Embedded - 523: Bad Experience With Donuts

Episode Date: April 2, 2026

Chris and Elecia chat about Leapfrog toys, things they like, large company politics, awards, and open source governance.  The Toy Story 5 Trailer with LilyPad toy which is suspiciously similar to the LeapFrog LeapPad tablet. Which is different from the original LeapPad which had cartridges and capacitive touch (capacitive touch was used on the globe as well… the latest globe also has a screen). Why does Elecia want an award? Who knows? But right now, she's getting ready for a listener to nominate the show (Chris and Elecia) for IEEE's Meritorious Achievement Award in Outreach and Informal Education. Probably. But we've got nominators and endorsers so that's mostly sorted. She also signed Embedded up for the Women Podcasters Award which is a popularity contest. You can vote here: www.womenpodcasters.com/awards-voting. The show is under the Science Podcasters category. Some things we like: Ctrl-R: In a command shell, ctrl-r searches your history. Better than ! because you don't have to remember as much. Data bars in Excel: This can create a plot of your data in the column. Merlin Bird ID: Want to know what bird is making that sound? Want to know the name of the bird you just saw? Merlin Bird ID is a free app that is amazing. Plucky Cards: Want to have a 1:1 where you talk about more than your status? Choose a card, any card. Or maybe just look through and have a 1:1 by yourself Just reading about Bunnie Huang's new RISCV board Dabao Evaluation Board for Baochip-1x taught us things! We're not sure what we'd use it for yet but it does spark a few ideas.  The Embedded.fm Patreon Slack book club is reading Pragmatic Programmer 20th Anniversary Edition. Talking about open source projects and governance models, we referenced three contributing guidelines: Valetudo, ESPHome, and Zephyr. Some later research led to Leadership and Governance | Open Source Guides and presentation by Cornelius Schumacher – The spectrum of FOSS governance models (Slides). The link between the politics associated with the size of companies and the open source governance models clearly needs a bit more thought. Transcript

Welcome to Embedded. I am Aalya White here with Christopher White. This week, it is just us, and I have a list of things for us to talk about. Great. And we're sorry we missed the last episode, but we were tired. Indeed. Indeed. So tired. All right. So, let's get on with the list. Well, I'm going to start with Gary Brown, who we don't really get to talk to. But he did send us. Some information in the email transmitted through many layers. What are your thoughts about the new toy story trailer? In particular, I believe Gary was referencing the leapfrog toy that is the villain of the story.

The very, very, not very veiled leapfrog toy. It was a, it didn't say leapfrog, they can't do that. It was a fraud, and it was nominally educational. Oh, yeah. No, I don't think we're fooling anybody here. No. Yeah, so if you haven't watched the trailer, it's Toy Story 5, and this time technology is the villain,

and the technology is, I remember you had, so Leap Frog, maybe you need to explain what Leaprog kind of was, because I don't know that they exist really anymore. They were bought by V-Tech and they are still available. Okay. But it's not like there's new developments that I know of. Okay. Okay, so Leaprog, I worked at Leaprogg around year 2000.

And when I got there, it was like 600 people-ish. Big. And they made educational toys. And before Leaprogg came on the scene in Toiland, There were toys that would be educational, like a bus that had letters on it. But because, you know, 26 is hard. 26 letters. 26 letters.

They would, like, stop at, you know, I or J or wherever it felt convenient to stop. Leap rug was always very much. You have to use the whole alphabet. It's not just decoration. And they made the leap pad, which used a capacitive sensing technology. where you would run a pen over a piece of paper and it would read to you and it would sound things out. And that was probably their flagship project and really fantastic. And this wasn't a tablet. You put a book in it.

You put a paper book. You put a paper book in it. And a cartridge. And the cartridge in it knew where in the book you were based on, I'd forget how it worked. Based on some capacitive sensing and with a page of the book. Okay. And then you could scroll over the page of the book, the actual words on the book with pen. And if you did it slowly, it would sound them out with you. If you did it quickly,

it would read it to you. There were buttons on the page that would let you do different things. So before we get into Toy Story, since this is the Embedded Show, I forgot how those were actually made, because, like, this was 1999, 2000, right? So were there any microcontrollers besides 8051s? Was it an 8051 or was it something more exciting? It was a custom ASIC. Oh, right, right, right. And it was an 8051 core. Okay. But there were plenty of other microcontrollers at that time.

Well, yeah, but that's the one that was the most. I had been mostly working with like, analog devices, sharks, and TI, C-2,000s. I think of sharks' purpose-built DSP is not microcontrollers. Well, I had been doing a lot of microcontroller work. DSP work at that point. But we were using them as microcontrollers to control devices. There's sharks just a couple of feet from you in that thing.

Okay, so it was like Sun Plus was the processor? Sun Plus was another processor that we would use for the cheaper toys. And where the 8051 had a bespoke Kyle-based compiler. Okay. The Sun Plus had a assembly language. We didn't usually program it. We just told them what we wanted. Gotcha.

And part of the LeapFrog goodness or Secret Sauce was that they had their own audio generation system that was based on putting pieces of audio together, not at a word level, but at a sound level. Right. Okay. And that library, they built up over many years. And it was so fun to play with, because everybody got mad at you. And I'm like, you added a variable. And I'd be like, yeah, but I also put it in the register. You guys weren't using the registers enough.

It was so fun. But usually I got to build toys. They had purchased a company that had the Odyssey Globe. Right, I remember the Odyssey Globe. Which was, again, a capacitive sensing with a pen, and you would touch a country, and it would tell you about the country. Maybe it would tell you the capital, the country, the population, the currency, all kinds of information. And the great thing about that was it had a game. would be like, find a Burundi.

My geography got so much better as I tested that. So I'm looking back to the topic. I'm looking back at Leapfrog's website or they have a store. They still exist. The brand does. And there is a LeapPad Ultimate, which is a tablet for three to nine-year-olds with a screen, a quad-core processor. And a bunch of, it's basically a, it's Wi-Fi. It's what's in the movie.

Yeah. Except it doesn't look quite like a frog, but it is green. And it is called Leapad Ultimate, so they kept that branding. So this sort of exists, so I guess Toy Story went forth. What do you think about your lineage, your heritage, your legacy being used as a villain in a Toy Story? I mean, I haven't seen the movie. Maybe they all make friends in the end. And I do think that there is a definite, when do we give kids screens?

Right, right. It's a good question. It's a hard-to-answer question. Is it worth giving them screens earlier if it gets in the alphabet earlier? Three to five? I mean, yes, they need the alphabet going into kindergarten these days, but I don't know. Surprise, they don't need quantum mechanics going into kindergarten. Open web access with parental approval.

No. Yeah, that doesn't sound like a great feature, but it would be a feature everybody would ask for. You cannot have it. So, I mean, I think, I hope that the movie is good and that they do figure out a little bit about screens versus imaginative play versus educational play. And maybe they talk a little bit about how imaginative play has just as much value, if not more, at certain ages, than screenplay. Yeah, I don't know. I mean, I don't remember when one of the leapfrog toys was in The Simpsons, and I was so excited because it was one that I had worked on.

I was just like so excited. I didn't care what they did to it. It was just so cool. It's like when everyone was wearing FitBets. Oh, I wrote that screen on the president's wrist or whatever. Yeah. So at the risk of indulging nostalgia, do you miss that kind of really small device embedded programming trying to get the devices we had at the time,

the microcontrolls we had at the time with their much more limited resources to do stuff. And the manufacturing aspect of, if I can save the whole system a penny, I've paid my salary for the year. Well, I mean, I think that's still true. That was cool. That was cool.

I haven't been doing a lot of high volume lately, so. I wasn't at Leapfrog for that long, but it was a job that was really important to me. It's a job that I have missed many times. Some of it was the community. The people who worked there knew they were working on toys and really enjoyed it. And they were working on educational toys. So, you know, you go home at the end of the night feeling good.

I got to do a lot of trying to make more platforms where instead of making one toy at a time, we would try to bundle a series of toys so that the way the buttons were handled happened similarly. similarly and generally write less code. Yeah. But I also, you know, did a few really bespoke things that couldn't fall in my lines of toys grand plan. I don't remember the question, but I do look back at LeapFrock with deep nostalgia. That's good answer. There's actually a meetup happening this month and I am so excited.

If I was in Bash, what does Control R do? Control R in Bash gives you an interactive history. So you can scroll, you can search through your history. So where normally I just push up arrow to get to the last command I wanted, which might be 20 in the stack. I can instead push Control R and then type Python. for the thing that I last wrote Python. Or, you know, if you have a Git command that was complicated

and you remember the argument, you could just look for the argument. So I'm searching through my history. And I could technically just search through my history file if I wanted, Grep. Sure. But I never do. Right. Okay. I think it does other stuff too.

But, and if you keep hitting stuff, it'll keep giving you another option. Yeah. But, yeah, I forgot about that. Somebody mentioned that again. It's not just Bash. I think it's all, and most shells have it. Yeah, it's a useful feature.

I came across a useful Excel feature, which for me is kind of weird because I am definitely at the high end of Excel features. I've used conditional formatting. Additional formatting. So let's say you run 10 tests, and they all come out with a different number. Or say, actually, say you're a T-

teacher. Okay. And you have the assignments across the top and students down as bros. So Amy gets an A on test one and a B on test two. Okay. And let's say you want to know how everybody did as an average on test two. You highlight that column. You go to conditional data and then you are given the option of formatting it so that they go red to green. So like all of the Fs are red and all of the A's are green. And in between, you get this nice curve, basically, but it's a color curve. Okay, I see.

And it's nice to be able to tell, am I looking at the same data? Am I, are my average is good? Do I have, is everybody yellow, which means everybody got exactly the same score? It's a nice way to look at data curves, basically, without having to plot everything because sometimes you don't want to plot everything. And if you had that sort of thing in a classroom sheet, you could see, oh, Amy always gets A's and B's,

because hers are always greens. And Betty has sometimes gets A's and sometimes gets F's because hers go red green, red green, maybe there's help that can be made. So I've been doing a lot of data analysis lately. and I did not realize that there was this thing called data bars, which is similar to the conditional formatting of the red greens. Okay. Except instead of making it red or green the whole square,

it fills up how much of the square should be full. So it makes a little embedded graph chart. It makes a little embedded graph chart. In your data. Oh, that's pretty clever. In the data. It's very clever. It's been really useful and yet also really not useful because I do it when I shouldn't, just to see if it's interesting.

Some of my data has chunks. Like I'm running this test with these five parameters and then this test with those same five parameters. And so I can see the patterns, which is fascinating. Useless, but fascinating. Okay, so those are the two tips everyone should know, I guess. Oh, there was another one. tell me about Merlin Bird ID.

Merlin Bird ID. Do you think there are people who don't know about it? I can't believe that, but I will tell them anyway. Okay. Merlin Bird ID is an app for your phone. I believe they have it for iOS and Android, but it uses, it's basically a bird identifier app, and it uses birds song or vocalizations

to identify the birds through around you using machine learning techniques, old machine learning techniques that have been around forever. And yeah, it's really cool. If you go for a walk, you can just turn it on. And you'll even go in the background and just record all the things it hears. And then you can look at the end of your walk

and see all the birds that were near you. Or if there's something weird you hear, you can turn it on and it'll tell you that was a great horned owl screeching or... And if you see a bird, and you don't know what it is and want to know. Yeah. You can say it was gray and smaller than a goose,

but bigger than a raven. Yeah. and then it will give you a picture, and you can choose between the pictures. Yeah. You have to download the model for your region because they can be quite sizable. So get North American birds in North America.

But it's pretty neat. And it's free. Yes. It's from it. It's Cornell University's ornithology program. And it's really fun, and it makes me feel like a complete nerd. They do ask for donation sometimes.

But yeah, and they've added some new features. Like you can keep a catalog of all the birds you've seen. It's the life list for birdwatchers. Oh, you probably have to log in for that. You don't. I don't think. Yeah, it just keeps it in the app. I know it wants me to log in, but I don't even know if I've made an account.

Yeah. Anyway, it's a fun app. It is a fun app. Okay, but we should talk about technology things, I guess. It's technology. Birds. Birds are technology, as we know. They're robots.

Did you look at Bunny's new board? A little bit, yeah. This was the bow chip and the deba bow board. Is that right? I'm reading that URL. It's a Risk 5 board. Yeah, it has some interesting features. It's a ground-up sort of design for a chip.

And one of the things I guess they really wanted to emphasize was kind of verifiability, inspectability. So as a feature I've not really heard of before, but everything's open source, but also inspectable by non-invasive sensing means, like infrared scanning. So you can look through all the layers and confirm that somebody hasn't screwed with, you've got the actual tip you thought you did, and somebody hasn't screwed with the metal layers or stuff, which is interesting. infrared in situ inspection, iris inspection, a non-destructive way to look at the silicon and confirm you've got the right chip based on the pattern of transistors on the silicon itself. Which, you know, I don't generally think about, but I guess for certain applications, that's important. Certain end users want to know what they've got. Yeah, it's a micro, it's got a Risk V stuff, it's got Raspberry Pi style,

hardware acceleration, I think, which actually runs, where the Raspberry Pi has its own kind of bespoke thing. I think this was based on Risk 5 itself, so it's got some of the peripherals are run by Risk 5 architecture, so you can generate your own bespoke peripherals in a PIO kind of way from Raspberry 325, but with Risk 5. So, interesting. A lot of cool. cool stuff. Lots of features. Yeah. Not expensive. Available from crowd supply. Second quarter, 2026 is when they start production, which is now-ish. Yeah. June 2026, chips will become available for first 3,000. So yeah, it's interesting. I applaud people thinking in different ways about some of these problems because we do tend to

build on the architectures we have and just glom on more features to them without maybe taking a clean sheet to things when sometimes a clean sheet actually enables a lot more functionality and security and other kinds of desirable things. And that iris inspection, I think if you could try it with a board you understood what it was supposed to look like, it would make it more useful to look at other boards. Yeah, yeah. can you experience with it. It's apparently fairly simple. You just need a microscope camera and an infrared LED lamp. Yeah. I want to try that. But it's going to depend on the packaging. Some will probably not permit the light to penetrate.

That makes a lot of sense. You'd have to decap it, which is just sanding it down just the right amount, like they did in the open circuits book. Let's see. Awards. Awards. Okay, so I don't know why I want an award for the show, but I do. And I've accepted that for whatever reason, deep, dark in my heart, I want an award for the show. The thing is, there are the popularity contests, which I did join one. it's the women

podcasts. I don't know. I'll put a link. They promise not to send you email. But if you could vote for us, that would be awesome. But if you don't want to, that's fine.

It was a low cost. There are some that are basically you pay for. And it's like a thousand bucks for some of them. Or you pay 300 bucks and then they keep encouraging you to enter more

categories. Like, yes, I could enter as a host, but I can also enter as a producer for just the same amount of money. But anyway, I haven't done a lot of those because, I mean, it's just buying something. And then I did start the I-Triple-E EAB Award for Outreach and Informal Education. Okay. Sounds. I'm not sure I got all that right. Sounds reasonably well aligned.

That's what I thought. It seemed like. it could be pretty similar. Of course, now that I've read more about it, as opposed to just the stuff on the outside, you have to, once I've entered, I've pretended to nominate people, which gives you more information, which is kind of irritating, but...

So what you're saying is you had to go through the nomination process of some made-up thing to see more information about how to nominate? Yes. Okay. And now I'm like, this seems to be for like museum docents. Hmm.

That seems not as well alike. Even though, I mean, it's for individual or small groups of individuals, us, who have made a significant voluntary contribution to the informal education ecosystem. Okay. Using their professional expertise to increase public engagement, awareness, and understanding of IEE fields of interest. Yeah. So, so to do that. I need a nominator and I think I found one.

Several people have offered, which is great because it's non-trivial. And I need endorsements and several people have offered. So I think I'm good on that. I-Triple-E member endorsements are better than not. So I'm good on that. And then I made up talking points for people to mention in their endorsements if they want to and pointed out shows that might be good to talk about. And then I had I-Triple-E write the rest.

No, I'm sorry. I had a, I don't even think I was using chat GPT. I had some AI thing write the rest. Mm-hmm. And again, it was going through the nomination for myself. And the truth is, I think I had to, because there's no way I would write such nice things about us. Okay. Did you know that we have created a persistent.

informal curriculum. Curriculation is a great word. Curriculum. With a library of over 500 episodes. I did know that. They tackle the most difficult middle ground concepts in engineering. Debatable.

We humanize the profession and foster retention. Fine. We bridge industry, ethics, and safety. It really does sound like AIRota. I'm sorry. There's no way. way I would have managed to say such nice things about us. The thing that they actually wrote that I wanted to talk about, sorry, I'm not going to read this to you because it is kind of irritating and I need to go through and write it properly.

You could go look at our survey. People have a lot of pull quotes from that. There were, yes. But the thing that the clanker identified that I hadn't was mentorship at scale. Hmm. What does that mean? The show intends to, and I think actually does, provide the sort of mentorship I wish I had. Even if it wasn't direct one-on-one, it is mentorship at scale. It's providing the information of what the industry is like and talking about the good and the bad and showing the differences, different paths that are available,

both in and out of the industry. It's a very technology way of phrasing that, mentorship at scale. I'm not wild about that phrase, but I do get the concept, and I think that's the nice concept. I really liked the phrase. I don't know why. I guess I just, it sounds marketing. Well, yes, that's what all of the awards are, marketing.

Yeah, but it sounds tech marketing. It's to ICCA. Who else would you... It sounds VC marketing. VC people are always saying scale, this, scale that, yeah. They're always about scales. I think there must be snakes. I didn't say that on the podcast.

You can't prove anything. Our transcripts are done by AI and could be mistaken at any time. They're also done by humans, so they're corrected. So what I just said is not actually accurate. Let's see. Thank you to the Patreon folks for having such a nice slack that they're part of. and also to Nathan Jones in particular for actually doing the reading on our book club

whereas the rest of us are like weeks behind. Nathan is out there talking about the cool parts of the book, talking about the parts he disagrees with, and somebody soon, might be me, will talk back and say, yeah, I think you're right or no, Nathan, I'm sorry. I totally disagree with what you just said. What's the book this time?

Pragmatic Programmer. Okay. What's the shtick of Pragmatic Programmer, a book I have not read? Well, it's actually the 20th anniversary edition. I'll get to it soon, I guess. And I read the original when it came out. And it's funny how different they are. And how much I dislike the differences.

Well, I mean, the industry has changed. Pragmatic programmer. Git didn't even exist. Yeah, and they talk about source control, too. It's pretty funny. Did it? I guess it did. The idea is what is the easiest way, the best way, best and easiest way, the most pragmatic way to get your job done?

It doesn't talk about AI, so we're just going to put that off to the side. AI is the furthest thing from my mind right now. I know. It usually is. Unless you'd look really angry in the night, I know. No, I'm doing great. I'm doing great. I'm making my peace with it.

So the thing I remembered most from pragmatic programming, pragmatic programmer the first time was the dry principle that don't repeat yourself. It was a very tactical, if you have written this line of code three times, you should have made a function for it. Well, not a line of code, but these four lines of code. And if you have only made one change, one typo change.

You've broken it in four places. You've broken it in four places. So let's just do this. And if you need a parameter, make a parameter, it's... That is the first thing I look for in code reviews. Yeah. Has been for a long time. It's like, why?

Why? Why? Why is this block of code everywhere? You're going to make a mistake. And we're all going to suffer for it. So, but it's advice like that. And I liked that piece of advice because it was very tight. It was very, I am coding right now. What should I be looking for?

Okay. They had some other things like you should use version control, which now it's like, of course you should. But there are still people who believe that. I'm changing my mind on version control. I think it's a mistake. I think we should. Shared drives are not version control. I think we should live with the ephemeral.

Did I say that to you recently? I know I thought about it recently. No, I just said that. All right. Apparently we're just in the same spot. Okay, so is the, so is it just kind of a compendium of best practices or is there kind of deeper how to approach thinking about being a programmer? There's some how to approach thinking about being a programmer. It's programmer programmer, which is interesting choice of words.

No, I'm not even sure that's true. Well, it must be because it's alliterative. Oh, it's definitely, yeah. I think that's an interesting choice of words because programmer to me feels like the minimum code a technologist, the person who is the implementer, the implementer, not necessarily the designer, not necessarily the architect, not necessarily even an engineer, the programmer. I know that word is weird and it just gets applied in various places, but we don't hear programmer that much anymore. like I'm a computer programmer is not something anybody says. That's what I would have said when I was a kid or at an earlier stage of my career,

but it's not something you hear anymore. It's kind of archaic. Whereas I hear software engineer or software developer. And, yeah. Because there are many countries you can't say software engineer. Because the engineer word is protected. I thought because of the gangs, the engineer gangs. The ganges.

Yeah, yeah. No, it's just a weird thought I had because it's a weird thought I had. that's not a word that you hear much anymore outside of the word programming, the act of the final act of, you know, implementing something. It talks about architecture and continuous learning and quality. It takes a hard line towards personal responsibility. If there's something wrong in the code, does it really matter whose fault it is? Or should you clean it up because it needs to get cleaned up?

Well, you know, there are commands called Git Blame, so I think that's not helping. It's not Git, it's not get, we're all on this together. Right. So I think it goes beyond programmer. Yeah, okay. But, you know, pragmatic software developer has zero alliteration. Zero. The standard software developer, the self-conscious software developer. Whereas I would have tried to alliterate it with SSDD and had that be the acronym.

Super. Same stuff, different day. Oh, okay. Okay, well, is it a book that's interesting for senior people or do senior people just kind of know this stuff? Well, if you don't know it, you should. And it's got no arc if you wanted to read. It's broken into topics that are like a page and a half each. So if you want to read this every time you get a cup of coffee, you could.

So it's like a calendar. Yeah. Okay. I think there are a few things that you should read one before another, but nothing big. deal. If you just want to flip through, that's fine. But like I said, I, I am weeks behind. I was, I was on track, but then I kind of got busy at work and fell off. And I'm going to declare bankruptcy and start with the next chapters. There was another big topic you wanted to talk about today. Yeah, large company politics. Oh, God. Well, uh, okay. So, on a job posting,

recently, there was the desired qualifications or some section talked about wanting the applicant to be familiar with large company politics. Okay. And someone asked me what that meant. What that meant. The person who asked, was somebody who has mostly worked at a company that tends to have two or three developers on a team? They go through projects. They get finished. It's mostly software.

There's some hardware. But this hardware is often similar. And so it's a design house, so things are just kind of flowing through. We're trying to grow this in a certain way. We're trying to go IPO. We're trying to broaden our product portfolio. There's a lot of stuff that happens in non-consulting companies that doesn't happen in sort of design houses.

True. Although you can get the same. Sure. same environment in smaller companies. It is very different than like LeapFrog where we did have some small teams, but they were all very interdisciplinary. You had a mechanical engineer who did the mechanical part. You had the manufacturing person who would come in and tear apart the mechanicals to make them

cheaper, the electrical person and the software person and the content person and the QA person. And so it was a team, but we all had our roles, and they were all very well-defined roles. And occasionally, you'd get some overlap. Like the QA person would come and say, can I automate this? And I'd be like, yes, please. Or I would go back to the double E and say, can we add these things? Because it will make it easier for the software and it will respond better. So as you grow teams, how does that work?

and then as you grow even bigger teams so that you can't have just one of each type of engineering, how do you organize that? And where do you go from big company politics to just process? When you say politics, what do you mean by that? What do you think the ad meant? And what do you mean by that? Getting things done.

the methodologies associated with getting things done. Not the processes. That sounds like processes to me, whereas politics sounds like we have departments with their own agendas that need to negotiate, compromise, develop their own power bases and compete, and you need to be able to navigate that sort of space. That's what that means to me.

Well, I mean, okay, maybe I should have said getting things done for all the things that aren't written down how to get things done. I mean, if you wanted to launch a new product at HP, it wasn't like you just went to your boss and said, I have this great idea, it'll make us millions. There was a whole process. And part of that process was finding a product engineer who either liked you enough or really thought it was a good idea enough. to put their effort into it. Okay. And there's also, you know, fiefdom kind of stuff where maybe the QA team, maybe the electrical

engineering team is headed by, I don't know, somebody named Dan Smith. Sorry, Dan Smith, if you exist. And, you know, they have 50 people and they like their people and they want to protect their budget. And software wants to hire 10 more people, but that's going to impinge on double E's budget, but the actual overall arc of the company is it needs more software people, but now you have this internal conflict. That's the sort of thing that I think of as politics a lot at the time is, oh, I'm protecting my section at the company, maybe at the expense of company's success overall. One of the things that I brought up was that the terms of success for a small company are simpler than the terms for success of a large company.

Yeah, certainly. A small company, your cycle time is hopefully faster and you know if you're selling things. A large company, sometimes the sales pipeline is very different and sometimes it's much longer and you don't really see what's happening or it's a sales pipeline to inside the company. So you really... Yeah, sure. Yeah, there's a lot that goes into it. and the fiefdoms are part of it because the actions and desired results are not as attached

as they might be. If everybody was at the company to make it so that the company made more money, if that was the sole requirement of everybody at the company, everybody's going in the same direction, they're yoked to the same chain, and there might be differences on how we make the most amount of money, but our goal is to make the most amount of money. that is very different than what you're talking about with fiefdoms, where, yeah, I want the company to do well, but, you know, the company is big enough.

It doesn't really matter. What matters to me is my power or making sure my people are safe. And even if that's at the expense of the company. Yeah. So, I mean, would you rather be, some people would rather be the vice president of a middling company than the director of a successful company. and therefore, you know, that will dictate how their politics within the company operate. Yes.

And so how do you deal with large company politics? I wasn't even ready to talk about that. I was just ready to talk about identifying what kind of politics might exist. Right. What is a large company? How does a large company differ? And therefore, what politics might come out of that, right? Right.

And if you think about, you know, you have a small team. You have a startup. You're like five people. And that includes a marketing person and engineer, somebody going out to get funding. I thought you were going to say donuts. Sorry, bad experience with donuts this morning. It's not our fault. I forgot in my five people.

Somebody to go get funding. But, yes, there's small team. Small team. And then they get bigger and they're successful. And now they have two products. Yeah. And so they have not quite twice as many people. And then they have enough people that HR becomes a problem.

They need to be able to manage all of this. Instead of one person taking time out of their normal job to onboard somebody, now it's going to be somebody's job to onboard people and to take care of benefits and all of the little errors that crop up. Bureaucracy. And then now we need somebody who is in charge of security. It's not going to be somebody's job. It's not going to be somebody's part-time job to be in charge of security.

Somebody is going to be the security SAR. And now we have 50 people. And then we have 100 people. And we have maybe six products in two different product lines. And when we were five people, we all knew what the other person is working on. But now there were 100 person, there are people in the company. that I don't think I've ever met. Yeah.

And I don't know what they do. And I really have no idea what the accounting people do or why they said we couldn't write in purple ink. That's just crazy. But why would I spend all of my time making fun of that, fighting that? Because I never wrote in purple ink to start with. So all of these little weird processes get written down for what seemed like no reason. And, okay, so 100 is still really. really manageable. You can meet 100 people. You can know what everybody does. And if you're the

CEO, you do know what everybody does. Or you should. But at a thousand people. And remember, this is still a small company. A thousand people is still a pretty small company. You're not talking HP, IBM. HP shouldn't be on my list anymore. I'm sorry, I still think of old HP. You're not talking Google or Apple. Yeah, they have tens of thousands of people. At a thousand people, the person at the top can't know what everybody does. It would take all of their time and they would not get to do anything else. Even at 100, the person at the top can't know what everybody does to any big degree.

They have to rely on their managers and middle managers. And then if you get to, you know, 5,000, you're in different divisions, you're in different, you don't work together, you, there are people you will really honestly never meet in the company. And everything goes through channels. You have your management channel. And whether that's an interdisciplinary team or whether that's a software team that works with an EE team. And both of those are 20 people each. and you work with your software team, you don't work with other people. Outside your software team.

This is such a different concept. Yeah, siloing and stuff like that. Exactly, the siloing. And then there's the management becomes different. You can start talking about matrix management where you have, where before you had a manager who told you what to do, and then maybe once a year told you if you were doing it well. Now you can maybe have a project manager who tells you what you're supposed to be doing because they're trying to get a schedule of everything to come out at the right time.

And you may have a people manager who is separate, who tells you if you're doing a good job by talking to your project manager and to your coworkers. Yeah. And so that person is in charge of making sure you're happy and you're educated and you aren't screwing up too much and they write your end of year review, but they don't tell you what to do. The program manager tells you what to do. And they don't even define what the project is.

They're mostly just there to make sure that you're a good human being and continue to stay working. And then there's the program manager who is the person who sets the direction for what this whole subsystem is going to do and they don't necessarily set the schedules because they don't know how all of cogs go together,

all the people work together. But they're in charge of staying fixed on, we are going to build this and not letting anybody tell them, oh, no, you can't do that. Because, of course, everybody says you can't do that. As you get into a larger company, you get a lot more support in some ways

because there are other software engineers around. Going from Agilent to Crospo, where at Agilent, I was in a team of software engineers. I worked with a team of electrical engineers, and I knew most of those teams, but I didn't know the mechanical engineers very well, other than they always knew the best places to go to lunch. And I didn't know, like, I knew occasional people,

but not, I didn't know most of the groups. Yeah. where I was the only software person. We had an algorithms person who could compile things and wrote the entire algorithm in an interrupt handler. The entire thing in an interrupt handler, I tell you. What more do you need? If that's all you're doing, I'm just kidding.

Yeah, so there was no support. They asked me, should we be using version control? And I'm like, I don't even understand the question. I don't understand the question because the answer is so obviously yes. Should we use electricity for this product? So, yeah, it is different. And if you have never been in both, it's hard to see the huge differences.

Well, and the thing that's, you can speak in generalities like we are. Right. But every company's different. Every company's got different people with different agendas. different structures, different theories and practices and ways of going about doing all of this bureaucracy stuff. Some of it boils down to kind of the same stuff. You have the same terms, project manager, director, division, first line manager. But sometimes they have lines of business. Sometimes they're organized along lines of business. And that's a little bit different than

organized by discipline. Line of business might be, oh, we have the router division and we have the switch division and they are divided up. in the same groups. There's the EE team for the routers division, the EE team for the switch division, and we make kind of similar stuff, but we have duplication of the entire company throughout that structure. And there's a lot of companies organized that way, where it's not just there's the electrical engineering department. It's you have five electrical engineering departments. And wow, what happens when you have a cross-disciplinary project across lines of business? Now you have two EE departments that are, yeah, it can get very complicated in large companies where it's not just about who's doing what. It's who's doing what for what product line.

Right. And there does tend to be process. More processes you get to be a larger company. But different divisions, different lines of business may have different processes. Yeah. And so when the router and the switch divisions, you know, try to work together. Right. And they say, we have to follow the process.

they're each talking about their own process, and there's only some overlap. Yeah. So I wanted to demonstrate. I wanted to better explain this. I mean, it's fine all in words. But it's fine to say this is my experience, but I wanted to be able to point to and say, you can go look at different experiences. Because it's true.

It's not going to be the same. for everybody. And there's books, you know, there's books histories of various companies where you can get a sense for how each company operated. I mean, HP has a lot of books about its past, and Apple certainly does, Microsoft, so Google. All of those giant companies, you know, you can read about their mid-sized days when they were large, but not enormous companies and get a sense for how these sorts of things. I mean, I know the HP way, but did the other companies have?

Yeah, I can't think of them off the top of my head, but there's definitely been books about Apple over the years, either written by people who worked there. I think Guy Kawasaki wrote one. I know there's stuff about the history of Google and stuff. Okay. I wanted to point to something and say, it's like that. And what I came up with was to talk about open source. Okay. Open source projects are pretty similar to standard engineering projects, at least in some ways.

Can be very large. One of the big differences is that open source projects need to entice developers sometimes. It may not be, many of the developers may not be being paid by the project. And when they're not paying you, it's usually because you're doing it for fun or passion or giving back. But the management aspects, which I realized that in open source, it's called the governance models, they can be really similar. Yeah. And then I went on Patreon, or I went on our Slack and asked a few questions and then possibly poorly put together the results. There's a company called Valitudo.

Valitudo? Valitudo. Okay, we'll go with that. It's a robot vacuum software. Okay. It is... Open source. Open source.

Okay. It seems to run on different robot vacuums. Specifics are not important here. I looked at the contributing guidelines because that's how I was going to be able to look at their governance model. And my takeaway from what was written on their pages was that this is a project run by a benevolent tyrant who does what he wants. And it's a solo project. He may take contributions, but only if he feels like it. Yeah, and there's plenty of open source projects that are organized that way.

Why not? I mean, if I'm going to have an open source project where I put my heart and soul into it, I don't necessarily want to play with other people. I can think of one large example that operated that way for quite a long time. It starts with a L and ends with Linux. Exactly. Part of this whole idea with the open source was about Linux. Because they do have a governance model and it's strict. Yeah.

You don't just randomly put things in the kernel. No, I can assure you you do not. And even 13 years ago, you do not. And even if you are passionate about it and are excited about it and have written a feature you really care about. They can say, please stop bothering as kid. Because they're huge.

And they have a methodology for getting things in place. And part of that methodology is asking permission first through RFCs proposals. Okay, so there was Valitudo, which benevolent tyrant. Ken... It wasn't Kenneth. It was Keith. Keith suggested ESP Home. which seems to be a bunch of people make a hub system for ESP devices.

Okay. So if I wanted to make my own light controllers, this seems to be one of the things that would make that a lot easier. And it seemed like it was a smaller group and a lot of people were willing to help you. Okay. The chances are that if you were a contributor, you were part of the community. Yeah. You know, you weren't. I'm a user, not just a developer.

Right. Or I'm a developer not just a, yeah. Either way. Both. But this is a project I'm interested in because I want to use it. Yes. And so you're passionate about the thing. And you're talking to other people who are passionate about it.

Everybody wants to succeed. Everybody wants everybody else to succeed. It's basically a brilliant co-op. And then this is where I might have gone wrong. I read the Zephyrs contributing guidelines. Remember, I'm using the contributing guidelines as a way to look at the governance model here. Did you compare them to Linuxes? I wonder if they just copied Linuxes.

I did not compare them to Linuxes. They did not copy them directly. Okay. I didn't do it side by side to her. Yeah, that's fair. But I have read Linuxes more recently enough that these were different. These were smaller. Zephyrs were smaller.

It shows a much different big company-like system. There are four or five documents they want you to read, are mandated reads before you start messing with the code, which is fair. They don't want to help everybody who wants to randomly change code. And I don't want them to accept code that is written by somebody who may be trying to obfuscate a hole. Or just, you know, doing something that's not in line with the goals of the project because the project exists as its own thing. that's not your personal. Oh, I would like Zephyr to have built-in support for, you know, I don't know, something silly I can't think of right now.

Snake, I want a built-in snake game from Nokia that can run on every microcontroller that's just there by default, you know, whatever. That's a great idea, actually. I'm going to go write that up. Based on their contributing guidelines, they have been burned by AI code. Yeah. And maybe by people trying to submit not great things. And I say that because most processes are created in reaction to somebody trying to do something.

A lot of laws that work that way too. Oops, let's try to ban this or clarify how this should be done. So if I see something, yeah, clarifying how you submit AI code, I'm going to assume that somebody tried to submit AI code with a license. And since they specifically say AI code cannot be made by a human and humans have licenses, therefore you can't submit AI code with a license. I'm going to assume somebody tried. The Zephyr people on their site, on their Discord,

might be brilliantly welcoming and super nice. Their contributing guidelines page made it seem like if you wanted to do much more than simplify the code or add a few comments, you should write an RFC plan and you should be prepared to work for it. They expected a level of self-sufficiency and professionalism that wasn't in the other two I talked about, the ESP Home and the VALATUO.

Now, I think both of those would accept it. They would be happy if you were very, here's everything that you need, here's how to debug it, here's how to test it. But they would also handhold the newbies. Oh, no, I'm sorry. ESP Home might handhold the newbies. Valitudo would have just given you two and been done.

Do you think that's valid? Do you think it's valid to say, okay, I can tell you as much as I can about big company politics, but one thing you can go look at are open source projects because they are all kinds of models. Some of them are one person who's in charge, and most of the stuff goes through them, and people complain about how they're the only person in charge, and yet their vision gets done. It depends on the goal. I mean, of course, the easiest way to learn about all this is to live through it. It takes a long time, though. It does take a long time. And so if you want beyond, you know, monographs about here's, I worked at this company between 1998 and 2004 and here's how it was developed and I, you know, whatever, that's helpful. But if you want to see something live in action, then yeah, I guess open source is good because they often have, they're often governed on mailing lists still. So you can see. sign up for the mailing list and see how people are community.

Go look at the archives and see how people communicate about issues. And, you know, you can look at, they're organized by, like for Linux, they're organized by group. There's a networking mailing list. There's this subsystem mailing list. And that's a good example of siloing. Yes. And you can see how people communicate and who exerts power over certain things.

Not sure how much time you would want to spend on that because they do. tend to be idiosyncratic and personality driven once they're at that level. And it's not exactly like a company because you don't have management. You don't have people who are, it's more gatekeeping than directing a lot of the time. They're taking a, they're kind of corraling a group of people who want to make contributions, not dictating you need to be working on these contributions for the goals of the company. There's still, there's plans and stuff. that there's routing steering groups and stuff,

but that exists on a different level than the day-to-day mailing list I'm talking about. So you may not have visibility into the Linux steering group or something like that, except for their own publications. I mean, if I worked at Google as an engineer, I wouldn't have any insight into a division level. Division-level plans. You might hear stuff, but yeah. They have all-hand meetings where they kind of talk about stuff.

Yes, but it would be. on the order of, it's the information that trickles down and is... Right, but I'm just saying it's not, you're not getting a special insight necessarily. There's some of the open source stuff, but it is a live thing that you can go look at. One of the things with being able to do a cross-section study of the open source governance models is that they are very different. If the same, so Zephyr had a lot of well written and thought out,

here's how you contribute guidelines. And so, you know, I got an impression from that. And if that, if that same level of, I don't want to think difficulty, level of process was in somebody's, somebody's origami generation system, I wouldn't contribute to the origami generation system because it's too much work.

But I might do that in Zephyr because either my company is paying me, or I believe it is important enough. So, yeah, if anybody has other open source governance models or thinks that this is a good idea to noodle on or a terrible idea to have noodle on. Let me know. I'm still thinking about

how to explain big company politics and I don't think I've gotten there. And then the question becomes, do I really want to bother? But that's a separate question. We have time for... Time for whatever.

Listener question. Andreas has asked us this multiple times. Let's do it. What do you think of using Ethernet to connect MCUs with each other and with an embedded Linux device.

Hardware cost versus ease of debugging versus ease of development. Let's invent a system because it's too hard to think in general realities here. Let's have a robot and let's have one MCU in charge of walking, one MCU in charge of emoting,

and one MCU in charge of data collection. I think it depends on your requirements. So Ethernet is very fast, can be very fast, but there's a lot of machinery that has to happen with the Internet. You have to have the physical termination, which is a big connector, and requires special termination stuff. You have to have the software to support it,

which is an IP stack all the way up, unless you want to do something kind of, Ethernet layer and build your own, but that's a waste of time. It's always been a cutoff from me. If you want Ethernet, you get an operating system. You get TCPIP, which means an operating system. There's the physical reality of it, which is cabling, unless you're doing something nifty that I don't understand.

Yeah, it's, so it depends on your requirements. Like, how fast do you need these things to talk to you? And there's other considerations with Ethernet, too, that come along with Internet protocol being designed for the internet. So maybe you have collision stuff happening or latency or things that are happening in the TCPIP stack that affect your real-time behavior that you're going to need to tune and figure out that perhaps a stream of data over spy or something else that's quite simple that isn't required for you to think about.

And you do probably have to have some sort of switching. So you have to embed, you know, if you have enough things in there, they're going to have to go through a switch or a hub at the very least. I mean, you could have it have a single brain, which then becomes the router, which why did you invent the router? Congratulations. Oh, you don't need a router for any of this, but. But if you had things that need to go through other,

like if you weren't connecting them all so that they all can see each other. Oh, now you want to relay things? Well, yeah, okay, now you're going to store and forward packets, which requires buffering, which increases latency. Yeah, I always come back with questions like this, like, what are you trying to do? Yeah, so this is a little too general. I mean... Do you need 100 megabits or a gigabit between things? Okay. And now does your MCU, is your MCU even capable of driving that? So, so let me take it as a yes, we should. there are times you should when you have multiple things that are very discrete, very modular,

thank you, that's the word. You can test it entirely. Oh, yeah, yeah. And you don't need it to be in the rest of the system. That's a good candidate. I also am wondering, no, why Ethernet, why not Can? Right, that goes to the requirements. Like, okay, is Can not sufficient for your banter?

with needs. Can's a little simpler, usually. And then you're like, okay, well, why not ice squirtsie? But never ice squirts. Now you've gone too far because ice squirtsey really is. Never ice squirtzy. Well, and the other question is like, okay, am I using off the shelf component? Is this a bunch of raspberry pies, which have Ethernet on them already. And that might be cheaper than rolling your own board. Obviously, just do that. Obviously, just do that. But if you're rolling your own board and you're choosing an MCU and stuff, I add a lot of cost to the MCU as well as the hard bar. I tend to think that Ethernet is a big hammer for a lot of projects. And if what you have are a bunch of nails, then, yeah, sure, use that hammer.

And I'm not super familiar with the power requirements. I tend to think it's a little bit more greedy than... It is a little. No, it doesn't mean you can't use, like, twisted pair for wiring if that's convenient for something else. So that gets down to the physical versus the software kind of software and electronics question is like, Do you want Ethernet because of cabling? Okay, but you can put anything you want over 10 base T or whatever, Cat 5, almost. So let me sum up.

Ethernet cost is usually higher. The ease of debugging. That's pretty good. That actually does get a plus on that one, because being able to stick something else in there and query. everything or sniffing everything. And it's very robust to noise and things like that. So that's another consideration that you don't have with just bare cereal connections

going place to place. And ease of development. And I think that's a wash. It can be easy. It can be quite easy. And depending on what you needed to do. Sprinkle some raspberry pies around. That's easy.

But if you've got to react in a microsecond to something one module saying to another, that's when you start to have to think about things. with Ethernet. All right. Andreas also asks, what is your experience with Renaissance MCUs? Have I used a renaissance MCU?

I don't think I have. I have. Okay. What's your experience with Renaissance? MCUs? Their documentation was weak. Weaker than NXPs.

Much weaker than STMs. Weaker than NXPs worries me. Yeah, it should. And at the time, they had an eclipse-based IDE that was... Eclipse. On the low end of effort put in to make it easy. But that's been more than 10 years. So this is all pretty speculative.

Did they buy somebody? I don't remember. I have never used them, so I can't comment. Yeah. Sorry. We don't know that one. Let's see. I have...

We don't have any sponsors. right now. We need sponsors. We need sponsors. If you want to sponsor us, hit me up or contact sponsorship at impeda.fm. And we are overdue for a special episode for the Patreon members. So we will. We just celebrated a wedding anniversary. I tried to convince him to talk about that because it was kind of our wedding was kind of a disaster. It was fun, but it was a disaster. It doesn't seem appropriate. So anyway, we're cooking on that. So that'll be in the next little bit. And? Let's see.

Other things on my list, talking about having a library at your company and what technical books are useful, but I don't really think we need to talk about that. And the plucky cards. The plucky cards. The plucky cards. What's a plucky card?

Mark Omo suggested these cards, and they're like super large playing cards, and they have different questions on them. The idea is that if you get the management pack, then during a one-on-one, you hold up your cards like you're playing Go Fish. Okay. And the person chooses one.

Okay. unfulfilled. Who has the wrong amount of work? And how can you tell? I have the wrong of work because I always feel tired. I think that your dog has the wrong amount of work because she's currently exploring the room instead of napping in her bed like she's supposed to. Maybe she'll get stuck in a bass drum. I think I have the ones that, I think I picked up the ones that are four managers. I think if somebody took those seriously, they could be quite useful. And there's one for mentoring. There's one for I think managing managers, which is the one that I picked up this morning. But it was a, you know, a lot of people are like, what do I talk about in a one-on-one?

These cards answer that. That's nice. And they are in different sections. So you can, you know, you can not just choose randomly. You can say, okay, well, this week we should talk about your teams. Next week we should talk about leadership skills. So yeah, I was kind of amused by them. I haven't figured out how to change this into a more fun game, though. You know, everything can be turned into a drinking game somehow. And when I tried to play solitaire with them, they were just not great. I couldn't get them to line up at all.

Yeah. All right. All right. Well, that's a solid show. The dog has already gotten up, so she's decided we're done. I don't know that we've talked much about our new dog. It's been more than a year now. It's been two and a half years.

It's been two and a half years. It's been two and a half years. Yeah. Jojo is the first dog we've gotten who really, truly honestly likes us. Yeah, it's suspicious. The thing about getting a beagle is they think you're dumb because you're not rolling. Well, she thinks we're dumb for not petting her constantly. I don't think she thinks we're dumb.

She just thinks we're irresponsible. Okay. Well, that's an improvement, I guess. Thank you for listening to this show all the way to this point. Why did you continue? But thank you for doing it anyway. Thank you to our Patreon supporters for their questions and their support and reading the book club and just generally having good ideas and discussions. Thank you to Christopher for producing and co-hosting.

Thank you to Those of you who have emailed in If you would like to email us It's show at embedded.fm And I do enjoy reading those Okay All right, that's it Let's talk about poo

We need to poo. Like the last episode Different kind of poo I tried so hard Not to make jokes It was so hard. When we last saw Winning the Pooh, they were, I believe, going on an expedition to the North Pole. Oh, said Pooh again.

What is the North Pole? he asked. It's just a thing you discover, said Christopher Robin carelessly, not being quite sure himself. Oh, I see, said Pooh. Are bears any good at discovering it? Of course they are. and Rabbit and Kanka and all of you. It's an expedition. That's what an expedition means.

It's a long line of everybody. You better tell the others to get ready, while I see if my gun's all right, and we must bring provisions. Bring what? Things to eat. Oh, said Pooh happily. I thought you said provisions.

I'll go and tell them. And he's dumped off. The first person he met was Rabbit. Hello, Rabbit. He said. Is that you? Let's pretend it isn't, said Rabbit, and see what happens. I've got a message for you.

I'll give it to him. We're all going on an expedition with Christopher Robin. What is it when we're on it? A sort of boat, I think, said Pooh. Oh, that sort. Yes. And we're going to discover a pole or something, or was it a mole? Anyhow, we're going to discover it.

We are, are we? said Rabbit. it? Yes. And we've got to bring po-po-po-bo things to eat with us in case we want to eat them. Now I'm going down to piglets. Tell Kanga, would you?