Embedded - 60: Fun Things You Can Make out of Beagles

Episode Date: July 23, 2014

Jason Kridner (@Jadon) joined us to talk about the BeagleBone Black... and other things. Some good books for Beagle : Bad to the Bone: Crafting Electronics Systems with Beaglebone and BeagleBone Black(co-authored by Jason) Getting Started with BeagleBone: Linux-Powered Electronic Projects With Python and JavaScript Programming the BeagleBone Black: Getting Started with JavaScript and BoneScript More comprehensive list of BeagleBone resources BotSpeak - A programming language for internet endpoints To contact Jason about ordering a bunch of units for your OEM use, see his contact info on BeagleBoard.org's About page.

Welcome to Embedded, the show for people who love gadgets. I'm Elysia White, alongside Christopher White. Hello. And Jason Kreidner of the BeagleBoard.org Foundation and Texas Instruments. Hi, Jason. Thank you for joining us. Hi. Nice to get a chance to talk to you. Can you tell us about yourself? Sure. I'm probably your typical engine nerd, right?

My mom bought a TRS-80 for her business. She was in tax returns and she bought a computer when I was probably about eight and gave me a book on how to program and I learned to program in BASIC. Told me I could do whatever I wanted as long as I didn't open it up or take a hammer to it or otherwise physically damage it. I could do whatever I wanted in software. But pretty much everything else in my world I took apart. That one was just a little too expensive.

But everything else I pretty much took apart and was pretty hit or miss if it ever got back together. I really wasn't really all about getting it back together. I was really all't really about all about getting it back together. I was really all about trying to understand how it works. Once I got that, I never bothered. Then I spent, you know, I wanted to know how the computer worked. And I ended up mowing lawns and spending all my money at Radio Shack. And I'm sure a lot of other kids did. Got the Forrest M. Mims book, Getting Started in Electronics, and built everything out of it. A few times over, probably, and changing things. It was nice just giving an intuition for electronics,

which is pretty much what I was always looking for, is intuitions. Maybe not every last little detail, but just kind of get the feeling for things. I still don't have that. Yeah, I don't think you'd ever get it entirely but I tried I'm still trying I'm still trying to get a feel for a lot of things when I was in school it did not serve me well

I just hated college I loved all the stuff that was available to me this was before the web I got on the internet and just started you know downloading all the software I could and playing with things and but I was more interested like I wrote a an EM wave simulator and spent more time playing with my little simulator and trying to build up intuitions and stuff about you know how this stuff worked and I did actually learning how

to answer the questions on the test so I never never did that all that great. So I've always been really interested in electronics, always knew I wanted to be an engineer, and I've wanted to share that fun with other people, and I think that's kind of who I am. I think your experience matches a lot of people, I mean, what you're saying. I'm nodding my head quietly over here. I'm very familiar, including the inability to tolerate the structured learning process in college after having done so much exploration on your own. Yeah, because it's not what gave you the passion to kind of get into it in the first place. I mean, you need the theory. I actually enjoyed a lot of the math and

a lot of the theory, but learning to take the test and pass the test was just, really took a lot of the fun out of it. And it is a game. As someone who is pretty good at taking tests, and I am easily amused just by tests, I recognize that it's a totally different mindset. I should have spent more time finding the fun in that. It is a special sort of fun. I mean, we're all building on the backs of others that came before us. And those professors spend a lot of time thinking about the right things for you to learn.

And the way to learn them and the organization. I mean, it's sort of their passion to help you be enthusiastic. And it's, you know, we have two math professors in the family. And one of them, it is her calling. I mean, she knows that her students often hate math, and yet she just wants them to love it even a little bit. And it's amazing. out the magic formula to share your excitement and passion about things like that, like math, like engineering, like all these

things that serve you fantastic in life. And I want to know the magic. I've been looking for it. I keep searching. There is no one true path. You'd be disappointed if you found it.

You're right. You're right. We're always creating new problems. That's what Engine Nerds do. Okay, wait a minute. We have totally gotten off track. This is the first time we've gotten off track in the first two minutes. I think it's probably my fault. No, no. I think it's Jason's fault. You are famous for something. You are famous for something. You are famous for the BeagleBoard and your involvement with that. So that was why I wanted to ask.

I have a bunch of questions about the BeagleBoard too. You worked for TI for a long while, and then you did some open source-y things. Let's skip just TI and open source-y things. Skip all of TI Because I've been at TI for 21 years Wow, that's great So what did you start doing there? I guess we shouldn't talk too much about it

But I started doing more hardware stuff So I actually started out in product engineering Characterizing chips and monitoring production stuff. But it was all really just because I wanted to get into DSP. Yes, because that's where the fun stuff is. The fun math.

Yeah, FOIA is awesome. Math does stuff. Yes. And yeah, so I wanted to get into DSP, got in through product engineering. Like I said, my grades weren't super. But co-opping is amazing.

It's like, and once I got into actually working and doing engineering stuff, I was so happy. I just was so happy I was doing it. I couldn't get out of school fast enough. But, yeah, so started out doing project engineering, then started doing board design stuff, FPGAs, ASICs. You know, took the FPGA work and, you know, translated that into ASICs. So I was doing VHDL and board design. And then, you know, built different analog, like interfaces to analog front ends, like for some of the first ADSL modems.

So a lot of the original hardware work and the ASIC work was going into modems and to game cards. And then kind of switched to software actually around MP3 players, and actually we called them digital audio players, because we were the first one to do a really programmable one with DSPs. So we were the first one to launch a reprogrammable digital audio player. I'm trying to remember. I had one of those boards. Was it the C55 family? We actually did the C54s. The C54, oh, that makes sense, yeah. Yeah, so you actually played around with the C54s. That's cool.

I worked with Audible for a little while, porting their engines to different systems. And so I got one of those hideously expensive dev kits. But, of course, it was always something you borrowed for like six months at a time. Yeah. Well, there was this thing called a DSK or a DSP starter kit.

Keith Larson was the guy that made that. And that was actually a bit of inspiration for the Beagle board. Because there's a whole BBS community, right? This was before internet and everybody had the web stuff. So it was a BBS based community. I'm dating myself.

Anyway, the, there was a cheap, uh, C, C five X. It wasn't C five four. It was before the five fours.

Okay. I didn't have those. I had one that when I got it, they said if you don't return it to us, it will cost you like $15,000. And the best part about it was that it came with a really, really cheap pair of earbuds. That was the model because you sold 50 of them, right? Or whatever it is, and that's it, right? So the idea that you're going to make a dev kit and sell 10,000 of them was not normal.

No, the dev kit's purpose was so that that chip would be sold. And so the DevKit was really expensive. And it was only given or loaned to a small number of people who could then influence how many of those chips were bought. Yeah, absolutely. Because nobody ever actually paid that huge price. No, of course not. If you were actually doing something real with the chips and making a product,

the field app engineer always ended up giving it to you anyway. Yeah, and it was like mine where it wasn't given to me. It was just loaned to me for an indefinite, unspecified amount of time. Did they ever collect it back from you? No, they did not. And it's just not that it wasn't worth anything and that they didn't want to not waste that money and give it away. It's just that by the

time you were done with it, there was something new designed and they weren't even bothering with pushing that kit anymore, right? It was like, time to move on to the next thing and so they'll forget about it. It's always on loan, but it never comes back and based on christopher's hand gestures i guess my next next question should be that is so different from what the beagle board is like today

with its i mean as far as dev kits go and you it has to be a great processor i mean people get the beagle boards and they use them and then they design their own boards with them. So why did these things ever cost $15,000? That's a hard question, but that was the model, right? It worked. This is better. Yeah, it is so much better.

It's certainly more democratic. You don't have to be a big company with a contact to a field engineer or a sales engineer and say, oh, we're thinking about doing this project. We're going to sell a million of your chips or buy a million of your chips for our product. Can we have one of these big boards? It's not easy because that's still the model.

You say that it used to be that way. No, it's still that way. I mean, most of the dev kits are still insanely expensive. Well, it's so strange. I worked with the ST ARM Core recently, and they have the $15, $30 board set that you can buy on DigiKey. But if you want to get just a few more pins or you want to do this little thing, their dev kits go up to $500 and then up to $3,000 really quickly.

And it's not like the $3,000 one does more than about 15% more than the $30 one was. You know as well as anybody of the economies of scale, right? Exactly, and that is what the problem is. When we did the first BeagleBoard, at the time I was the chief technologist for portable audio and video business. So that's an extension of MP3 work. It was. It was. It was coming out of the MP3, but it was mostly video players and stuff at the time.

You know, like the Arcos jukeboxes. That was probably one of the more notable ones. I remember all the players at the time were starting to add video. And so, yeah, I was the... so that's what i was doing right so i was we were looking at the what was coming out of the the kind of the cell phone chips so you know ti was was big into making cell phone chips and those you know economies of scale, you're selling by the millions. We were actually making a million ARM processors a day. Wow, that's kind of cool.

It's a nice number. A day, yeah. So lots of economies of scale going for you. And it made sense to start using some of that for the media players. And the economy for the cell phones was changing, right? If you look at it now, there's very few players and most of them are insourcing for their, you know, Apple and Samsung, right? Those are your big guys, right?

Everybody has an iPhone or a Galaxy. And even the media players that exist are usually stripped-down versions of their phones. It used to go the other way. A phone that had a media player was like a jumped-up. Junk, yeah. And now you actually use the cell phone service to get your music. Yeah.

So you always have your entire music collection at your fingertips. Even if you only got 16 gig in your phone, you're now using cloud services and accessing your 50 gig music collection from home. Maybe not everybody has that. Unless you're trapped in Carmel with your in-laws with really, really crummy internet service after having broken your phone. They don't listen to this podcast, do they?

Actually, my father-in-law does. So now you're just trying to guilt him into upgrading their internet service, right? Yeah, so I do have some problems with phone and music. Chris doesn't put music on his phone, but I always have just a ton of music on my phone. I'm sorry, we are still not even close. It's like a road we keep crossing. Okay, okay. Okay, MP3 players. What is on track for you?

You tell me. I'll stay on track. I can do that. No, I don't really mind. So you were doing MP3 players, and the boards got bigger, and we started using more of these chips and a lot of them. And how did you flip over into Open Platform Evangelist, which I think is what LinkedIn said your title was?

So it was all about kind of seeing what the benefits were of particularly Linux, right? I think that Linux is just the community around Linux is incredible. And I wanted to do something that kind of scratched my own itch. I wanted something that could kind of help give people that passion and interest in electronics and stuff that I had. And it also made good business sense.

So it all just kind of crossed over there. And getting to the Linux community, trying to do something that appealed to them and something that they could use to scratch their own itches, right? To serve their own interests, to be able to advance the state of Linux on ARM because it was kind of at that point where these phones were getting and these embedded processors were getting powerful enough to run full-on operating systems and enabling to use high-level languages and do development directly on the platform

and just all these things that were, what an embedded system was, was starting to get really blurry. It's, you know, they've always been, you know, microcontrollers are computers, right? They're this, you know, they've always been, you know, microcontrollers are computers, right? They're this, you know, it's programmable architecture, runs instructions, but this is the point where, wow, we're actually running desktop systems here.

You can actually run window managers and you have all these operating system abstractions. So around 2005, I remember TI came out with this OMAP system. And I don't think it was called the BeagleBoard then, but it ran Linux and it was a single board computer of unusual processing power. I mean, is that a direct? You're talking about the OSK, osk the the omap starter kit

yes and and it wasn't very expensive but it was it wasn't but it was it was yeah it was it was cheaper actually gerald coley the guy that designed the original beagleborn um did that um before me right that was actually, he developed the OMAP starter kit. And it was fairly affordable. So he made a whole lot of sense for being the guy to do the BeagleBoard. So what was the impetus for the BeagleBoard? It certainly sounds like a lot of things came together. There was the Linux, there were good processors.

But who sat down and said, you know what we need? We need something really cheap and really good. Yeah, so Gerald and I had lunch in Houston one day. And he was trying to do all-map broad market stuff. And I was coming from the Port Bolivian video, but people told him that I had interesting ideas. And so he flew down from Dallas to Houston to have lunch, and that was where we birthed the idea. It was... I told him that what we wanted to do was to build something that appealed to Linux developers.

That was... The maker thing came later, right? We had no idea what a maker was. This was not really a thing. Arduino was not really a thing at the time. It got to be big before the BeagleBoard got to be big.

But when we were coming up with this, nobody had ever talked about open hardware. There were certainly some small Linux systems out there, but the capability of the chip was so much higher than everything else. And we wanted it to be affordable and appealing to the Linux hacker. And that was different. Usually you're marketing to people that you know are going to embed this into a billion different products, not somebody that just wants to have fun with technology.

And then everybody wanted to embed it into their products. Jump a little bit forward, yeah. And BeagleBoard was meant to be something to hack with, and it was a reference platform, right? Not a commercial off-the-shelf board. But now people drop it into products everywhere. Yeah, I mean, I usually do consumer things, but I have at times worked on systems that were going to sell less than 1,000 units.

And they were expensive, and then there were service costs around it. But gunshot location, I think it's over 1,000 units now. But it wasn't a million a day. It wasn't a million a year. It was maybe we'll get to 50,000 over the lifetime, and that'll be awesome. And it's different. The expense of designing your own board and your own system, and then you have to do all of this software laundry-like chores,

like how do you update these things and how do you make sure it's all consistent and how do you have a good error logging subsystem and yet now now i just go buy a big old bone black because the nre is so expensive to get all of that going that you might as well just buy something and drop it in. We're definitely at a point where we kind of have to share a lot of our technology in order to get the right economies of scale. We can't go build everything on our own, right? You got to build on the backs of others.

And Beagle allows us to do a lot of that because it meets the needs of so many different people. We all get to share in it. Did you expect that to happen? Yeah. What I didn't expect was the it's going to come up, but what I didn't expect was the Raspberry Pi.

It was a lot of some of the same thinking. But, I mean, those guys are smart. I mean, they did the job of putting it into words. The power of this thinking, right? The power of giving people a general purpose computing platform that can run Linux and is cheap. So let me get the history right,

just to make sure I understand. It was BeagleBoard was first, and then Raspberry Pi, and then BeagleBone Black. Yeah, but there was a BeagleBone before the Raspberry Pi. So we, in all of this, right, so there was the original BeagleBoard,

then the BeagleBoard XM, which just kind of, the XM just added more USB ports, went from one USB host port to four. That sounds familiar. Speaking of Raspberry Pi. Right, it's kind of the, and so it went from 720 megahertz to a gigahertz. But the whole time, the original BeagleBoard was really targeted at Linux hackers, right?

And it looked more like a desktop computer. And it was for people who were experienced. It wasn't necessarily targeted towards newbies. Well, because targeting hackers doesn't always mean make it easy. It's much more about making it possible. And smart people will bring solutions. My vision was always, though, that if we engage this Linux community, they'll help us, you know, put together something because they have a lot of the same passions. They want people to learn about the technology

they're creating, right? I mean, Linux is a community of sharing and that, you know, eventually it would get to be this educational platform and that so that was that was always the the vision there um what raspberry pi did was they spelled it out right i mean they just um they did a great job of it too i mean in the last three months four months i have gotten to play with my first BeagleBone and my first Raspberry Pi. Because normally I work at a much lower level. I don't deal with these whole operating system things.

But with the BeagleBone Black, which I borrowed from a listener, thank you, Philip, my first impulse was to recompile the kernel because that's what I wanted to learn how to do and it was just, let's go see how to recompile this kernel. But for Py, I plugged it in, I plugged it into my TV, I played Scratch with it, played with the Scratch language,

I goofed off with Python, I dorked around with the camera. It was, the Big Oone Black was utilitarian and functional, and it was going to help my career, and the Pi was just to play with. What was the difference between them that led you to do those different things? I don't know. Because they're basically, I mean,

they're both Linux platforms, right? Beagle's just faster and more capable. Yes, of course, the Beagle. It seems like a more serious device. You have to treat it more seriously? I don't know. Well, okay, some of it was the web page. You know, the BeagleBone, it tells you how to get into the depths pretty quickly.

And the Raspberry Pi tells, the introduction stuff was, here's how you play with it. Okay. And then there are books. There aren't as many BeagleBone books, although I guess I shouldn't. There's about six out there right now.

In the O'Reilly Safari, their book library thing that I have, there are a lot of Raspberry books and I didn't see a lot of BeagleBone ones. But you wrote one. You wrote Bad to the Bone, Crafting Electronic Systems with BeagleBone and BeagleBone Black. That's a very long title you have. Yeah, my apologies there. So that was with Stephen Barrett.

And Stephen Barrett did all the heavy lifting. He's a professor out of the University of Wyoming who's written a lot of books on using microcontrollers. And it's published through Morgan and Claypool. And they have a really fantastic system for universities. So a lot of universities just subscribe to Morgan and Claypool. And they have a really fantastic system for universities. So a lot of universities just, they subscribe to Morgan and Claypool and they get access to everything in their catalog. Sort of like me with O'Reilly. So yeah, I understand.

Yeah. So yeah, you've got Safari. So this is kind of similar to that. I think Safari is largely good at engineering professionals and working in Claypool. I hope I'm not speaking too much out of turn, but I think it's more focused on education. So educators. Well, if you write another book, if you are an O'Reilly author, you get access to Safari for free. So that's almost worth

all the lost weekends. You should write another book on Beagle. That's what I think. I see. Get rid of all that stuffy impression that this is all for business and not for fun.

Beagle is definitely about fun. I have fun with it all the time. What should people do with their Beagle boards when they unbox them? There's the webpage which tells you a little bit, but what is the most fun thing to do first? Well, fun thing, the hello world for electronics is blink an LED.

I think you'd be hard pressed to find a platform that you can blink the LED faster in than a BeagleBone Black I guess I did I find blinking and LED boring that's because I've been doing that go ahead, what were we going to do? I think one of the things I did do on the Beagle board

before I started recompiling the kernel was look at the I squared C because I was surprised at how easy it was to just plug in something with I squared C and it found the address and it read my accelerometer. It was pretty amusing. Yeah.

It's actually really easy to use. I think most people find the Beagles actually easier to use and easier to get started with than a Raspberry Pi. And a lot of that is because it comes with software already on it, so the first thing you don't have to go and figure out how to flash

an SD card. It's because we've been around longer, we've had more experience at solving a lot of these first challenges with users. So we started embedding Flash on the board so that it wasn't the first experience. How do I program Linux on this? We actually shipped with the working Debian Linux distribution on it. That first experience, the blinking LED thing, I mock it a little bit, but for somebody who hasn't done it, it's a very visceral experience.

It's like I did something that actually made hardware change, right? If you understand what's going on, and sometimes you just shouldn't make it too easy. I think in some ways it's bad that we make it too easy because if you go through the process and understand that you're setting a register and you're making these things change and a register and a processor actually equates to some physical thing that's going out on a pin and that's very empowering and exciting for people who haven't had that before. It is. Yeah, it's the difference between I made this happen and something made this happen.

Yeah. And my software can touch the world? Yeah. I'm constantly, that is what makes me an embedded systems programmer is I'm constantly fascinated by that. But tippy-typing away can actually change some physical thing over there.

And we make it really really easy to do that and and i'm honestly i kick myself in some ways it's tricky to figure out what you need to make easy and what you should believe being hard because some things should be difficult so that you have to go through the process of really understanding it in order to to get the the enjoyment out of it. And that satisfaction that you've taught yourself something. If you just go to the, literally when you plug in the board into your computer, you point your web browser to the, well, there's a, it shows up as a flash disk, a flash drive, a flash drive.

And there's a start.htm. You double-click it, and it brings up an interface. And it starts telling you how to do basic I.O. stuff in JavaScript and actually editing it there in the web page. Yeah, there was the script on the web page. That was pretty neat. Yeah. So that's some, you know, because a lot of what

I'd like to do is try to attract all these folks who are doing web development and they can create web pages and for some reason they're very reluctant to mess with hardware at all.

I came from software so I have great sympathy. A lot of times the software engineers are told if you touch something, it electrocutes you. Yeah. And I found that that isn't true. Because they broke other people's stuff. Yeah. One time.

And that's the beauty of the BeagleBone, BeagleBoard Black, BeagleBone, now I've totally confused myself, the BBB, is that you can accept it's not that expensive. If I drop this out of my drone, it's okay. If I touch it and it shocks itself, yes, I'm going to be sad, but no, it's not going to be my $3,000 computer. Right.

Yeah, well, I think that's a lot of people of our generation's experience with electronics learning is, you know, oh, here's this expensive thing, and, you know, I don't want to have an ESD problem and kill this chip and then ruin it. You know, this is going back to what Jason said at the beginning. What Jason said was about the TRS-80, right? You can take anything apart, just not this. You can do software, but you can't touch the hardware of this TRS-80.

And I had the same thing with an Apple II. It was like, well, you know, don't touch anything. And this floppy disk, you have to be very careful and hold it right here. And, you know, don't breathe wrong. And you can type. So even in initial work experiences, well, here's this prototype. It's, you know, maybe $100,000 worth of equipment.

So sit here with your static thing on your arm and don't look at it wrong. Whereas now you've got such cheap hardware, but it's hard to get over. Oh, I don't want to break it, you know? Right. And the floppy disk was actually a fantastic thing because now with people's hard drives, you don't even want to let kids play with your desktop computer because it has all the family portraits on it. Everything is on there and you're always scared of losing your data because most people don't understand that you have to back up everything to begin with.

I'm amazed at the people that still don't back up their computers. Chris finally has converted me to CrashPlan. Finally. To any backup. I always believed in backup, but now I'm very much off-site backup. You believed in it, you didn't actually practice it. One of the first things I want to teach people is Git. I mean, the version control tool

that's used by the Linux kernel maintainers. It is just amazing technology. You can't lose your work. And it's up there and it's free. And if you put it public, it acts as a resume for you, which is strange and cool and scary and strange. Yeah, Git is great. If that is how you're introducing people to backup, I don't know.

It's not really backup, but it's collaboration. It's not really backup. It's collaboration. It's collaboration. It's workflow. It's history. Most people using computers don't have a notion that they could save history of stuff that they work on forever. And every change they make could be able to be rolled back. And shared with everyone. Think of files as these things that they're either up to date or deleted.

And you can control it, right? I mean, you can keep your private repos, you can push things public, and I don't know. All this technology just needs to be demystified. And I couldn't be more excited about being involved in Beagle and being a part of that demystification. And I love what Pi has done so much to actually increase the Beagle community.

And even though I can get a little bit jealous of all the numbers that they've done, I still think what we're doing is so incredibly important. And I think what we enable is also very different, right? I mean, if you want to play with Scratch, hey, Raspberry Pi, fantastic. You want to actually build a robot? You're going to be a whole lot better off with a BeagleBone. And I'm happy about that, right? So it's a big role. There's lots of room for us to play together. I have heard some Pi fanatic say that the black is selling at a loss

in order to encourage people to use TI chips. Man, that's incorrect. So CircuitCode does get favorable pricing, but CircuitCode pays more for the TI chips than some of the high-volume customers. So they're paying for every chip. I never wanted this to be something that would, when TI would get bored of it,

because as you know, eventually the dev boards, you know, they're not designing in that chip anymore, right? So there's something newer, there's something cooler, there's something sexier, so all the people designing new products aren't going to choose to use that chip anymore. So we wanted a purchase agreement with TI. Okay, yes, I'm a TI employee, but Yeah, you're going back and forth as to who you are. Yeah, well, Beagle works outside of TI. It's a separate entity, and it is not TI.

Does it help TI? Absolutely. So I don't want to make it sound like there's no TI motive for it. But at the same time, it stands in the penitent of TI. And I've personally always tried to make it satisfy those broader goals. And we don't take subsidies because I knew exactly what would happen if that was the path we went down. If what we said is we're just going to take chips on consignment, sell everything at a loss, that's great for the first year, and then it goes away. Yeah. It's not sustainable.

No, and we've been around for six years. We've been around for six years. And people can still get a hold of the original boards. Demand goes down. They don't get built as often, but they still get built. We can still source all the components on them. We've had inquiries. Over the last week, I've had inquiries for both original BeagleBoards and BeagleBoardXMs.

And both of those have build schedules and still get built. And yet there were no BeagleBoard blacks for the two months that I was randomly trying to buy one. There were. That's the crazy thing is that hundreds of boards were still being made every single day. Hundreds. I know, they needed thousands, but there were hundreds of boards being still being made every single day hundreds i know not that needed thousands but

there were hundreds of boards being built every day and um and even during the the time where absolutely nobody can find them and we still get through waves of of nobody being able to find them and and that's because it's um well i mean there's there mean, there's a few reasons. If you want to get a board, the thing that you have to do is, like, Adafruit has this nice thing where they'll send you an email when a board comes back in stock. I watched the boards go up there. You know, usually they get a shipment. When they get a shipment, you know, they get like 500 or 600 boards at a time.

And I would know how many boards they got. And by about, you know, 10 o'clock in the, or so in the morning, when they get under a hundred, you actually see the, the, the number show up on the Adafruit website of how many boards they have left in stock. And then, you know, I would watch it and it would actually go down at the rate of about two to three a minute. So in the next half an hour or so, all the boards would be gone. Yes, yes, exactly. That was my experience. But it wasn't that you couldn't get them.

When you got the email notification, you had to get to the website and place your order. And so there have been boards all the time being made. It wasn't that there was some loss of supply. But this still seems like a logistics foul-up. I mean, you have people who actually want to buy these things. I mean, I bought one from SparkFun and then was on the back order forever, and they got some, and I still didn't get my prices are probably

just too low our prices are too low um and yeah so that's um there's there's a lot of truth to that so um you know it's not subsidized but at the same time um you know the the market demands people sort of a few things all happened at once. It took a while for people to start designing these into products. And once they did, they started sucking up all the supply. And it's fine if they want to embed it in a product. You know, there's a terms of use, and it implies that it's not useful or not to be embedded in our product.

It's not that we have anything against people embedding BeagleBone blocks in a product. It's what it does to the supply that's so awful. We need to get it planned. You need to call up CircuitCo and tell them how many boards that you want to buy and schedule it out and, you know, share the risk in sourcing the components, right? You know, you go and go buy all these parts, right? That puts CircuitCo out of a lot of money to go buy all the components. And it's not like they're making a huge amount of money off it. they're making money. Everybody in the chain is making a little bit of money. It's lower profit margin for the distributors. So like the Raspberry Pi,

what the distributor, their MSRP is their cost of finished goods. So it's a loss leader for every distributor who sells the Raspberry Pi. They did it smart in a lot of ways because there's only two distributors that sell the Raspberry Pi. And then everybody else buys it through those two distributors and then resells it, and that's how it gets everywhere. And that was just a relatively smart business model. I mean, we're much more capitalistic in our approach, and they're kind of more socialistic in our approach and you know they're kind of more um socialistic in their approach and that's not a that's um not entirely a judgment call well no their goal is education

i mean they're very upfront with we we are all about the education and i i what does the raspberry pi compute module have to do with you know i have i was just about to say i Raspberry Pi compute module have to do with education? You know, I was just about to say, I have a compute module sitting in a box on my desk and I haven't opened it yet. And I don't, I am trying, there's a little bit of dissonance here. Yeah, well, and... Oh, and for people who haven't heard that... There's a little bit of it, but I think that we've just been a little bit more honest up front

about, you know about our motives. Sure, and to be fair, they've had their volume problems as well. I remember several periods where you couldn't. And they pretty well solved it. And I think we've got the formula to solve it on Beagle as well, but we need distributors to place orders and to share the risk of sourcing all the components because we can't go and just ramp up production.

We don't know how big the demand is because a lot of the numbers are people embedding it into products. Now, I know we have a very broad-based demand. There's a lot of new people coming and finding it that, know, I need something with these quadrature encoders, right, that I can actually measure the speed of the motors that are turning and something with the 32-bit PRUs, these microcontrollers that are built in. I don't want to, you know, go and buy a Raspberry Pi plus an Arduino plus all this other stuff, you know, when BeagleBone's already got it all. So there are a lot of more people coming in and getting started with BeagleBone now.

But we don't know how many. And that's very tough to take that bet and make that guess. So we're, you know, it's part of just how we've built up. We've built up, you know, slow and methodically along the whole way. But now there is so much demand from people putting into products that there are people coming in and starting to build these exact clones to go and satisfy the OEM demand. Well, you did make it open source.

It's open source, unlike some other boards. I like the Raspberry Pi. I like the BeagleBone. I have a Raspberry Pi on my desk because I could get one. And I will have a BeagleBone on my black as soon as I push add to cart on SparkFun, which currently has a few in stock.

Yay! Awesome, awesome. You can get them on Amazon.com. currently has a few in stock. Yay. Awesome. Awesome. You can get them on Amazon.com. They usually don't run out. Yeah, they come in and out. Special Computing is one. To get off of this topic. All right.

When you were mentioning Git, I actually had a question sort of pop into my head. Is this going to be about rebasing no i have so many questions about it was just uh it led me back to linux and i was wondering during the development of this how how linked in with the the linux community you guys got how much you contributed back um did you have to contribute kernel patches to support various things on your board? Is this the sort of thing where you've done the work to get BeagleBone Black fully supported or BeagleBone, any version fully supported in the kernel

so that you can just build straight out of kernel.org? It's always been a goal. Today, if you have an original BeagleBoard or BeagleBoardXM, it's always been a goal today if you have an original BeagleBoard or BeagleBoard XM you can build straight out of mainline and it works great with BeagleBone and BeagleBone Black

you can build straight out of the mainline but some of the key elements are still a work in progress there's this you know a lot of people heard about the, like,

we get it right on the forefront and the battle lines of Linux, right? So I think that, you know, projects that aren't,

it isn't their goal to be mainline kind of stay out of some of the fray and we've been right in the middle of a lot of it. Like,

device tree. So if you wanted to be an ARM been right in the middle of a lot of it, like device tree. So if you wanted to be an ARM device supported in the main line, Linus got really sick of all the ARM processors, all their variations, just way, way, way too many variations. Not as like every PC looks like a PC looks like a PC. And so we said, well, you know, need to move to this data-driven style

so that instead of having all the variations of software, you're just having variations of data. And that's what DeviceTree is. So we adopted DeviceTree very early. And so people have been learning how to deal with that in the Beagle world. So we're helping to solve some of these problems early, and then everybody gets to benefit from that, from us working with it in the mainline community. But things like the device tree fragments or overlays, where you're adjusting the device tree at runtime. So we're still using a bit of an older kernel

in what we ship with the boards in order to enable some additional features. And that's a lot of work. I mean, that's a lot of thankless work in a lot of ways that people don't realize is happening behind the scenes. You're making a hardware product, but you're also advancing the state of the art of Linux.

And, you know, I know from personal experience, there's a lot of back and forth that goes on there with the kernel maintainers. And it's, yeah, it's, thank you. Well, thank the community because, I mean, we have a very technical and engaged community that's, you know, it's evolved somewhat over time as we become a little bit more maker focused and a little less focused on the Linux kernel devs themselves. And I think

we need to kind of get back a little bit to our roots and pay back some of those guys. But we have some really fantastic people in the community. Robert Nelson, he happens to be an apps engineer for DigiKey, but he's doing all of our kernel maintenance right now for the kernels that we actually ship with the board. And Kun Koi, who's now at Lanaro, used to do a lot of that work, and we built a lot of the success of Beagle off of the work that he did in the past. So we've, you know, our community is just fantastic. And that's definitely the community's goal is to see everything upstream in the Linux mainline. And, you know, it's a lot of work. And I'm happy it's continuing lot of work. And I'm happy.

It's continuing to move forward. Will it ever move forward fast enough for everybody to be thrilled? No. Of course not. Because a lot of the work is built on the backs of volunteers. And I don't know.

They're doing fantastic work. Can't thank them enough. Yay for open source. Yay. So back to my, what do I do with a BeagleBorn fantastic work. Can't thank him enough. Yay for open source. So back to my, what do I do with a BeagleBorn when I unbox it? Is it a good way to learn Linux

and computers in general or is it a good way to learn embedded systems? All of the above. It's good to learn Linux because it's cheap hardware that you can play with and it's good to learn Linux because it's cheap hardware that you can play with and it's outside of your desktops. You're not reformatting

your hard drive and you get to play with Linux. And that's a great technology. Linux is something that everybody should be learning today. It's everywhere. So maybe I don't have to do that sales pitch, but

it's a great way to learn Linux. It's also a great way to learn embedded systems. It's a lot more than just a Linux SBC. The hardware, the chip we used for the BeagleBone, like the original BeagleBoard was built essentially off of a cell phone chip. And for the BeagleBone, it was actually an industrial chip. So something that was used in navigation, point of sale, but also a lot of just machines, right, that build other machines in factories

and communicate between them all. And that means it has some pretty fantastic embedded peripherals like pulse width modulators and quadrature encoders and these real-time microcontrollers. And that's stuff that... That's the fun part. It is. That's the fun part. And a lot of stuff isn't super, super easy to use these days.

It doesn't have necessarily all of the fantastic Linux abstractions and some of those are still coming. Some of those are still a work in progress. It's getting easier and easier all the time. Have you ever heard of Google Summer of Code? Yes, that's where they get students usually and they pay them to work on open source projects with mentorship from people like the BeagleBoard.org Foundation, right? Absolutely.

So we're one of the mentoring organizations for Google Summer of Code. Yeah. Do you have any students assigned? What are they doing? Yeah, we're currently going through six projects. We did six projects last year, and we've got another six. Was it five or six last year? And then we have six this year that are going on right now.

And they're doing some really fantastic things. And a lot of the things, they're going to make a lot of those other new cool features a lot easier to use. And they're doing their best to have, have that, again, that mainline focus, right? So that everybody can build off of this stuff. But one of the projects is an interpreter for this thing called, called BotSpeak. It's a bit like Fermata. I don't know how many of your listeners will be familiar with Fermata, but it's a firmware for Arduinos that allows you to use higher level interfaces for hooking up sensors, right? It's kind of a standard firmware load for hooking up to a bunch of sensors.

And BotSpeak is a little bit like Fermata. You could do the exact same thing with Fermata What was it called? Botspeak, B-O-T B-O-T-speak, so you'll find that's out of Tufts University were the folks that developed it, Chris Rogers

he does a lot of stuff with Lego. I think that that's what he's kind of well-known for. And so he does a lot of things with LabVIEW. And in order to provide a common target for LabVIEW, he's kind of just abstracted microcontrollers as a whole with his Botspike language. And it's essentially just a high-level assembly, which learning to program an assembly is a really,

really, really fantastic thing if you're doing microcontrollers. Because you understand that here's the instruction, it executes, it does this thing. But instead, his instructions are things like digital IO, right? So digital read, digital write, sort of the same level of abstraction as programming Arduinos in sketches or C. But here the syntax looks more like assembly.

So there's an interpreter for this bot speak language. So you can just open up a web browser and start typing in. So there's an interpreter for this BotSpeak language. So you can just open up a web browser and start typing in. Because it's an interpreter, you just type in these commands directly, or you can build up a script and have them execute in a hard real-time loop. Because it's not running Linux, it's just sitting here running this interpreter and absolutely nothing else. You've got the peer-use subsystem here going and doing things. So I think it's going to be a fantastic way for people to just start learning about doing

real-time programming, as well as a very practical solution for people that don't want to understand the details of the subsystem and just want to start coding up some things in real-time. So certainly if you've ever done anything in assembly before, it's super, super easy. But the commands are at such a high level, it's almost like basic, right? Because it's just set DIO high, which DIO high, analog read this.

So I think these projects are doing a lot to improve the state of for for people doing beagle um as well as just general education right so well and they also support arduino and and that other processor we decided we weren't going to talk about anymore they do they do they support a lot of different platforms um and and a lot of a lot of different platforms. And a lot of what we take approaches with Beagle to abstract things, but at the same time kind of leverage the things that are unique, right? We don't have just one agenda in Beagle land. Yeah. Yeah.

When I asked about do you use it to learn Linux or embedded systems? And you said both. Yeah, I know. And I do get that. Everything in the kitchen. And that's where, you know, that's where, even though that other one kind of says they're doing one thing in particular, right? Which is to learn Linux. Which is smart.

And this just comes down to a lot of my engineering mindset right i love to i want to know it all i want to know it all it's just just a little bit we're not going to tell you what to do with this you do something yeah but that does make it a little difficult to get started are there books you have a default thing right and my my little javascript thing you know just because you know i probably get a lot more attention than a lot of other people in Beagle Land, and I've got my JavaScript thing. Or I want to make it easy for a web developer to create hardware. And if I have to find one mission, that's it.

How do I make hardware easy for web developers? And we've got Node.js on here. We ship with it running and a web server running using Express and this kind of Arduino-like language. There's a built-in IDE, Cloud9 IDE, and it's written in JavaScript itself. But there is a lot. I mean, I agree, and I think you're doing a great job, in case I failed to say that before.

I really am. Yes, but. Yes, but. But I did have a hard time getting started because there is so much. Well, but it's a general purpose thing, right? I know. So you can do anything with it. I was thinking up.

Did you try going down the JavaScript route? Yes, but then I ended up compiling the kernel. So how much of that has that been? I totally agree. It had a lot of interest on me because you didn't, it wasn't like saying that, well, you know, thou shalt play with this JavaScript thing first and have this great positive first impression.

No, no. It was because I didn't have anything to do with it. I had an idea of what I wanted to do with the kernel recompile. But what I needed was a getting started or here are the top five interesting projects you can do first. How did you miss the getting started? And lots of people do. Did you go through the getting started guy there's one big the one that pops up on the web page when you connect to it

yeah yeah it did that and and at the end i still i had what a blinking light and a script did you did you blink the leds i mean did you there's like hook up a motion sensor, there's, you know, hook up a motor, there's hook up a button, you know, hook up an LED. Did you? But they didn't, they didn't end in anything. They didn't end in a... Do we build a bot? Do we, what do we, do we, do we start building something?

I mean... Yes. Build a, build a motion sensor you can leave somewhere. Build a light dimmer. Dirty dish detector. A dirty, that would have been really useful. Not just, I mean, blinking a light is great, but it's a start and not an end. And I guess since I didn't have any, you know, I actually didn't blink the light because I was too lazy to wire up the LED because I knew how it would work.

The truth comes out. Well, I did play with the I squared C when I realized that was pretty easy. But I didn't end up with a project. Yeah. I didn't find a project compelling until I started wondering, and I was reading a book about Linux, and they said how hard it was to compile the kernel, so I took that as a challenge.

And did you succeed? Oh, yeah, I did succeed. Although there was some swearing dealing with what the machine equals is, because there's the machine equals BeagleBoard, and machine equals BeagleBoard. And that was just bad. But that's totally a detail we can talk about later. The Linux community, as fantastic as they are at being open and solving problems, ease of use is not a primary concern. That's one of the places where there's a bit of an impedance mismatch, right?

Because they want to make things easy for people who know, not the people who don't know. So the getting started was too, it stopped too soon. Yeah. And what I'm looking for is the, the 200 page book that I only get through a hundred pages of, you know,

the, the, somebody thought about how to put this information together and introduce you to each subsystem sort of thing? Does that exist? Or is that, you know, I should go write a proposal for a rather. The book.

So the crafting. Is that what your book does? The crafting book does a lot of that, right? It introduces you to the major subsystems, but it still might be at two, you know it's it's coming at almost an intro to engineering level that's that's that's for students it's

largely at at you know yeah intro to engineering that that's you know not that you know you could be a high school student that's wanting to learn more about electronics it's it's still not so much on what you can make with it. I mean, although there's lots of projects on building robots, building things with object recognition and some of these other sort of technologies. It is project-oriented, but I don't know that it has the killer project,

unless you consider robots the killer project. Why do you? Because robots... We don't want a killer robot. Because I think it's like the greatest platform for building robots, really. And a robot isn't just something that walks and talks and does things like that, because it's not always mobile robots. The most fantastic robots are like 3d printers or robots and it's

a really really fantastic platform for building robots that build other that build things and it's it's killer for a few reasons these it's got the ability to do the real-time stuff so you want to do real-time motor control but at the same time you know you can be on the internet and you can directly interface to Thingiverse and print from Thingiverse. Okay, so now I want to rephrase my question. Where is the best tutorial to get my BeagleBone black robot started?

It's so hard um the yeah because a lot of it's still technology focused like like the machine kit is this fantastic linux distro for building machines but does it have the here's how to build a machine tutorial no it gets you the technology right it gets you um this this fantastic software that that allows you to allows you to create motion paths and create smart machines. Does it say, here's how to build your bot? Not really. The book, the crafting with electronics with BeagleBone Black does, right? It tells you here's how to build a robot.

So that's the book I should have started with? Maybe so, maybe so. There's three other kind of application books that I think are really nice, all from Packet Publishing. There's Building a Home Security System. There's another robotics book that's specifically targeted robots. So how to build robots with a BeagleBone Black. And then there's a book on home automation. So automating things around your home.

And O'Reilly just came out with getting started with BeagleBone Python and JavaScript that's the make book instead of O'Reilly and that's a nice that one's not going to tell you how to build anything in particular but it gives you some ideas of types of things you can build and it's a nice sort of

out of box getting familiar sort of book the crafting goes a little bit deeper you know it's also you know kind of you know doesn't assume that you really know how to program doesn't you know there's um there's also a book on bonescript which is that that javascript library um so simon monk who is i think, a really great author. He's written a lot of books on Arduino and Pi and lots of other cool tech. Has written a book on Bonescrub.

And look, in the table of contents, it has a roving rover. Roving rover. Oh, a roving robot, sorry. Man, I can't even read these days. But I'm addicted to all of it and that's that's it's too hard when i start but now it looks like i have a better shot of doing something more fun um than recompiling the carnal which was sort of fun but you know with side

projects they have to be amusing or i get bored and go watch TV and read a book. So there's those fantastic books that tell you how to make particularly cool things that the BeagleBone is particularly fantastically suited for. The tutorials themselves are also getting a lot better. We're improving the out-of-box tutorial framework. We're making it something. It's another Google Summer of Code project where we're actually making it more collaborative where everybody can fork. And there's these kind of training card decks

where you can put your projects into these card decks. And they're all maintained through GitHub. And everybody can kind of collect their own exciting card decks of things that they can build and how to build them. And, you know, getting that out of box to start getting into those peer use, I think, is the other sort of really, really cool thing that you can do with the Google Summer Code project. So that interpreter is going to be there in the box

probably starting in the September sort of time frame. So she'll just be able to start typing in real things from that same interactive web environment. So many things going on. The Raspberry Pi just released a few new boards. The Pi Compute they want people to embed and the B Plus with its lower power and more usb and more gpios and all of its nifty things what do you have coming up for us can you

tell me um i can't give you a whole lot of details right now but we we certainly do have some things in the works. There's... For some reason, I'm getting... How come when you can say nobody's listening? So we've got... We're certainly still outperforming those guys and doing a lot more. But I think we want to... Maybe the next step is to do something more exciting on the high end and really, really capture the high end and back. So I'd be on the lookout for us doing something more in the high performance sort of area and capturing that back,

which also might be on more of the expensive side, but still be a lot of fun for hackers like yourself. The other thing is because we're open hardware, we're already seeing a lot of people doing interesting things on making derived clones. I saw an announcement recently of a blue steel board. That one actually dropped some components off the board. You'll see people doing variations of the hardware, and I think you'll see a lot more other people doing innovation around the hardware design. That's something that, you know, because we're open

hardware, we can have that. And because you work for TI, it's extra bonus points. Yeah, extra bonus points. I make absolutely zero dollars and zero cents off of every BeagleBone sold. Well, there's still brownie points. And the karma. It's a good karma. It's a good product. So any last thoughts you'd like to leave us with? Well, the only thing

I think of is just the cool things that I've been able to play with lately on Beagle. I'm always having fun hacking on different Beagle projects. These Google Summer of Code ones have been a lot of fun. I've been building LED walls. So I'll be showing off at Maker Faire Detroit

a four foot by eight foot LED wall that I built up using some stuff out of NYC Resistor, some software out of those guys. And some open hardware out of those guys. But I didn't really understand how the LEDs were refreshing and stuff, so I took one of the Google Summer Code projects that actually turns the BeagleBone Black into a 100 megahertz logic analyzer. And I wrote some C code to analyze the waveform coming out of it and actually reproduce the LED images that were being scanned out.

So, you know, what cool could you make out of it? There's so many, so many things. You know, get on the mailing list, you know, start looking at the G Plus and stuff and following on. And there's an endless stream of fun, cool things you can make out of Beagles. I believe you. It does seem like something we all should be interested in.

My guest... Oh, sorry. My guest has been Jason Kreidner of the Beagle Board Foundation and Texas Instruments. Thank you so much for speaking with us. Thanks, Jason. Thanks, Chris. Thanks, Alicia. And thank you for listening.

If you are building a system around a BeagleBoard, please stop ordering them from distributors and order them properly by contacting Jason directly. Wait, what? I can give the right contact. That works fine. If you like the show, you know, this show, I mean, if you like the show you know this show i mean if you like beaglebird that's great but if you like this show please tell someone about it maybe use it to start a

conversation with that shy person at work or if you are that shy person then with the other shy person and finally thank you for to christopher white for both co-hosting and producing. The final thought for this week comes from Charles Schultz and probably is a favorite to all Beagle owners, of which we are too. And that is that Snoopy didn't start off being a Beagle. It's just that Beagle is a funny word.