How'd you like to listen to dot NetRocks with no ads? Easy? Become a patron for just five dollars a month. You get access to a private RSS feed where all the shows have no ads. Twenty dollars a month, we'll get you that and a special dot net Rocks patron mug. Sign up now at Patreon dot dot NetRocks dot com. Hey, welcome back to dot net Rocks. I'm Carl Franklin and Amateury Cavil and Dylan Beatty is with us. You can you can jump in anytime Dylan during this.
Hello friends, tell people listening to us on the internet, how are you all doing out there?
We will give you a formal introduction after all this stuff that we do at the beginning, starting with the show number year. So this is episode two thousand and two and two let's talk about what happened in two thousand and two.
Congratulations on two thousand episodes, by the way, thanks buddy, thank you. That is a significant milestone. Well done.
And this is a kind of a strange numbers show because two thousand and two is the year of the first dot net rock.
Yeah, we are fully entering recursion now.
And it dawns on me that we really only have twenty four more of these. What happened in the year.
Yeah, well, it's because there's certain group people are like, please, don't do this anymore, and it's like, you know it'll end.
Guarantee you.
We're not going to predict the future, right.
Come on, you got to do that. You got to be like, listen, this is on predictions for twenty thirty two.
Here you go.
But a couple of other things happened in two thousand and two, like the euro Yeah, congratulations, Dylan, that affects you personally.
Not anymore.
Oh yeah, that's right, not anymore. Where are you now?
We have no right now. I'm in the UK, I'm a home and a break from tradition, I'm.
In my house. Yeah, okay, great.
And I do have a lot of euros, a whole lot. I during the whole kind of COVID lockdowns, I sorted out all the loose change that came back from all the trips, and I had one hundred and eighty seven euros in cash, about half of it in coins wow coin. So I went to Tallinn and I went to the Depeche Mode cafe and I spent eighty five euros and coins playing pinball.
Awesome, glorious day. It's almost as great as sleeping to ten thirty. There you go, all right, So other things. The US went into Afghanistan in March, some other warry type things.
They didn't know they'd stayed for twenty years.
Yeah, well they figured. You know, hey, the Russians left tail between their legs. Yeah, went so well for them. East Timor gained independence. There were some riots in India in Gujarat. The International Criminal Court was established. Let's see the stars outbreak begins in China in Guangdong.
Yeah, well, we won't figure that out till later. Right.
Switzerland joined the UN. Yeah, how about that? A Winter Olympics Salt Lake Winter Olympics are held in Salt Lake City. Brazil wins the FIFA World Cup in South Korea and Japan. Hollie Barry became the first black woman to win Academy Award for Best Actress. It took her till two thousand and two.
It's kind of nuts.
Do you know what she did with that Academy award is she took it with her when she went in person to collect her Golden Raspberry Award for Catwoman, and she walks out on the stage and she gets a golden Razzie in one hand and her actual oscar in the other hand.
Legend, legend.
Yeah, Star Wars two, Attack of the Clones is released. If you hadn't had enough of jar Jar Banks, here's another.
One in episode two. That's right, and that's a feature. You see him, but he never speaks.
Tongue got cut out or something like that.
I think they figured out that that that whole voice characterization was a mistake.
Yeah, and just never had him speak. Yeah, good, I hated that guy. All right. Let's talk about the top ten movies. The Bourne Identity Signs, Chicago Ice Age, Die Another Day, Men in Black two because he didn't get enough of Men in Black and Men in Black one.
Well, I don't even remember Men in Black one.
Spider Man, Harry Potter and The Chamber of Secrets. And finally, the top grossing movie of two thousand and two is The Lord of the Rings. The Two Towers, absolutely the second one. It's awesome. Yeah, that's where Gollum really looks awesome in that one. Okay, Well, that's enough of that, Richard. Let's talk about science and tech start in space.
In space in in February of two thousand and two is the first time they used a new quest airlock without the space station there. I've been brought in in July of two thousand and one, and they're starting to assemble the space station. So there's lots of spacewalks in two thousand and two, more than there had been in any other time. They'll hold a number keep going up. There are five different Space Shuttle missions. Four of them
are all bringing stuff to the space station. The S zero trusts crew rotations, the S one trust, the P one trust like building in the building. But the one exception is in March where Columbia, who could not get to the space station because she was overweight, does go for the Hubble servicing mission, which also had five spacewalks in it. So this is the only satellite to ever get major servicing was Hubble. Right, it's kind of a
maging what they did. So five spacewalks of two astronauts each spending something like forty hours of total work, and they put in new solar panels, a new power system, and to set a new sensor. It's just a complete upgrade Tubble.
You know, I recently saw online picture of Neptune that was taken with Hubble next to a picture of Neptune that was taken with a JW. James Webb, you know, James Webb telescope, and the difference that you can see rings around Neptune.
Oh sure, well Hubble. Hubble is optical and James Webb is infra red, so you're seeing different frequencies and they're colorizing it to make them look similar. But it just shows that not the optical range is not always the most necessary. And as much as we all adore Hubble, and it's definitely on its last legs. Again, there's a whole debate about whether it's worth servicing it or not.
The Nancy Grace Romans satellite's about to go up, which is also in the high infra red near optical range, but has and has a similar size mirror, but such so much better electronics that will observe one hundred times more space than Hubble in the same amount of time. Wow. So just you know, texts come a long way. Also based on an old spy satellite.
Now, if we could just get a telescope that would show us what dark matter is. I think that would be us. We're trying to figure that out. That's one of Mission Nancy Grace Roman's missions, so we're working on it a couple other space things. In two thousand and two, after the Challenger disaster, the Air Force had pulled out of the Space Shuttle missions pretty much entirely, and they started what they called the New Evolved Launch Vehicle System.
And in two thousand and two both those rockets fly for the first time, the Atlas five and the Delta four, which today are both retired. This is when they began also a first flight for the new European rocket at the time, the area En five. It was so big it actually would carry two satellites at a time, which means when the first flight failed catastrophically, it lost two satellites instead of one. Wow.
And let's see what else in April SpaceX, Yeah, we'll get there. In April, we get our last messages from Pioneer ten. You know, we always talk about the Voyager missions, which are way you know now passed out of the heliopause, but Pioneer ten was actually ahead of them and launch in seventy two. The Voyagers were in seventy seven and
it was firsting to go to Jupiter. It only had one hundred and fifty five watts of RTG power, but by two thousand and one was down to sixty five wow in it radio finally packed it in while it was eighty astronomical units about twelve billion kilometers away from Jeez, and you're right. In May, this pseudo billionaire at the time named Elon Musk forms a company called SpaceX after he was unable to buy a deneeper rocket from the Russians to do his Green Mars mission. He figured he'd
just try and do it himself. Last time I looked, still hadn't flown anything to Mars.
Right.
Also in May, the Mars Odyssey orbiter does the first sensing of actual water ice deposits on Mars. We've always seen that there were white poles on Mars and the presumption that they were frozen carbon dioxidecause the air pressure is so low. But there's also water ice there as well. Yeah, Okay, over on the computing side, yeah. In August, the first episode of dot netX polishes a guy named Pat Hines, who we consider are Lucky Rabbit's foot first on most
of our shows, back all of our shows. I think yeah is out there. This is of course, it's actually a slow year for comput because it's two thousand and two. In the dot com boom has ended, the bust is on full force. It's not a lot going on. But it's also the year that the Phoenix browser is released. So this is actually the Mozilla community, which was formed by Netscape in ninety eight just before they were acquired by AOL, and a bunch of those folks built their
first built. They wanted an open source, independent browser in the legacy of Netscape, but not Netscape. They called it Phoenix, which they weren't allowed to do because there were a company called Phoenix, so they renamed it the following year two Firefox.
Well, first they renamed it the fire Bud because and then someone else said you can't use that either it's an embedded database. And there there's a fork of Firefox from way back then. Well, you could basically you could pick an element and an animal and it just changed it every time, so you could be like earth oh or ice, weeezl or.
Whatever you like.
Those are commands in Rockstar the language.
Also speaking of AOL, they had that was the year they announced that they were going to integrate AIM and ICQ.
Wow.
So the great messenger conflicts of the early two thousand. So ICQ I think was pretty much the first out in ninety six and got acquired by AOL in ninety eight, like because AOL was buying everything back then. They had made their own messaging service in nineteen ninety seven, and so now they had two. They figured, you know, five years on they try and put them together. I remember that's all back.
It does seem you used to run astonishing, you know, with the big gift of hindsight that AOL, the company who gave away the CD ROMs at the gas station, bought Time Warner. Yeah, like Warner Brothers, Time magazine, that conglomerate the gas station CD people bought them. Yeah, because that was they were the biggest hitter in that fight at the time. And the AOL Time one became a thing and then they went, I actually about the whole AOL Nah, you're good.
Oh well, so we used to use a program called trillion for messaging back. Yes, I don't know if it was after that, but it was around that time after where it had three different you know, messaging platforms.
Because the rest remember it was also yeahoo messenger and MSN messenger.
It was yeah who messenger, There was MSN, there was ICQ, there was it had an io C client. Uh go one real old school. Yeah. Trillion was great.
Yeah, yeah, crazy, Yeah, I know at first version of trillion is two thousand, so wow, it was way back there. Last, but not least, a guy named Reid Hoffman forms a company called LinkedIn. Wow, in December, not a lot going on in the hardware side, So I'll just bring up the two things that I found from two thousand and two.
Well, wait a minute before you move on to hardware. Yeah, there's one big thing you forgot. What's that dot Net was released to manufacturing in two thousand and two.
You know, you're right, it was absolutely It was after almost a year delay for security fixes.
Yeah.
So I remember going to the UK headquarters in Reading that year with a couple of the guys I was working with at the time, and just we got in the car, we drove up there. We're like, hey, what's this dot net thing? And they're like, this is c shop and this is web forms and this is wind forms. And then they gave us the Visual Studio dot Net like Preview three release on a stack of DVDs and we all took it back to work and we're like, this is the future. And yeah, it eventually caught up.
Well, I mean, and you know you're you're you're right, Richard about the end of the dot com boom, and that's why the Moniker dot net was so you know, this is the next thing, right right, yeah, all right, hardware go for it.
Only a couple one was the Handspring Trio. So back in ninety eight, they the guys who made Palm, they got acquired by three com, so they bailed, They took
their money and they ran. They didn't want to work for three com and they but they did license the Palm os that they had developed back from three Com and immediately set about to make PDAs, though Palm had their own PDAs, And in two thousand and two they released their first one was called the Trio ninety with one hundred and sixty one hundred and sixty display and
a keyboard on it but no phone. Later they'd make the Trio one eighty, which would have GSM wu WO and they would only last a year because within a year three Com would come back and buy them. So it's like we took our money from selling to you, made another thing and you bought us again. Good deal for them. But it's the PDA twitches and last but
not least, the Rumba by I robot. So this idea had been rattling around since the eighties, but in nineteen ninety nine sc Johnson kicked it a bunch of funding, I guess because of the dot com boom to build this robot. It was a vacuum that could avoid obstacles and stairs, not dog poop.
And old Hansman was the first one to post videos of his room. But he was so happy about it. So and you know, designed to go under beds and things like that. It was very low and lean and so forth. It made great cat videos for sure. Oh yeah, they're much more advanced now like current Gen ten's are impressive devices. But yeah, that's when that started. We have a rumber and Kelly has my wife has a habit
of naming everything George, so it was Rumba George. And she had a little beagle, stuffed beagle who she called puppy George. And she put Puppy George on Rumba George and took a video of it driving around the house and it was the coolest, stupid but cool. Yeah. I can't believe I wasted that thirty seconds on that little story, but I'm sorry, LRD, I apologize, all.
Right, That's what we got her instrum.
So we got two.
Two, we got we got two more. Two very important cultural milestones from two thousand and two. The first gay in the Battlefield franchise Battlefield nineteen forty two. That was two thousand and two, and I remember it because everyone I know was hooked on that game for about six months. Wow online clan play. It had the best gameplay mechanics of I think any like you know, deathmatch style online
shooter and the other one. It was the first audio Slave record, and this being dot net Rocks, I thought we should get just a little bit of rock and roll in there, because very cool. When we're not talking about Rockstar today, we're talking about something else. But yeah, Audio Slave two thousand and two one of the great kind of crossover supergroups of that generation.
So cool.
All right, So before we start talking to Dylan about what he thinks we should talk about, let's do better no framework, roll the music.
Awesome, what do you got?
Well? By now people listening to dutnet rocks this should be old news. But in case you haven't heard, you can now use your your own LLLM models and gethub co pilots cli. Whereas before you could only go through their network. Right getthub network. So now you can bring your own apikey and guess what it supports Olama, which is using the open AI protocol.
Nice.
I have a big Olama machine sitting right next to me, blinky lights, You're going local, You're going I'm going local. So I tried it. It works by Jeff Fritz and I did an episode of Code It with AI episode twenty seven where I demonstrated it, and man, is it fast? And I'm using a quin coder thirty billion library nut library model on it. Just really fast and really good. However, I found that it fell down a little bit when asked to do things that required a little more architecture
and thought and wouldn't figure those things out. But really really good. And I'll tell you a little story that I told on that show. The first time I ran it, I noticed that when after I hit the enter key on the prompt I hear this high pitched wine, babe? What the heck is that? As soon as the text came back, it stopped and look at my uninterruptible power supply that it's plugged into. And the UPS was basic equally saying, ah, I can't take that power train. Oh wow,
So yeah, they caused my UPS to freak out. And granted it's a small UPS, but yeah, plugged it right into the wall, no problems.
So you need a bigger UPS.
So now it's not about tokens or you know, requests, it's.
About power and I have Now you've got the power problem.
Yeah, well but I've got a roof full of solar panels, so I'm not at all worried about It's going to be a good summer. Let's put it that way. And you know, get get hub Co pilots raising their price in June. This is the perfect time for me to put more cycles into the local LM. Nice, that's it, all right, who's talking to us?
Richard grabbed a comment off show nineteen ninety nine and the one we called how We Beat the Wine two K Bug, Yeah, which came out really nicely, and Rob Howard had this nice comedy, says Hi, guys, great show. It was good to take a trip down memory lane and hear the war stories. My own experience was nowhere near as exciting as your guests. For my customers, I'm sure they felt it was just as important. People were worried, right,
it had been hyped up a lot. I ended up on the phone at eleven pm Australia time on New Year's Eve and that would be some of the first time zones to roll over, right, yeah, with one client talking them through a last minute emergency system wide backup. What a one hour before rollover? These people look crazy, Yeah,
because you know they wanted it just in case. I was able to getting through the process and feeling a little bit calmer, was still enough time to watch the Sydney fire works, fun times beautiful, you go, yeah, I love it, you know. It's just I'm just great to reach out remind people that we did a thing back then and get some feedback from that. So ready to hear. Thank you so much. You good to hear from Rob, Yeah for sure, so Rob, thank you so much for your comment. Copy of music co Buy is on its
way to you. If you'd like a copy of music. Co buy I write a calm on the website at dot netroocks dot com or on the facebooks. We publish every show there, and if you comment there and I read it on the show, we'll send your copy of music.
Go buy music to code by of course, Still Going Strong twenty two tracks, twenty five minutes each, designed to keep you in this state of flow while you're writing code. Many many, many happy customers. All right with that, let's introduce this guy who we've been talking to kind of casually. Dylan Beattie is a consultant software developer and international keynote speaker. He's been building web apps since the nineties and works primarily on Microsoft dot Net, hddp APIs ux design and
distributed systems. Dylan lives and works in London, and when he's not writing code, he plays guitar and writes songs. Oh yeah, he also came up with that language called rock Star, which you heard about in previous versions of the episodes of Rocks where he was here. Welcome back, Dylan, Hey, it's great to be back. Yeah, what are you up to these days?
Well, most immediately like right now, talking to you guys earlier this afternoon, I was porting a bunch of web code from next js to Astro, hanging out with Rendall, doing live streams about claud code and getting up and running with AI building the odd MCP server and dot learning about CSS and all of the amazing things that that can do now. And yeah, it's worse than the problem of focus in tech is worse than ever because the number of shiny things just keeps going up and up and up.
And you know what, the finest memory.
Have such good add tools.
You know, the finest memory I have of doing things with you, Dylan, was when we were all in Oslo. I think it was with Rendell and Richard and I and you did a sort of a show that was like a quiz show based on QI Yep, which I had never heard of. At the time, I had never heard of QI and now I've binged watched every episode. It's like my favorite British production.
There is definitely a very a very English kind of panel game that folks who have grown up here or you know, listen to a lot of BBC Radio and BBC TV stuff, are familiar with. And you know, we're always the latest one. We've been talking about is you've seen the last one laughing yes, where Yeah, So we're wondering would that work as a conference panel, Like you get a couple you obviously you've got to pick you know, people you know are going to have a good time
with it. But you have a panel of four people on stage. The audience is can ask the questions, no holdspot, and if you laugh, you're out and somebody else gets up and takes your spot.
What I liked about that show we did is it was very Hollywood Squares in that the contestants got to see the questions before beforehand and make up funny answers. Yeah right, and that it was just hysterical. Anyway, No more about that.
Good time, So go to another one of those one of these days.
In a while.
Yeah.
We could even do it online.
Yeah right.
So use what Works.
Yeah, so this is what we're going to be talking about today is a thing that a bunch of us has been working on. So we'll start with the headline. There is a website at use what works dot org where a bunch of people are swinging by and they are there is a manifesto there that you can sign, and there's about I think thirty five people now have swung by and taken a look at that, and this is has basically emerged from over the last I mean
a long time. You know, I've spent many years talking about open source and free software and licensing and just all of these kind of complexities of this idea of there are programs on the Internet that you can run without having to give anybody any money, and how, you know,
all the one how did we get there? Because this is an amazing thing that you know, we have used to build some astonishing products and platforms and services, and but once you start doing this at any significant scale, you start to realize that there are actually all kinds of different ways that these models can break down. And so this came out of a a lot of people over the last year or two have been kind of standing up and doing talks about, you know, what does
the future of open source look like? And the news stories that I think started people looking into this were the ones that end up on Reddit with people talking about an open source rug pull, and perhaps I think the most high profile one was probably reddis And you know, Reddis was a really interesting case study because they built this, you know, fantastic high performance key value store, really really useful building block for creating high performance cloud systems. And
they basically gave it away for free. And you know, you could, you could pay them in all kinds of ways for maintenance and supporting that kind of thing. But if you didn't want it, you didn't have to. And you know, to simplify the whole thing enormously, what happened is companies like Amazon turn around and go, oh, well, yeah, if you want reddis, we'll run reddis for you for
fifty bucks a month, hundred bucks a month, whatever. Just you put your master card in and here is your connection string, and there go, you've got a redesting in the cloud.
So radis is free, but if you want to integrated with our cloud system, it's going to cost you a little money.
Yes, So this is this is what reddis did in the first instance, and they weren't the only one. And you know, I'd absolutely kind of see the reasoning behind it is they go, well, look, you're you're absolutely free to integrate reddis into your own product unless the only thing your product is doing is selling our product to your customers. Doing that, you can't do that anymore.
That sort and skimming, right, Yeah, it's you know, hey, how would you like to buy this glass of water? Yeah? For five bucks? Yeah.
And eventually the people who are given out the free water are like, you've got a trillion dollars and we don't, and that's not right. And you know, I think a part of it is, hey, we'd also like a trillion dollars. But another part of it is nobody should have a trillion dollars, especially not if they're can by selling free stuff. Right, And so reddis did a two to two and a half years ago they said, right, we're changing our license. It's not completely you know, free as in free beer anymore.
And this caused a massive backlash and you know, the press and the media picked up on it, and social media picked up on it, and you know, all kinds of people. Microsoft, I believe it's Garnet. Microsoft Garnet was there. I think it was called was there? Like it's a drop in reddest compatible clone. So if you're running reddicin as you're and you don't want to pay for it,
you use this one instead. Some folks took the last reddis code base that had been released before the license change and they forked it and they went, well, hey,
we've we got this thing and this one's free. And then a year later Reddis turn around and went, actually, you know, we're going to go all the way in the other direction, and reddis is now released under the I believe it's under the GNU General Public License, which is basically the one that says, hey, you can use this, but if you touch it, you're going to show everybody what you did, right, which in a way is even more effective to trent to companies like you know, Amazon
and Google Cloud, because they're like, well, we're not going to open source our you know, metrics and dashboards and billing systems that we used to charge people for the free software, so maybe we should pay for a license. And you know, this is this is one incident there's been. We had a great session in the London Donet user group last year with Mike James from Avalonia who's talking about the same you know problem of a sustainability and
they got a team of engineers working on this. You know, incredibly powerful, incredibly capable. It's an open source drop in replacement for WPF effectively. You know, there's there's more to it than that. But if you've got a company who've got expertise and investment on WPF, and now you need to go across platform. You want to be able to run on Linux, mac Os, those kinds of things. Microsoft has never offered a really compelling migration path for desktop
applications on non Windows platforms. You know, wind Forms are still out there. Wind Forms are still kicking. It's still, i think the best UI framework they have ever shipped. But you know, WPF is a little bit long in the tooth. The Maui Wing Ui, all these different things that all kind of they were all going to be the future, and then they all weren't the future. And now there's Blazer, and if you want Blazer on the desktop,
you put a browser in an electron app. And meanwhile, you know, Mike and his team are out there with Avalonia, going well, actually we're just going to solve the problem. But it's quite expensive solving the problem, and so you should probably give us money if you're using this to run your business. And you know, there are multiple examples of this kind of thing going on.
To bring it back to what we use what works the big picture, here is the paragraph that's right right on the front page. Organizations should generally use existing, proven, and sustainable solutions to common problems rather than developing, maintaining, and running their own. In other words, use open source projects, right yeah, yeah.
Well as long as they're maintained, right, Like, yeah, I get that Microsoft Research spun off Garnet and basically gave you reddis without all licensing constraints. But are they keeping it up? Like that's always the question. We don't go in for open source software, going for maintained software.
Yeah, this actually you hearing you read that, I need to merge a pr on this because there's been some discussion about replacing existing software with established In that specific paragraph, because I think it was Udi Dehan pointed out that software can come into existence overnight and resistence is no indicator of quality but establishment, you know, community and some
continuity around it. But you know what we started doing so around probably September October last year, a whole bunch of folks working in and around open source, and a lot of them from the dotnet space because a lot of us have you know, very strong networks in that community, you know, sort of started chatting to each other and being like, should we put something together that is maybe just a little bit more structure than a whole bunch
of people giving conference talks about the same things and doing podcasts about the same things. And I came up with this idea of you know, like all the good names, we came up with about fifteen, and then we went for the one where we could get the dot org and all the social media handles.
Right, So the best name is the one you can actually get the content for us.
Yes, but you know, conversation has been around what are the what are the the first point where somebody goes down a path which is going to lead to one of these scenarios that we all think we should try to avoid as an industry, right, And so one of them is, you know, maybe counterintuitively, it's going out to companies and going you should pay money for free software.
And you know that as my time as an engineer and as a tech lead and a CTO, getting budget for free software was always you know, it was not a pleasant experience. And partly I think it because it cuts orthogonally to the reason a lot of people got into development in the first place. Is we want to
build stuff. We don't want to have budget meetings and yeah, yeah there, you know, with your team one morning and you're going through, okay, what's in the sprint and somebody's like, ayeah, authentication and you're like, well, we could find one, and oh but that one costs money so we'd have to have a budget meeting for it. Or we could use that one up, but that one's really complicated and there's a learning curve. Yeah, we could just build our own.
You know, how hard can it be? Use the name password, password reset? Okay?
And then yeah, well do you are you finding are you finding that with the in the new AI coding landscape that you know, where you have teams and agents and stuff and squads.
That more people are considering building their own stuff that they're paying for now because they have this power.
Yes, I mean it's uh, it is not my experience on this has been going for a lot longer than AI has been a viable way to deliver software. Uh, And I see, I'm gonna go on. I'm here. I see no reason to believe that AI is fundamentally going to solve all the problems that we didn't solve with structured programming, and we didn't solve with functional and we didn't solve with object orientation, and we didn't solve with scrum, and we didn't solve with that child, and we didn't
solve with canban and and and cloud. And you know, the challenges are still there, and the challenges are around, you know, the confidence in what is being deployed into production, and that's something I got to say, you know, whether whether by a coincidence or a causation causality, the uptime of some of the major providers in this space, the kind of numbers we've been trying to get away from
for decades. You know, we're talking like uptime that doesn't have any nines in anymore, because it's like eighty seven percent a lot across the past three months, and it's not just the AI platforms themselves. I think we're starting to see some kind of knock on effects of that. But you know, I think the point here is that the act of creating the first working prototype or the MVP, the minimum viable product for a piece of software that
is often relatively easy. You know, most developers I know could build a login system in a couple of days if all they care about is user name, password. Was the password right or wrong? Yeah, fine, okay, cool, and then the like, we need a password reseid, Well, that's easy, we'll bolt in a SMTP relay. We'll send an email
for that thing and it's done. And then you know the but if you get something off the shelf that does this, they're like, well, yeah, we have two factor, Yeah, we have a worth two you want login with Google, login with Microsoft, login with LinkedIn, login with whatever we got that. You know, like, what's a timing attack? You don't know what a timing attack is. And you've built your own password security all right.
Yeah, I hope you're not protecting anything important withs that, And you know the I think the difficult thing is that the you get a level of expertise.
If you have built your own security system as I have done, and you have supported it in production over a number of years as I have done, you learn a lot more than you would just getting one off the shelf and plugging it in. Sure, but the biggest lesson of all is that you should have just got one off the shelf and plugged it in.
Right, let's take a break. I think it's a good time and we'll be right back after these very important messages and we're back. It's Dot at Rocks. I'm Carl Franklin. That's my buddy Richard Campbell, Hey, and our buddy Dylan Beatdi is here talking about using what works. Hey, hey, making sure that I guess it's you know, do a cost benefit analysis.
Of well, you know, the joke here is it takes long enough to understand a problem well enough to know it's a hard problem. Yeah, you know, it's easier to just stay ignorant in some ways.
Yeah, and so you know, the the angles at which we sort of came at this whole use what works thing, you know, one of them is, let's see if we can create a critical mass of you know, case studies and supporting material for people who are about to have that conversation, you know, team leads and tech leads who are about to go to that boss or go to that CFO and be like, we need money for free software,
and here are the reasons why. And yeah, it's free, but it's not free for us because we're making fifteen million dollars a year and we should be passing, you know, paying some of that forward into the ecosystem which allows us to be here in the first place. And part of it is, you know, encouraging the developers who are working on this stuff and reinforcing. Yeah, it's great fun
building your own relational database. You'll have a blast, but you can't take it with you, and you are going to be more in demand as an engineer if you know how to get a database off the shelf and set it up and tune it and optimize it. And you know, what's your do you understand that your rotational backup strategies? And do you know that in SQL light if you just copy the dB file, if you're using right ahead locks, that's not actually going to back up
your database. And you know, those are the skills which are transferable, and it's maybe cuts to a sort of you know, an interesting is dichotomy The word it's one of those words I keep meaning to look up a lot of us folks, you know. I my degree is in computer science. They didn't teach me to use anything.
They teach me. Taught me how to build software out of raw algorithms, you know, which is a little bit like you know, expecting physicists to design apartment building gave you a pair of chromium and users to put the zeros in ones on the disk in the right or and you know the I think that the difficulty here is that a really valuable engineer is one who has reams of knowledge that they know they shouldn't apply at work until they absolutely have to. And that means pretty good.
Yeah.
One, you're going to be spending a lot of time learning stuff that you should not be then going and doing in your job.
That's right.
You know, I teach.
I teach distributed systems, And one of the things I say to folks at the beginning of day one is, look everything I'm going to show you. You do not go back to work next week and do this. This is stuff that you play with and you learn how it works, you familiarize yourself with the patterns. But also
I'm going to teach you what other the signs. It's time to think about doing this for real, right because you know, if you start doing this stuff then just just for fun or just because it looks good on your CV, you're going to leave behind a legacy of software which is way harder to maintain than it really needed to be.
And you sometimes complexity is necessary, and when it isn't, don't introduce it.
Yeah, and if you can, you know, the complexity is there,
but somebody out there will have tackled that complexity. And you know, whether you're you know, container orchestration, or you're building message queueing software, or you're building APIs or MCP serves, any of this kind of stuff, you know, the one if you're the developer, don't just be like, yeah, I'm just going to build my own, because you probably will, and you know, you may even get a nice little bonus for the amount of money you saved by not
buying it. But sooner or later that thing is going to need more, more maintenance. It'll stop being fun, and then after that it'll start being a problem. And then after that you're going to be the one who's sat there at the office at three o'clock in the morning because there's a new GDPR data retention compliance deadline coming
up and the only way to do it. So but then, you know, the other perspective on all of this is the journey that a lot of the open source maintainers themselves have been on from Hey, I made a cool thing at work and I think it might be useful. I'll put it on GitHub. Yeah, I got a new release, I got a new feature, I got pull requests, I got people are helping me. This is awesome. Yes, of course it's free. It's under the GPL or the MIT
or you know, the Apache license. And then it gets to the you know, the point, and this happens a lot of oh maybe this could make some money. I know, I'll sell maintenance contracts, all right, now I have twenty dollars. Oh but now the people who gave me twenty dollars think that I should be doing exactly what they want for their very very specific use case because they gave me twenty dollars, and that's not really how this works.
And so, you know, encouraging the second somebody out there is like, I have some code, I have a solution that I think might be useful to other people. Is that's the point where you're like, what is this going to cost in the end?
Right?
And you know, when I go public with this thing, am I just gonna yeah? Yolo GPL. Everyone loves the go New Public License because Linux is awesome, and yeah,
you know it is. But also a lot of companies with deep pockets have got it where it is today, and unless you are in touch with those people and have access to their pockets, you're gonna end up creating a situation where something which may become a you know, vital part of somebody else's infrastructure is the thing that you do for fun on your weekend, right And at that point, that's when the sustainability just becomes a real, you know, a situation. None of us, I think, wants
to be on any side of of either. You know, you're like, well, we got fifteen developers up and running on this thing last year because we thought it wasn't going to cost any money, and now it's gonna be five thousand dollars per developer per year. And that's a hard conversation to have because if you're the one who went for it because it was free, Now you're like.
Are there any success stories of popular open source packages being sponsored by companies with deep pockets that depend on them and don't want them to go away?
So there are the one that I know off the top of my head that is sort of specifically probably still here because of a single sponsorship deal is Avalonia.
Yeah. We just had them on talking about that and they've got a huge pot.
Yeah, yeah, so that they had one major corporate sponsor who was like, look, we basically we need your thing to keep going so that we can keep going, and we're happy to support that. And you know, there have been numerous examples over many, many years of companies that have done, like you know, a one time grant or a one time donation or some kind of short term sponsorship.
But sponsorship is very fickle, and the consensus that seems to be emerging slowly is that if big companies with a lot of money want to use your thing, you just need a way of selling it to them, not donation, you know, trying to get your company to donate money. A lot of them they don't have a form for that, Like there is there is.
Yeah, it's the first item on the budget that gets cut when when times are tight.
But you know, you got to somebody and like, look, here's how it works. We can put the whole team on this, probably take us six weeks to do it, and then we're on the hook for a week's worth of maintenance every three months for the rest of the life of the company. Or we have a support agreement with these people and it costs you five thousand bucks a year or whatever. It costs. Put it on the spreadsheet.
That's what it's going to cost. We're going to pay them, they're going to look after us, and that's how the thing is going to go. And you know, the other sort of interesting aspect to a lot of this is how stringent a lot of these these you know, open source projects that have now incorporated or embraced a commercial aspect. I remember one of the first ones that I know about to do this was a thing called service Stack, which went from being I think BSD licensed to being
like a hybrid license in twenty fourteen. And so that's you know, that's twelve years. That's ancient history in open source. And yeah, they switched it. They said we can't give this away for free anymore. And the guy who uh demis Bellot, who sort of created and ran it, he said, look, we're not going to send lawyers out. We know who you are, we know that you're running it without a license. We have better things to do than try to hunt
you down. But the people who are you know, doing the right thing, either through their conscience or because they're legal, team is like, ah, no, no, no, no, no, come on. That has made it a you know profitable, sustainable, there's got people on board their ship and features and they're doing great. And the interesting case. I'm always interested, you know, if something interesting is happening today, I'm like, did anyone
do this five years ago, ten years ago? Do we have any precedent for what this looks like long term other than like, look at the crystal ball and try to predict the future. And when service stack went commercial, they forked it. A bunch of people on the internet were like, well, we're going to take the last version that was BSD licensed or MIT licensed and we'll fork it because you can do that because the license says you can take this version of the code and do
whatever you like with it. You know, you can even you can turn it into a product, and you can sell it as long as you don't take our names off the license, and you know, as long as you're acknowledge authorship. And then service Kit was like, you know, three months of great excitement about yay, software should be free.
And what happened is the same thing happens every time is all the people who are serious and are prepared to put their hands in their pockets, they go with the upstream fork and all the people who think they're entitled to a free lunch. And by the way, you've got to do this and this and this and this
and this, because that's what our project needs. They go with the fork, and so very very quickly the fork will find itself inundated with support requests from people who are never going to be prepared to support, or sponsor or otherwise contribute to the maintenance of what they go.
Oh, and that's the problem, right. Sponsorship usually comes with strings attached, and companies that spend money in this way often consider themselves you know, well, you're going to do this for us because we give you money. And it could also be at the exclusion of their you know, their competition. So it gets really dicey when you start to take money. You have to if you want to keep it, you have to play by the rules.
I think there's also always been an implication with sponsorship and donations, picularly that the people providing the money are like, well, we're not paying you to use the software because we could have done that for free, right, So we're giving you money because we want to be special, we want priority, we want you to listen, right, And you know, I don't think that's necessarily premeditated. But there's a hey, this is working well for us, let's let's sling them a
couple of bucks. And then later you're like, oh, well, yeah, it would be good if it would do this, and we did give them all that money. Yeah, and so yeah, And it's like I said, the whole thing is about looking at these different perspectives on the problem of open source sustainability.
You know.
One of them is from the folks who are maybe they have a successful project out there and they're thinking, do I need to you know, switch my licensing basis or how can I get this to start paying for itself. Some of it as the folks who aren't there yet but they're thinking about releasing something. Some of it is the developers and the tech leads. Some of it is business owners and you know, encourage a more holistic understanding of the total cost of ownership of all this stuff.
And it it's getting some traction.
You know.
I think it's a it's a topic which a lot of people have been circling around for a long while now, and it's I think that you know, to me, the biggest thing that happened is open source was great when you were downloading it and running it on your three eight six and your four eight six and your penteum and then yeah, but you know, cloud came along and fundamentally, you know, moved the needle in terms of how you
make money out of software. Sure, and you know, when when you can charge people to run it, the question of who gets paid to write it becomes very very murky indeed, and that I think is the sort of the instigator of all of these questions which a lot of us out here are now grappling with and seeing what all this looks like.
A lot of questions, not a lot of answers.
Yeah, that's simple, but it's like, it's not like we didn't have enough to do. The more software that can be maintained by others, profumalarly with some level of support, the better off we are. It's always going to cost less than doing it yourself.
I mean, there's a there's a mantra which I've heard lots of people. The formulation that I enjoy, because I think I came up with it was if you can't sell it by it right, and you know, if you have a bunch of bunch of engineers on your team who could be building the thing that your company sells, or you know, it's not necessarily thing your company sells. There are folks out there working in nonprofits, you know, charities, humanitarian organizations, but there will be a reason why that
organization exists. There is a change that you want to make in the world, and if you are adding GDPR compliance to your home brewed logging system, you're probably not getting any closer to that the objective that got you all there in the first place. Yeah, and you know that's a you look at all the things that you could be building. You like, right, which ones? Which ones are the what Chris woodruff Woody, He goes, what's the
secret source? What's the thing that nobody else is doing that If we get it right, we are going to be successful. We're going to solve more problems. We're going to cure more malaria, We're gonna shift more units. Whatever it is that you're trying to do in the current financial year. What's going to move that needle?
Yep, that's going to that's going to make the biggest difference out only. Yeah, I think we always underestimate all the other work we have to do too. Right, like, it's easier to gnaw on the log in problem than it is to law on the whole problem in front of you.
It's also, you know, I will admit to this as readily as anyone else. It is tremendously good fun reinventing the wheel, that's sure, you know, sure as a as a developer, I think certainly me and I know a lot of other folks that I've worked with over the years. Our happy place is when we understand the requirements and we just get to sitt in, you know, write code and build screens and build tables and you know, create stuff.
And the best set of requirements in the world is one where someone else has already done all the hard work of figuring out what does it look like, how does it work, how long can the passwords be? What happens when you click this button? And you know, there's the number of folks out there who have you know, they've they've built Tetris as a learning exercise.
Right, well, a couple couple of shows ago, I don't know, four or five we the better know a framework was a site that had build your own X and it had all of these things for learning how to write code. These tutorials for how to build your own oh and the list was huge. Yeah, you know, build your own database was one of them, of course, And you don't do it because you want a free database at the end. You do it to learn how to write code, and by doing that you understand how these things work.
It's very cool and I think that's a you know, incredibly valuable one. It aligns strongly with the idea of code as a craft, which is getting a lot of attention at the moment because there's a you know, ground swell of opinion going out. Code doesn't matter anymore. LLLM writes it, LLLM reviews it, if it's good, goes to production, you never look at it. It doesn't matter if it's if
it's messy. And you know, the leak recently of I still haven't quite worked out did the clawed code thing actually get leaked or did they leak the prompt or
did they leak the debug tables or something. But I recall reading some commentary from people going this is not very good, and you're like, well, it is very good, because you know, I mean, they're losing more money than they're making, but they are bringing in a lot of money, and a lot of people are using it, a lot of people are very excited about what it can do. So by a you know, a comfortable set of metrics, the code here is is probably not the certainly not
a deal breaker for it. But how much better could it be if it wasn't like that? Yeah, but no the idea because the other thing, you know, it's obviously this is a show about AI. It's a show about sustainable.
Openly, all shows are about AI lately, Pan avoid the topic.
It's actually I.
Think it was you, Carl, I remember many years ago talking about Cloud said something like technology is interesting for three months when it's brand new, and then for three months when when people are making money with it, and in between it's just a lot of people getting excited
about the same thing all over again. And it feels to me like we we've had the three months where it's new, and now there's a sort of little bit of all right, let's wait and see what is going to be the groundbreaking and you know, to me, it's putting AI in things is not exciting. I want to see something fundamentally new that we couldn't have built without
these engineering capabilities and it's a yeah, interesting times. But one thing that I'll see a lot of folks talk about is that if companies go down the road of using coding assistance instead of junior developers, then we are drying up the pipeline because where are the next generation of seniors and architects going to come from? And you know, you look at the crafts like you know, furniture, instrument building, carpentry, all those kinds of things. People do it by creating
work pieces. They build a chair just like thousands of millions of chairs have been built before. They do electrical installations, They build guitars, they build banjos, and you know, I think that legitimacy for that as a way of honing one's capabilities and experience as a developer. Yeah, build tetris. You're not gonna be able to sell it. Somebody already did that, but you will learn a lot in the course of doing it.
Yeah, if not about programming, then about geometry.
Yes, you'll learn about wall kicks, which are the point where everyone thinks, I know how Tetris works. And they get in and they got the balls falling, the blocks falling down the pit, and then go left and right and they can rotate it. And then they get one up against the wall and they got to rotate it. They're like, oh, now that's not right. That's not how it was on the game boy, and then they go down the rabbit hole.
So it'd be a good way to teach math. Yeah, I know, So what else can we talk about before we sign off? Here a few minutes left, So.
You know, I guess on the use what works thing? I would love people to go and check that out. There's a whole bunch of us. It's not my thing. It was that thing that happens with you know, these sort of projects where I think about fifty people said yes, this sounds fantastic, let's do it, and then you know, forty of them got busy with other things, and there's now a critical mass of about maybe a dozen kind of checking in as and when they can and moving
things forward and sharing their own ideas. But you know, I guess it's it's a little bit like old dot net was back in the day. If this sounds like something you have been grappling with, we would love for you to come and you know, join us under this rallying cry and tell us you know what works for you, what doesn't work for you, Do have anything we can help you with? Do you have anything you can help
us with. Do you have a great idea or a case study or something that you'd like to get out there, And there's any way we can use this network to amplify that message and get more folks tuning into it. And you know, and if you think this is completely the wrong approach and not something we should be doing, we'd love to know why you think that. Yeah, so you know, head on over if you like it, to sign the manifesto because a big list of names always
looks nice, and yeah, we're open. We got a bunch of you know, Gethub issue tracking plugged in there and get Hub discussions is on there. Nice, and we would love to know what people think about the whole the whole wider question of how do we get a how do we get to a point where if you've got a great idea for a useful piece of software, you can put it out there in the world and know that when the time comes, it'll be able to pay the bills and you know, eventually hand over to somebody else.
That's the other big question with a lot of these, you know, small but critical open source projects. Right is if the one person who has been doing almost all the work for whatever reason does not get up tomorrow and want to do more of that than what, you know, what's the continuity for all the folks who are relying on it.
Yeah, good stuff, Dylan. It's been great talking to you as always.
Hey, it's always a pleasure.
Absolutely, and we will talk to you, dear listener next time on dot net for products.
Dot net Rocks is brought to you by Franklin's Net and produced by Pop Studio, a full service audio, video and post production facility located physically in New London, Connecticut, and of course in the cloud online at pwop dot com. Visit our website at d O T N E T R O c k S dot com for RSS feeds, downloads, mobile apps, comments, and access to the full archives going back to show number one, recorded in September two thousand and two. And make sure you check out our sponsors.
They keep us in business. Now, go write some code, see you next time.
You got a jam.
Vans and