Embedded - 197: Smell the Transistor

Episode Date: April 26, 2017

Chris and Elecia talk with each other about science fiction, advertising, ham radios, debugging tools, and programming languages. You can buy Embedded.fm t-shirts until May 18, 2017. You can always buy Elecia’s book: Making Embedded Systems. And don’t forget we have a Patreon if you’d like to support the show directly. Some science fiction we mentioned: Seveneves by Neal Stephenson, Nightfall and Last Question by Isaac Asimov, and the All This Time video from Jonathan Coulton. Digilent sent us goodies to review: one Analog Discovery 2 and two Digital Discovery units. So we did, though we didn't cover the high speed adapters and other nifty goodies. Check out Alvaro Prieto’s Troubleshooting tools HDDG talk for some additional information on the devices. For the giveaways, rules are in the show, hit the contact link to enter. Contest ends May 19th. Chris has been doing low-power ham radio contacts (WSPR) using an Ultimate 3S kit from QRP Labs. We talked about WSPR some with Ron Sparks in episode 76: Entropy Is For Wimps Make with Ada competition is back! It start May 15, 2017. We talked the 2016 competition with Fabien Chouteau in episode 158: Programming Is Too Difficult For Humans. Elecia is still fighting with Ubuntu before she can build her robot typist with her NVidia Jetson TX2 board. Philip Freidin sent in Stanford CS department’s reply to the lightning round question of “what language should you learn in the first college course?” Even better, he sent a link to a google spreadsheet showing how many schools answer the question. Elecia was on the Jennylyn Show. (I’ll update with a link to the specific episode on YouTube when it is available.) March Madness ended with PyBoard as the champion, more info on getting your winner’s hat soon.    

Welcome to Embedded. I am Elysia White, alongside Christopher White, and this week it will be just us. We have lots and lots to talk about. First, I want to let you know that if you want a t-shirt, they went on sale today. Today, today. What's today? Wednesday, April 26th 2017 weird it's like speaking from the future to me we have t-shirts again we have t-shirts again same design different t-shirts uh plenty

of cotton some long sleeves and of course there are women's sizes you guys are lucky there are even unisex sizes, should all be women's sizes okay and they're the same designs as last time same designs, yeah just a few different shirts different shirts okay

so that is the one thing people complain if I don't tell them we've told you let's go on to the complain if I don't tell them. We've told you. Let's go on to the things that you don't complain about. Okay. I've been asking people about books they have recently discovered or movies. You know, the whole something in 2016 you found interesting.

And I did that partially because people have asked us what our favorite short stories were. And then that led to an interesting conversation. And then on Twitter, there was the recommendation for the last question, one of the Asimov stories, when I noted that Nightfall was really cool. Do you think science fiction is an important part of engineering? Yes, absolutely. Because it gets people excited about engineering before they know part of engineering. Yes, absolutely. Because it gets people excited about engineering before they know anything about engineering or know they're excited about it. I truly believe that.

I believe that partially because I've seen other fields get people excited in the same way, like Top Gun in 1986. How many people... Excited about the Navy and going into the airplanes? Yeah, I mean, how many people do you think enlisted based on that two-hour-long commercial? Commercial. But, you know, I think definitely it was an influence to me. I read a ton of sci-fi as a kid, and Star Wars, I mean, that was one of the first visual images I have memory of. And, you know,

it's not engineering and science, but it's the same kinds of things. It's mechanisms and devices and artifacts of ingenuity being used to do things. And yeah, you don't think so? I do. But it led to my next question of, do you think it is important to continue reading science fiction? It does tend to bring a lot of us into the industry. But is it more important to read blogs about new languages or to read science fiction stories about possible futures? I don't know. I think once you're kind of established doing something, it depends on where your inspiration comes from, right? I go back to Arthur C. Clarke. He wrote some short stories and I think was widely credited with the idea of geostationary communication satellites and things like that. So I think there are concepts that come out of sci-fi that people pick up on and say, hey, let's try that.

And you have to be in a position to be able to try it. It leads to ideas and language as well. I mean, if I say build a tricorder, even if you don't know how to build one, you know what it is. Yeah, right, right. Some healthy, health-related thing that will look inside you and i think sci-fi authors are not restricted by the years of of working as an engineer or a scientist uh so they're you know a little more free to do things that may be impossible or come up with ideas that may be impossible but that's the fiction part right but even if they're

impossible they might be 98 impossible and well something a little less might work i'm or come up with ideas that may be impossible. Well, it's the fiction part. Right, but even if they're impossible, they might be 98% impossible and something a little less might work. I'm thinking about even Seveneves in the Neil Stevenson book. And toward the end of it, no spoilers, there was some stuff with just how to get to orbit and there were some novel mechanisms

that went way beyond space elevators to catapults and things like that. And, you know, it's all in the realm of plausible. There's no, there's nothing wrong with the physics of it. It seems hard to implement, but, you know, maybe in 500 years or 200 years, somebody will pull that off. So I, no, I do think, I do think it's still important. And, you know, It's the difference between doing your job and daydreaming a little bit too. There's a lot to be said for daydreaming if you're just doing your job all the time. Especially I feel like these days where jobs are extremely rigid, product focused. There's not a lot of room in the methodologies we tend to use for development

for kind of thinking or exploring tinkering i mean that's that's kind of moved off to the hobby land and out of normal r and d there's not a lot of r left i think particularly you and i have been on the d side and yeah gotten bored sometimes with just doing development, not doing the research. But even so, I mean, you think of... I mean, I'm just thinking of all the companies I've been at and how much research gets done. I mean, some of them leaned so heavily on the research side

that nothing ever got done, and that's another problem. But no, you know know daydreaming is good whether the source is sci-fi or some other source of inspiration i think i think yeah continuing to to feed your inspiration from areas that aren't just purely a part of academics or technology i i i believe that too so in violent agreement here this all came partially up because of the All This Time video from John Colton

and it's a song and a video it's on YouTube, I'll link it but it's sort of like playing the old type edit Infocom Zork style games and there's a song that goes with it and it's very pretty playing the old type edit Infocom Zork style games. Yeah, yeah. And there's a song that goes with it, and it's very pretty,

and it's not been an earworm. Instead, I've just listened to it on repeat a ridiculous number of times. There's something extremely charming about the video, and yet it isn't my normal thought path. Do you think people who hadn't encountered those text games would understand it? No. I mean, at what point, where's the age cut off

for people who understand Infocom games? Or, not understand, but remember what they were. I mean... Remember the frustration. You had to remember the frustration to get the video. Okay. Well, the next section, I wanted to talk about advertisements. Again?

In that we got contacted to have advertisements. And we're proud to announce that Disney is sponsoring this podcast. Disney, bringing you such films as Star Wars, Star Wars, Star Wars, and Star Wars. That's not so much an ad for Disney as it is a hope for Christopher. He's sort of starting the advertisements in hopes that they'll let him do things with the movies. You guys have a lot of money. Most of our listeners know of Star Wars. Anyway, I did say no, we're not looking for advertisers,

but I wanted to remind people that we have the Patreon and that they are welcome to donate money to us there and we will stay ad-free. Bob Iger, you can also sign up to our Patreon. I was this close to suggesting that the potential advertisers sign up to our Patreon. That doesn't make any sense. Did I say it wrong i did i forget it the patronage thing there i'll say it right if i say the patronage thing okay it sounds 17th century but it works and even as i say we're

not going to have advertisements i am going to go directly to people who sent me stuff and now we're going to talk about it. Okay. So, you used it more than I did. Maybe you should talk about it. So, Digilent kindly sent us a few pieces of gear to, as I understand, play with and then talk about. And then give away. And then give away. The piece I played with was the Analog Discovery 2-2, which if you're familiar with these, they're little USB pucks. And there's a variety of companies that make similar ones.

Salier makes one. Digitalit makes one. I think there's a couple of others. And they tend to be a mix of analog and digital in and out. And you can do various things, scope-like things or logic analyzers style things with them. And so yeah, I played with it mostly because I was building a piece of ham equipment and I have no idea what I'm doing. And so I wanted to see if I built certain filters correctly. And we didn't have any equipment around that could do that. And so I looked up on the webpage for the thing and said, oh, it's got a spectrum analyzer and a thing called a network analyzer, which a network

analyzer doesn't analyze like TCPIP. It analyzes how a particular circuit operates under various frequencies. So it gives you a plot of frequency response. So more like a resistor network analyzer instead of a...

Bluetooth network. Internet network analyzer. So yeah, I installed it and pulled it out, and I was able to pretty quickly look at the filters I'd made and see where I'd screwed up and where I hadn't. The software was good. It was traditionally Windows scopey. I encountered a few bugs here and there, but it, you know, it had a familiar layout to, you know, all of the

software USB scopes. I didn't try the logic analyzer part of it. You did. So I can't really compare it to Salier in terms of that, because Salier doesn't have these kinds of heavily analog analysis components. Well, I have used the Salier extensively, so there's some disadvantage to any new thing in that I understand their software. Having said that, and having tried to cheat my way out of actually trying this thing because I've been busy and stuff for the last few weeks, I asked Alvaro, who's been on the show, Alvaro Prieto, about how to use it if he'd ever tried it.

And his response was, didn't you listen to my hack my hddg talk which i had read his slides for before he gave it but no i i have to admit i didn't watch the video but i will link the video and he talks about the differences and one of the things he said to me was that the logic analyzer wasn't as good as the salier and so when i came up against a SPI issue, I went ahead and tried the analog discovery too. And its interface for the logic analyzer was less than optimal in that I don't think whoever designed the software has used a logic analyzer. When I switched over to scope mode, it was as I expected. You know, zooming was a pain and there was fiddling about, but I didn't have to read the manual.

It was what I, all the things I expected to see there were there somewhere. With the logic analyzer, I did eventually figure it out and I got what I wanted and it did the logic analyzing and it was a little odd in spots, but it worked. It worked pretty well. And for, you know, we got a Rigel scope, gigasample, blah, blah, blah. And it's a nice four channel scope with 16 channels of analyzer. I like it, but it's loud and it sits on my desk and takes up a lot of space.

So it doesn't get to sit on my desk all the time. This puck was, you know, palm size. There's a huge dynamic range difference between the sampling rate of that event. Exactly. And the, I mean, so. But if we'd had the analog discovery, I might not have gotten the oscilloscope. Pops out at 10 megahertz, so. Which means that when I'm debugging SPI, I sometimes have to turn down the clock.

But that's usually just one thing in my car. I don't know if that's true of the digital discovery, if there's a difference. I didn't manage to look that up. Because I think the analog discovery may have slightly derated digital capabilities since it's both together. The cool thing about it is the software

has lots of different components, like sophisticated stuff. Oh yeah, they're like different devices. It's a very well-featured spectrum analyzer, so you can do FFTs of your signal. You can even loop stuff back internally, so you can take, oh, it has a waveform generator. Well, that was the actual,

actually the other thing I really wanted to talk about that Alvaro pointed out was that it's got all these outputs and it's a waveform generator. Well, that was the actual, actually the other thing I really wanted to talk about that Alvaro pointed out was that it's got all these outputs and it's got power supplies. If I was designing a manufacturing test, I could easily put power here, measure here, and then write a Python script. It all goes back to a Python script, which is just awesome.

So that's a really neat piece. And you can route stuff inside it. so like i use the network analyzer to do some of the filter stuff and then i said well i could also route from the waveform generator and do a sweep into the spectrum analyzer and do a similar thing and so i did that uh and both worked fine so it's it's very full featured it is incredibly full-featured. It's $200 for the Digital Discover and I think $300-ish for the analog one. $279 for the analog Discovery. And they are really...

I'm impressed. I'm not super impressed with the Logic Analyzer software, but that will get better. But if you're a hobbyist, I think, and you're looking for a variety of instruments, when you get a scope, you often don't get spectrum analyzing. No, the only reason we got the scope we got was because the business paid for it. So to get a whole bunch of pieces like that,

that's a pretty good device. So I was happy with it. And so we are keeping our analog discovery. I'm sorry, we are not giving everything away. I have some more filters to build. But we are giving out the two digital discoveries. Okay. How are we doing that?

Gosh, I have no idea. Okay, how about we'll choose random numbers after the show, and one of the winners has to come from a.edu address. One of the winners? Well, I don't want to make it both. Okay. And we'll have two winners. I have a.edu address. Okay, try not to cheat. Try not

to use the one that you've had for 45 years. Wow! Oh, come out there, bro! Oh, I'm so sorry. College students, this is your chance. You can have one of these for the summer for your project, and it will come out even better. But this is the digital discovery. This is the digital discovery.

So you can only discover digitals. Well, you can also output lots of powers and patterns. Okay. So why don't we restate again what we want? We want random numbers. We want you to send us a random number between... 1 and 997. Thanks.

Okay, 1 and 997. Inclusive. I just want to be specific and make sure that, you know, somebody... And we'll pick. Based on what? We will pick whoever gets the closest. To whatever random number we choose. To whatever random number we choose.

And one of them will be an EDU. Right. Or whatever the equivalent is in your country. Tell me you're going to school, high school, college, either one. But not clown school. Even clown school. one but not clown school even clown school i'll take clown school i have no biases against clowns not like you yeah well you didn't see poltergeist as a six-year-old uh you mentioned doing ham stuff yeah tell me about your ham stuff this was hilarious because i said oh we talked about that on a show. And he said. No, no, that's not quite true. I remembered it.

You had to look it up. Well. Hi, Ron. Christopher's gotten interested in doing Whisper. And we did talk about this on a show a while ago. But now he's doing it, so it's a lot more exciting. I feel like this is a setup. Sorry.

Okay, sorry. Tell us about your harem radio stuff. I have had a license for a few months. General wife. And 10 years in a previous period. And I'm really doing nothing with it, except some video transmission from planes, which is what I originally got it for.

Drums. But I get the magazine, QST, as part of being an ARL member. And I paw through that occasionally. And I came across an article that was a whisper transmitter and receiver that somebody was just a guide to building, a very simple one. And I started thinking, well, I could do that. The components list is here, it's schematics. And then I realized how lazy I was. So I didn't want to kid it myself. So I went out and looked

on the internet to see what else was available to just buy. Somebody else had already kitted it. Right. So I got an Ultimate 3S transmitter kit from QRP Labs. and it's a kit that you have to build yourself. So that was a good luck little exercise for me, given how little soldering I tend to do. Uh, and it will transmit on just about any band from, I think, six meters up to now the two new, extremely long wave, uh, bands. I don't remember the exact things. I thought one of them was like two kilometers or something ridiculous. And yeah, it's kind of, it's not Arduino based as far as I know, but it's an AppMega chip and it drives a little synthesizer chip over SPI, or maybe I squared C,

and it commands that to generate the frequency for transmitting and that chip is capable of generating just about any frequency you like and yeah and it has firmware in it that somebody else wrote so once you piece it together and in theory uh you can do some transmitting and what kind of transmitting can you do uh it's this thing called WSPR, which we did talk about on a previous show, which stands for Weak Signal Propagation R. Okay, now I got to look it up. Weak Signal Propagation Reporting, maybe?

Yes. And the idea behind it is you try to see how far you can get with as little power as possible, which kind of excites me for strange reasons I don't understand. You know, anybody can go out and buy an expensive transmitter and throw 100 watts into a giant antenna and, you know, talk to anybody. But for some reason, being able to get across the country with just a little more power than an LED takes is kind of a challenge and interesting. And so this actually transmits a tiny amount of data, just a few bytes over several minutes. It's 50 bits over two minutes. So it's about one point something baud, 1.6 baud, I think. And basically it's enough to transmit your call sign and your location as a grid location and your power level in decibels.

And that's it. And then somebody else has a receiver, presumably somewhere in the world and if they uh they hear your signal it goes through whisper software which can pick all that out uh it's actually a pretty complicated uh encoding so i i tried to read about it because i thought about okay what would it take to implement the receiver part in software um and it's it's got a lot of error correcting and forward error correcting and frequency shift keying stuff.

So it's not just like Morse code that the computer's decoding. But the receiver software receives it, and then all the receiver software is hooked into a website called whispernet.org and it just uploads the contacts and there's a map there and so you basically once you're done you've got your antenna strung up you turn the thing on and it transmits every few minutes or tries to and uh you press reload on the web page and hope that somebody saw you uh and the first time i did it i wasn't expecting much but i managed to get to ph, which is pretty far from here. And then just before I was about to shut it down, I noticed a contact had gotten to Georgia.

Georgia, United States. Yeah. Which, given that I wasn't sure I'd built the thing correctly and my antenna was a complete hack. It involved shoving wires into BNC connectors. So as part of this, I built a little tuner, another kit I got, which is supposed to match the impedance of your antenna to your transmitter. And I don't know if I did that right, but it wasn't really working with it. So I ripped it out, and in a fit of frustration,

I jammed one end of the antenna into the BNC center and wrapped the other one around the ground on the RF part of the circuit. And that's when it started working. That's how electronics work anyway. I probably violated, no, I didn't violate anything. But yeah, it's just, I don't know why I like it. It's kind of like, you know, talking to Voyager or something.

It's like, how far can you get, you know, by the power equivalent of wiggling your finger in the air? It's just amazing. And I know the guy in Georgia probably, I ended up looking up his call sign. It seemed like he had some, you know, fairly sophisticated rig with a giant antenna. That's fine. You know, I don't care if the deep space network is required to hear me, but, um, so yeah, I'm still playing with that. I've now gotten far enough into it and

it hasn't worked enough past that first day that now I'm like, okay, I got to figure out what's wrong with my antenna. Cause it was just a, you know. I saw your antenna. It came in in a wire pile full of leaves, and I think you threw it away. No, I haven't thrown it away yet. There are some knots in it. But, you know, it's a long wire. I mean, that's what an antenna is, is a long wire.

But I didn't get it very vertical. I just threw it over a tree, and so there's lots of other stuff I could do, and I need to get that tuner working again. So, but yeah, it's really, it's,

I don't know why it's interesting, but it is. And you had fun soldering and this has all been learning. It took me a lot longer to build the kit than I was expecting. I was expecting like, okay, a long day. And I think it took me a couple of,

a couple of weeks of a few hours a day. Not every day, but part of the problem was I didn't realize that all the filters and things you had to hand wind, which is not that fun. It involves counting, and I'm not that good at counting, because especially when you're counting slowly, you know, okay you know rapid two and then distracted by the podcast i'm listening to seven oh no

yeah it's cool i'm enjoying it and the only thing that i saw you get frustrated on was trying to build your own cables and then i just gave you a pack of jumper wires and said don't build your own cables oh for inner the interconnect inside yeah because there are various modules i built a receiver module which i haven't really gotten to work yet maybe because my antenna sucks um but that hooks into the uh into the transmitter and it gets the same the clock it drives the same thing so you have to do that stuff. Yeah, so that helped with that. But no, I got frustrated with the receiver too because I was expecting to receive something. Nothing's working yet. I mean, I can always break down and just buy a complete set

that's just going to work, but what's the fun in that? Then you don't have the impetus to look up how it works because now you could change things. And I have learned a bit of more electronics just by going through the kit and looking how it's you know the sections are laid out and looking at the schematic and reading the instructions network analyzer worked i i thought network analyzers were super expensive oh for the digital yeah um yeah and building analog filters that's something i've never done before so the fact that i managed to actually transmit to anywhere with this thing it was a huge surprise so

not very embedded but it you know it has firmware sort of embedded uh let's see we got an email asking how to get into embedded development. I'm going to use that as a transition to say, hello, I wrote a book. It's about embedded software, doing embedded systems. And you should go buy it if you want to. Was this the question, how do I get my foot into embedded systems?

Yeah. I couldn't understand if the question was, how do I get a job or how do I get into the learning. Well, I think the learning you have to do first. That's probably true. And Will Metzler wrote a nice email that included a mention of my book. He wrote, he was just starting out in the embedded space and searching for a solid design methodology. He found that my book was a chapter-by-chapter documentation of his first few months' challenges developing greenhouse sensors.

And now he recommends it for new grads. Very nice. That was very nice. Thank you for writing, Will. Also, thank you to Michelle. She wrote a lovely email. She found us by hearing about npr embedded which is a very different show i know that some of you found us that way and i hope you're not waiting for our hard-hitting in-depth reporting about violence drugs and justice

because we're never getting there are we i't think so. That requires more research than we generally do. Well, we usually do a little more research than this show. And for those of you who subscribed after our ad on Overcast, welcome. Welcome also to the embedded folks who are wondering how they got here. But we talk about engineering. The question now is how else should we advertise? I mean, using NPR to advertise is probably not something we can control. But if you can think of a way that we should be finding new listeners, let us know.

Other than, of course, the way we always ask, which is for you to tell other people that you like our show and they might like our show too. Because you are our best advertisement stream. I don't think that's the word I would use. It's, you know, discovery. How can we get people to discover the show? Yes, actually. Advertising implies money. Yeah.

Yeah. But you went on television last night. You know, I didn't put that in the show notes at all in hopes that you wouldn't notice. Yeah, well. So, you went on television last night. And talked all about embedded systems and your book. And how was that? What was that like?

Tell us all about that. You were there. Oh. And they probably don't care. Who doesn't care? Whoever's listening. No? All right.

So I was invited to go on the Jenny Lin Show, which is about... About 27 minutes long. Exactly. On her website, she talks about it being for leaders, inspiring leaders. And I wanted, what is it like to be on a TV show? So I said yes. And I don't know what she thought we were talking about, and I didn't know what I thought we were talking about.

So it's kind of a miracle. I decided a couple days before recording that since I was so clueless, I was just going to start piling things into a box and they were going to be things that sort of represented my career in some way or another, something that you could point at and I could babble on for a while. And I put in some of my leapfrog toys. I put in an art light that Micah Scott made for us that I just like and thought it would be a nice intro to the show and how it's neat for me to be able to talk to interesting people. And I chucked in an Arduino and a race car and an IMU board and a copy of my book in Chinese and in English. The box is ridiculous.

It's just, it ended up being quite heavy. That was kind of cool. Thank you for carrying it. And when we got there, I think she looked at it and she was like, huh? And, you know, there were raw boards. The IMU was just on a board. And there were the BB-8s that were in pieces.

The assembled BB-8s. And she was like, did you work on this? And I'm like, no, I took it apart. And she was just like, because this wasn't a show for technology.

It was. But she's had Steve Wozniak on and other technologists amongst some other random people. But so, you know, it's not completely. Yeah, one recently about AI and data.

So it's not totally, I mean, it's Mountain View TV. So you have to assume there's some technology so it's not totally i mean it's mountain view tv so you have to assume there's local access there was a certain uh there's a certain wayne and garth vibe or uhf if you're familiar with that movie vibe to it um but it was actually pretty cool to see how much goes into a tv show even even even a half an hour cable access a lot of people a crew of at least seven or eight people at least uh they had a big control room with modern stuff they had three cameramen um yeah kind of impressive to watch coming from you know producing a tiny podcast it's like okay you know we throw up our gear and do some stuff afterward. But that requires, they had a lot of planning.

They had a sequence they went through and a lot of interesting electronics and stuff. And yeah, it was fun to watch. And you did a good job. Thank you. And I'm sure it'll be on YouTube for all of you to find very soon. It will be on YouTube and I will put the link in when I have it. I know she doesn't want anybody to see it. I so don't want you to see it.

If you didn't want anybody to see it, you shouldn't't want you if you didn't want anybody to see it you shouldn't have done it then there was that part where i pulled it so they did choose a few toys and i end up with just toys and my book um and the frog and so i had one of the leapfrog stuffed frogs that was she thought that was cool and i i she's asked something about embedded systems and so I pulled it over and I said, this is an example of an embedded system stuffed talking frog. Embedded system stuffed talking frog. And then for some reason, I started talking about ShotSpotter.

I had a stuffed frog in my lap and I start talking about gunshot location systems. I did think that was a bit of a tent a bit of a i don't even know why well you had a plan i had a plan and but your plan probably involves not picking up the frog first exactly um there were a few moments like that where i was like oh my god what the heck am i doing and i had max, which is the octopus that it will notify people if you don't pat it every day. We've talked about it on the show. And I loved the fact that she cuddled with that stuffed animal. I mean, it was one thing for me to have my stuffed animal in my lap,

but for her to have my stuffed animal and be petting it, I really liked that. So yeah, there were plenty of awkward moments. Mental note, if you're going to take toys on the show that light up and talk, put batteries in them because somebody might want them to light up and talk. Yeah, the director had some comments about that later. It was like turning off my phone. I didn't want it to light up and talk unintentionally.

Yeah. And yeah, I'm so sorry. So are we going to transition to YouTube? Do all videos? No, we have one listener who's been commenting and he put it on a recent show that we should be putting up the pictures of what we're talking about on the YouTube channel.

That sounds like work. That sounded like a lot of work to me. Video is hard. People who use the YouTube channel, you have to understand what the YouTube channel. That sounds like work. That sounded like a lot of work to me. Video is hard. People who use the YouTube channel, you have to understand what the YouTube channel is for. The YouTube channel is for people who want to listen to our show on YouTube. That is it. People asked, and we said okay, and we put it up, and there's a workflow that makes it

quite easy on us, but I don't really have the bandwidth to do any further exciting things with it so it's mostly just another avenue for audio and even if we took a video we'd have to edit it because i mean you do cut out if we have pauses or coughs or whatever and then it wouldn't be synced and that would be strange no and i'd have to buy a lot of cameras and then set up to yeah yeah clearly he doesn't want a lot of cameras right now so would you do it again not tonight would you have your own tv show no no i i don't think so what if you're on the other side i mean i like interviewing people but now people do tend i mean some people get nervous to come on the show. And I try to make it easy for them. And the people who mention they're nervous, I usually contact them a few weeks later and ask them if there was anything I could have done that would have made them feel better.

But going on a TV show, I was super nervous. I'm not nervous for other podcasts very much anymore, but I didn't, I guess I'm used to being the voice and not the face. And I, I didn't, the best part about being the face was, and the reason that I did the box was because I was like, okay, a lot of people haven't gotten to see many of these things. I should have the physical aspects. It shouldn't be like me talking about my book and waving my hands. I should have physical things. And I was a little sad I didn't have more of the physical things next to me, but I guess you can only cover so much in 27 minutes.

And I packed it in, man. I just talked fast. All right. I won't subject you to any more of that. Actually, I will bring up one of the things she brought up, which is I had a little bit of a rant about how embedded systems aren't just for nerdy engineers. It isn't just people who go to school.

I know a lot of you out there. If you're not, you're not professional engineers. Or if you are, you're not necessarily professional engineers doing embedded systems. But the people who are afraid of trying it, who are worried that they can't even follow the directions to make an Arduino thing. You can. I mean, Arduino is built for artists, and it's made so that you can touch it. It's not going to shock you. It's not going to hurt you.

And you may hurt it, and that's okay. It's $20 or less. It's not supposed to be expensive enough to matter. Yeah, that's the funny thing. I remember when I got first computer when I was a little kid and just how careful we all were around it with the disks and, you know, anything involving opening it up, grounding yourself. I mean, because it was really expensive and, you know, even do-it-yourself electronics back then was probably a lot more expensive than it is now. And so I don't know where that kind of hesitancy, I might break it, really comes from. Because who cares?

I think there's a lot of, I might feel dumb. And that is the thing that I would like to encourage people away from. Is yes, I will totally admit that I have two projects I have felt completely stupid about. I have projects I've worked professionally I have felt completely stupid about. I have projects I've worked professionally that I felt pretty stupid about. I mean, in the last day. In the last day? Yeah. Busy?

Well, we're going to talk about Jetson eventually. Oh. My ham kid, I blew up the power amplifier transistor within seconds of setup. Twice, I believe. Twice? No, just once. Oh. I think.

You were telling me that you thought me. Well, I didn't get any contacts, so I was blaming that, but I don't think so. But, you know, that was, I didn't quite read the instructions properly, and I turned a potentiometer until it went, until I could smell the transistor. And it turns out when you can smell the transistor, that doesn't mean it's working really well really yeah so yeah we all we've all blown stuff up i blew up uh an expensive display i think the first or second day i was contracting at uh my first real embedded systems job and uh you blew up a dna scanner i did um but getting back to the point of this

which is not how incompetent we are oh i was just enjoying talking about blowing stuff up it's not incompetence um people who are afraid people a lot of people are very afraid to even try. And now that we are entering this period where you really can build a lot of stuff yourself. Yeah. And there are tutorials and examples. Not just stuff, but things that help that you might want. Like the person monitor that Maxwell is. Or weather stations, or a thing to tell you when your mail's arrived.

These are easy. You can build these out of modules from SparkFun pretty easily and quickly now. And yeah. And I want people to not be afraid. Yeah. So I hope that was one of the things that I got across. I think you did. I think that was one of the things that I got across. I think you did.

I think that was. And yet I do have, do you want to talk about Jetson or do you want to talk about books? Let's get rid of Jetson. Okay. And then, yeah. Okay, so I have this idea for this robot typing thing, which is, I'm sorry, it is a terrible idea because one of the things I haven't told you is that in order to make the robotic hands type, I probably have to type to them. Or we have to do some sort of voice control. We all know that robot hands typing is pointless because that's just an extra middleman.

Right. It would be more efficient to do voice control straight to whatever it is. It's a learning project than an art project. Yes. It is not a good actual thing because you do have to tell the hands what to type. We've established it's pointless. Okay. Yeah.

But it's with Jetson, and I'm excited because I want to learn more about machine vision and machine learning. We should say what Jetson is again. Oh, Jetson is an NVIDIA board, and it is really way more powerful than anything I've ever used. I mean, my laptop is more powerful, but this is really way beyond Cortex-88. It's a tiny multi-core ARM board with gpu massive gpus and software to do

deep learning and so and and the tutorial which it said would take two days which i've been working on for about three weeks now um we'll get you through the computer vision and machine learning part. And I've had some plans for the motor hands. But the first thing you have to do is to flash the board with new firmware. And I hate, hate, hate, hate, hate, hate, hate that that is where I'm stuck on. You have to flash it with new firmware from Ubuntu. And not from a VirtualBox Ubuntu, but from a native Ubuntu. Which doesn't make any sense to me because I had a Jetson 1, which I did for a work project,

and I did it from a VM Ubuntu without any trouble. So something changed. I tried and tried. Or VirtualBox is crummy. I played with my USB. I did VirtualBox updates. Never tried it from my computer. We should still do that. We should still do that.

But I decided, you know, a lot of people are going to Linux development. It would be nice if I could do native Unix development. And I looked around for this live CD thing. And I got... Which is a hilarious misnomer these days. Because there's no CD involved. And I got an external drive, a very fast SSD, nice and big. I was just going to have it boot to that whenever I wanted Unix and those of you who are rolling your eyes and laughing at me I didn't know. I didn't know that was impossible. Because you can boot to Ubuntu

with a thumb drive and then you can say okay I don't want to install that I just want you to run and it will run. So I can do that but I can't use that to install the Jetson because reasons. Reasons that involve how big the thumb drive is. And I did figure out how to be super user in that, but wow, that was a pain in the neck. And there aren't enough drivers to do everything you need. It's just, I'm sitting there going, okay, I need to know how to do everything you need. It's just, I'm sitting there going, okay, I need to know how to do USB drivers on a run time live thumb drive of Ubuntu

in order to do flashing, in order to do... You were never going to do it from the live. You're never going to be able to do it from the live CD. You had to be able to install it something. I tried to install it with the SSD. And then you got stuck with Intel's protected UEFI, and now you can't install anything anywhere. So I can't boot.

So maybe somebody will tell us how to install Linux on an external drive in 2017, where apparently there's chips inside modern PCs that don't want you to install anything anywhere? I have a Dell XPS, and I don't want to hear about it if you're just Googling it. Because trust me, I have Googled it over and over again. I have done scary, scary things where I say, okay, it's okay if I don't get back to my Windows, and I don't want to do the scary, scary things anymore because I haven't managed to delete, and I don't want to do the scary, scary things anymore because I haven't managed to delete Windows,

and I don't want to. Which brings me up to Sam's question. Sam sent in a question. Do you use Windows as your development environment? Can you elaborate why you prefer Windows over using Linux for development? Well, Sam, you have chosen a bad, bad week to ask this question.

Linux is fine. I like Linux. If you were doing normal development, this wouldn't be a problem. But I currently use Eclipse and GCC. So it's not like I'm using Windows native things. I deeply miss using Kyle. What? and GCC. So it's not like I'm using Windows native things. Windows native things. I deeply miss using Kyle because it was all in one place and

the stability of the debugger was just so much better. I'm not using OpenOCD. I have used OpenOCD in the future. In the future? I've used it in the past and I hope never to use it in the future because it was awful yeah we've been down this road um but okay so the reason i use windows is because

that's where my compilers are and i know that gcc exists gdb is okay but i would like a modern IDE. I know that's spoiled of me but I really do prefer to walk through my code. You ever tried Visual Studio Code? I think I might have mentioned it once or five times. Is that in Linux? Yes. Is it? And Windows and Mac.

Oh, I thought it was just Windows. It's everything. Yeah. All right. Well, we should try that. I should try that. But mostly I prefer Windows because that is what my chip designers prefer. When Arm writes a compiler that is as fully featured as Kyle is for other operating systems.

I know you're making a face because Kyle sucks. Kyle totally sucks, and I 100% agree. But I don't want to spend my time writing giant make files, or even worse, porting giant make files from one stupid version of CMake to blah, blah, blah. CMake? What are you doing? I get so lost in the tools, and I don't want, I want to write embedded software.

I don't want to mess around with tools. And I don't want to mess around with versions of things. For professional engineers, most of the stuff, I mean, people use IR and they use Kyle and stuff like that, and that's just on Windows. And I also develop on Windows, much to my displeasure. I don't like it and I have to run into VM and lately it's been working worse and worse. And there's hours a day where Windows just decides it's not going to play. By that, I mean, it's

actually just not going to do anything probably because there's an update that is silently happening in the background and it's awful. But that's where the tools are. Yeah. And I would rather use the tools that everyone on my team can use. Yeah, but I mean, the Rube Goldberg machine that

I end up using on Windows is truly horrific with the hodgepodge of Python here and MinGW there and Sigwin over here. It's just all gross. That is totally gross. I don't think that's how you should be doing it. How else are you going to do it?

Well, I mean, we can't go into the details of your project. There's no other way. You can't just use IAR and then demand everybody else use IAR? No. I mean, isn't some of that because you've got like... There's very specific reasons why we can't do that. Because there are portions of the system that we cannot require IAR to be used for.

But, yeah. I think the more time you spend dorking around with tools, the less time you get to actually making a good product. Of course that's true, but that doesn't make the tools go away or any less awful. But that is why I tend to use Windows, because it means less time dorking around for tools for me. And it means more time trying to figure out what system is doing a 98% CPU for me. You aren't running native, are you? I think it would do the same thing if I was running native.

I rule! Well, Windows doesn't work that bad for me. I don't get stuck. USB goes nuts all the time. My Windows 10 USB driver is not very good. And yet, I don't get stuck like that.

I don't know. All right. So anyway, the O'Reilly Salary Survey, which is out for 2017, says that 71% of us use Linux to develop, 61% use Windows, and 48% use Mac. Wait, did you say 71% 61%? For all software not for embedded systems. Wait, that sounded like it added up to 200%.

It did. It's because you use Mac and Windows and I do use Linux and Windows. I use all three. So yes, it is going to add up to more than 100. And still so yes it is going to add up to more than 100 and still open OCD is an unstable piece of crap that's my final word on that

no it isn't you're going to bring it up in two shows I guarantee it you can't let it go it's so bad have you tried to read it recently I used it November so yes recently recently? I used it November. So, yes.

Recently. Gosh, you're so mad right now. I'm not that bad. Okay, good news. Oh, I mentioned this last week too. Andre has been explaining SPI or spy, depending on how you want to say that. And have you ever tried to explain SPI to someone?

It's got wires. They wiggle. There's a clock. Yeah. You read stuff when the clock is at one position. I ended up having to explain what an oscilloscope is because you do kind of need to know. What about a voltage?

You have to explain a voltage. I don't think you went that far. I think you did assume there was that. Pull-ups? That's for I squared C. You don't need them for spy pull downs no sorry pull sideways yes and then you do the hokey pokey and you spy yourself around um so yeah that's on the blog and it's he's doing pretty well also on the blog is march madness because somebody gets a hat i admit that i dropped ball on that but we dropped the ball on march madness i didn't mean for that um uh but the pie board actually won the pie board beat out the esp8266 in a hard-fought match. And for all of you shouting at the radio or your headphones right now,

I'm sorry. I was shocked too. But the judging criteria were random and relatively fair. Yeah, so I will have a post up this week that talks about how you score your sheets, and you can tell me if you won and then I will send you a hat. Isn't that exciting? They're baseball caps. They're very cute though. You're cutting that, right? Cutting that noise? No, not now.

People are still looking for internships. So if your company has not considered having an intern, please do. It's a good way to teach somebody something. Well, and it's a good way to find really good people. Yeah, it's like a three-month-long interview. I got a job for my first internship. Me too. And mostly if you are looking for an internship— Not that I'm a really good person

didn't mean that implication to hang out there most of the companies that are still looking are the startups who just realized oh my god summer's coming because they have been working hard and just looked up most of the big companies have already selected their interns but do keep looking. And if anybody needs an intern in CS and EE in Canada, which is, I know, a pretty big place, but still. I know a kid. He's probably pretty good.

You know we're at war with Canada now, right? We're not at war with Canada. Or we might be after this. Philip Frieden sent us an interesting article from Stanford about how their CS department is updating their introductory courses. That they're no longer using Java. And that included a spreadsheet of a large number of different universities and what they use in their introductory courses. Oh, very interesting. I did find that pretty interesting. And it was sort of, I mean, like I mentioned the O'Reilly salary survey that talked about different languages and salaries. And it was pretty cool.

But this is sort of the future. You want to know what the kids are learning? This is what they're learning. Apparently they're learning Python. Yeah, it really was pretty heavy on Python. I like that a lot of them were still using C++. Pascal! There's an outlier. So I will put those show links in the show notes.

Java, Java, Scala, Python, Python, Python. Gosh, my life would be so different if I learned Python first. It would have changed how I thought about software. It's different. I'm not sure it would have been good. Okay, now, emails from listeners. I know I've been sprinkling them in, but here's one from Oscar. He has been listening regularly for the last few years, and he has a question about starting a new firmware project. How do you go about choosing a language? Obvious technical points such as code size and performance, but what non-technical constraints do you consider? Regulatory familiarity with languages, the possibility and interest to learn available tools?

Yeah, all that. You know, it depends on the size of the firmware project. I think that's going to dictate it a lot. So if you have a very small processor, you're not going to be running Python, probably not going to be running C++, Java, that kind of thing.

So, you know, you have to get to the Cortex-M3 and above before you can start considering a little higher level language. And even then it's, okay, what performance are you looking for? How much RAM do you have? Because you're going to waste a lot of it for the runtime. You think C++ is a lot bigger and more wasteful? No, but if it's a really tiny chip, you probably wouldn't do C++. Even the M0s that do C++. I was talking smaller than M0.

8-bits? Well, I mean, Arduinos run C++. Fine. C++ is fine. Forget I said anything. Assembly on everything. But yeah, I mean, that's kind of the first thing. One of the constraints he didn't mention was example code.

I find that the vendors are giving us just a lot more code. And while some of it is terrible... There's a lot of it. There is a lot of it. The food is terrible in such small portions. But the code they give you, it probably runs to first degree, which is kind of terrible because reading code is very hard. But they did think about it.

The problem with that is you can get to a point where stuff works, but then you have to know how to get to the next step. Well, that's when you hire me. Wow. No, you do. And you have to be willing to read their code. You have to be willing to change their code. You have to be willing to understand the implications of,

oh, they're not doing this power efficiently. Right. Or, well, this is great, except it takes 30% of my RAM and I have no room for my application. Or they have three different examples and you have to merge them together in a way that they don't fight. Yeah. But that's going to drive you to C.

That does tend to drive you to C. There are some examples in C++, but that is a consideration for me is that i kind of want my code to look like the surrounding code and if like nordic gives you a ton of code so i tend to end up writing stuff similar to that even as i hate their style guide it's still consistency is more important to me um do you think we're heading toward the modularization of embedded software? I mean, it feels like that. It sure does.

ST and Nordic are both giving people just a ton of code, and it means that you're writing less, ideally, or at least you're using their examples as examples, as reference code, so that you don't have to invent everything yourself yeah um regulatory i don't think there's anything really regulatory per language oscar pointed out that uh he works on medical software yeah there's something with medical software there's no and there's msrc. So automotive would prefer MISRA C, even though there's a MISRA C++. Yeah, medical doesn't care, though.

I mean, they care. There's no extra brownie points, I think, for using MISRA. I'm not even sure that the FDA really knows about it. Maybe they do. Andre would probably yell at me. But I don't think there's any first order kind of, oh, you can't use Python in a medical device. No. I mean, it would be harder to start a site.

You can't use C Sharp in a medical device. I mean, you can. I think familiarity with language is probably one of the big ones for me. But that can be solved by sending your whole team to a class or sitting them all down and having them go through an online class together but i mean that would only be if oh god we have to use this language that we aren't familiar with for some reason no even you know sometimes there's that one person who's like this language is better i want to use it and you fire them immediately because they're disruptive well and then then there's this one person who's like oh no i'm using assembly like i've used for the last 45 years 45 years

i know i'm a little stuck on it uh and so classes are the way to figure out yeah even the online even the online courses as long as you go through something where you have a few projects that are that you have to do and that you know what your team knows and so if somebody wants to put in templates but that hasn't been discussed in the class then you know you're doing something odd and novel and maybe not good um so it's i really do think familiarity with language in your team. I don't discount the possibility of using the language that's nifty and new, Rust or Go. I would. Well, I don't discount it just on principle, but I don't know it. And if I was going to go on to a team that that was the goal, I would want the book they used, or the class they used. I think at least as far as rust goes, it's not

baked yet. I mean, there's still changes happening, and that's very unpleasant to deal with. Same thing with Swift. I mean, Apple's pushing Swift for applications development, and a lot of people are using it, but they're still in the, okay, we make major releases and break the language, and that's something you're going to have to get used to. So I wouldn't want to play with that in an embedded sort of world that makes things twice as hard. Well, I'm not, I probably would argue against Rust. But if there was a project that was using it, it's a little tempting. I'm interested in learning stuff.

But it's not really efficient to pay me to learn stuff except what i need to know to do your project anyway fabian shooto do you remember him yes i do he was the ada person we talked to yes he came on last year to talk to us about learning ada and about the make with ada competition That is going to start this year. Oh, cool. It's going to be in May 15th is when they plan to start it. I don't think their website is updated yet. So I will note when it is updated. Yeah, it sounded cool.

Oh, it was cool. And speaking of languages that people should try. That's one that learning more about that led me to better C and C++. It changed some of my thought processes in a good way. So, yeah. Do feel free to check that out. And it is okay with me if you cross-post your Hackaday Prize entries and your Make with Ada contest entries. We won't tell, I guess. We won't tell. Okay.

Although, if you do make the Hackaday Prize entries, you should have sent them to us so that we can retweet and everything and give you the skulls. Because the skulls are worth dollars. I don't know what you're talking about. The Hackaday Prize. I know. If you like it, you give it a skull. And then you trade in the skull for money? And then the person who receives the entirely fake digital skull, not an actual skull, can trade it in for money.

Okay. You seem skeptical. I just, the skulls, I'm tripping up on the skulls. Sorry, it's a Hackaday thing. Okay, well, that covers my list. I mentioned t-shirts, right? You mentioned t-shirts at the top. You can mention them here again.

Don't forget about the t-shirts. Don't forget about the t-shirts. Don't t-shirts at the top you can mention them here again don't don't forget about the t-shirts don't forget about the t-shirts don't come complaining to us in june when the campaign is over and we've talked about the t-shirts 17 times that's right i didn't talk about the survey but the whole show next time it's just us is likely to be the survey well that'll be a month or so yeah okay. Okay. Unless our guest bails next week, in which case, meh. It's always the danger.

I'll have to build something else to blow up. Yeah. Maybe we'll let everybody listen to me while I actually swear about Ubuntu

and installing it. Well, by then, your TV thing will be up and we can talk about that. Oh, God. No, don't watch it. Don't watch it. Don't watch it. All right watch it. Don't watch it.

All right. I think that is enough of a show this week. Thank you for listening. Thank you to Christopher for co-hosting and for producing. And hating Windows forever and ever. We like to have a balanced approach. We hate everything, really. And thank you, of course, for listening.

If you want t-shirts or to check out the Patreon or whatever. Subscribe to YouTube. Subscribe to YouTube. Subscribe to the newsletter. Read our blog. You can find that all on embedded.fm. And our TV show will be at the top of the website that all on embedded.fm and her tv show will be in the at the top of the website for a couple weeks we'll just put it up there in an embed no it will be in a tiny little

link that i'm going to link to a single character a dot one dot one pixel yeah micro dot yeah um but yes i encourage you to view source on the website then. Go find the dot. And a thought to leave you with. Of course, we're going to read Winnie the Pooh. Are you ready? Oh, Christopher's already napping. That's great.

So Winnie the Pooh went round to his friend, Christopher Robin, who lived behind a green door in another part of the forest. Good morning, Christopher Robin, he said. Good morning, Winnie the Pooh, you said. I wonder if you've got such a thing as a balloon about you. A balloon? Yes, I just said to myself coming along, I wonder if Christopher Robin has such a thing as a balloon about him.

I just said it to myself, thinking of balloons and wondering. What do you want a balloon for, you said. Winnie the Pooh looked around to see that nobody was listening, putting a paw to his mouth and said in a deep whisper, Honey. But you don't get honey with balloons. I do, said Pooh. If there are advertisements in the show, we did not put them there and do not receive money from them. At this time, our sponsors are Logical Elegance and listeners like you.