miniscule. But upfront, it's down, up. Hey, Ken, what's your operating system of choice today? I have for most of my life, because I was sort of born into it, run Apple. All right, now recently, meaning within the last five years, I've become more and more and more depressed. And what Apple is doing to something that should allow you to work is just atrocious. But they are taking a lot of space and time to do it. So it's OK. And I have come within the last month or two to say, even though I've invested, you know, zillion years in Apple, I'm throwing it away. And I'm going to Linux, to Raspbian in particular. Anyway, I'm half-transitioned now. For your music collection, would you consider donating it to a library like the Internet Archive? They also have tooling for, like, helping to repeat it. That question was answered by copyright. No, there's just, oh, you mean the actual disks? Yeah, the physical media. I don't know, maybe. I'm not sure anybody, you know, I want somebody who can use them, who wants them. Cool. I think a lot of people may have come here and thought, OK, Ken's going to talk a lot about what you're most known for and these sorts of things. And I'd say that's probably the common thing to do. But, you know, maybe it's too personal of a question. But in deciding what you're going to give to this group, is there any particular idea behind it? No, no, I was actually. Or is it just to say how quirky you were? I apologize. This is not on your, you know, your menu. I've worried about giving this talk. But in the past, I have very often given talks that have nothing to do with where I've given them. I don't know, have anybody heard the alligator talk? Or the, oh, what's another one? There was one other. Yeah, I'm sorry, Matt. So as the PAC, the encoding format that Rob Pike was trying to get publicized that some lawyer called him up and went no way. Yes, on a release of a Plan 9 CD, in the Plan 9 source was a PAC encoder and a PAC decoder. And it took up this, the Plan 9 took up about 100 meg of this 300 megs, 3, 400 meg CD. And I filled the rest with music. The rest was just filming. And it came down stomped on by the lawyers so bad that it, you know, so the answer is yes and no. So yeah, when building out the TL, I assume you did a lot of stuff with music metadata. Have you used music brains or contributed to music brains at all? No, I have used, yes, I've used it recently for the R&B in the country in Western. But my stuff predated that stuff. And I used it for that, and I'm very grateful. But the answer is no before that, outside of that. So now that you have the player piano, you have plans for a next edition. No, no, no. Awesome. I think we've got. OK, so how many Raspberry Pies do you have currently then? I'm sorry, it's too loud. I couldn't hear you. It's too loud. How many Raspberry Pies do you have currently? I have my bulk Raspberry Pies in a little stack of four with, you know, they're built up with little pieces of metal. And I have 12 of those stacks of four, all the same. They're Raspberry Pi 4s 4G, right? So that's the bulk of my bulk stuff. But then I've been collecting Raspberry Pies for a long time. The AMI is run by a Raspberry Pi A. It was a packet first. But then as soon as Raspberry Pi came in, the very first Raspberry Pi runs the AMI remote, the little jukebox. And I've got B's and C's. So I probably have not counting the bulk. I probably have 20 around the house doing things like UDP to the player piano, UDP from the network, you know, committee to the player piano, things like that. So I probably have, I don't know, is the answer, I guess. A lot, a lot, a lot. I use them for everything. Ken, I want to thank you very much for this talk. I mean, you didn't know this about me, but I have two player pianos, a Nickelodeon, a Wilcox and White Symphony player redorgan that also uses a roll, and a push-up, which is a unique instrument that I'll write you about. But I will also point out that since 1994, I've been giving a talk called, Why is Linux Like a Player Piano? It's very good. It's a lot of synergy in Linux and player pianos, and pianos in general, and I'd be happy to give that talk next year. Thank you. Thanks, Mad Dog. I think that's the first CFP submission for 2023, or 2024, that is. So I think we're up at four o'clock. I know there's lots of hands and lots of questions, and I want to make sure that we're being respectful of Ken Thompson's time. So with that, thank you again for joining us. Before you run off, I've got a few things for you. Oh, great. One of our traditions is to welcome our keynote speakers to the Scale family by our team with a jersey. Oh, cool. Got your name on the back, and the scale will go on the front. Where's my number? Huh? Thank you, thank you very much. So we hope you will wear that, and I don't know how to connect that one. Our keynote speaker this morning was doing electronic controlled clothing, and she said she was gonna find a way to make their jersey work in that. I don't know how to connect that jersey to your player piano just yet, but you are a creative, you seem to be very creative, and we'll probably come up with a way to do so. So with that, thank you very much for joining us for Scale 20.