movq.de phlog gopher://movq.de/1/phlog movq 2026-06-29T05:07:29+00:00 2026-06-29--on-not-publishing-software 2026-06-29T05:05:38+00:00 tag:movq.de,2026-06-29:phlog/on-not-publishing-software 2026-06-29 -- Regarding: Not publishing software anymore ======================================================== In the previous entry, I wrote that I don't publish new software anymore (at the moment). The reason is simple: Until 2026, I was able to live under the illusion that nobody cares about my code anyway. Or maybe a few like-minded peo- ple, but that's it. No company (or other malevolent actors) would ever grab my code and use it in proprietary systems, without asking, simply disregarding the license. That was just super unlikely. None of what I did was relevant enough for that to happen. Hence I never needed to really worry about licenses. Never needed to de- fend them. And now we have AI crawlers, they use *everything*, and they don't care about anything. (People in tech are so hyped that they even defend this behavior. "It's just like a human being learning how to program!" Sigh.) So, now, all of a sudden, my little naive software world collapsed. Long story short: I don't want my code to be used by corporations. I never wanted that. It doesn't even have to be *proprietary*. Just the fact that they benefit from using *my work* sets me off. Very few corpo- rations help make the world a better place, most of them are malevolent or, at best, useless. And I *especially* don't want AI companies to use my stuff -- they build new gas power plants and are just overall, sim- ply, bad. (I have a list here[1].) Hence the knee-jerk reaction of taking stuff offline. When I'm being honest, GPLv3+ isn't enough, either. It still allows com- mercial use. A hypothetical AI product could spit out my code or stuff based on it and always mention that it's GPL -- that would *still* be shitty (new power plants, drives people into dependencies, ...). I think what I'd like to see is a purely non-commercial software license. That's probably just as naive. I mean, there's The Anti-Capitalist Software License[2], which certainly goes in the right direction ... Should I use that instead of GPLv3+? But none of this even matters unless it's taken to court. And who are you up against? Billionaires. This is completely pointless. And that's why it's so frustrating, really makes me want to take everything offline right now. As soon as something is online, it's going to be abused. It's even worse: Now with AI crawlers, they don't even have to set up and run my programs. The code doesn't have to solve any actual problem. My programs could be extremely obscure and niche -- that would still be valuable. Using arcane programming languages would be even more valu- able. Just the mere fact that it's *code* or *data* is valuable. The same goes for blog posts or this phlog entry. It's a completely differ- ent level of abuse than it was before. Moving everything to Gopher or Gemini wouldn't help, either. If anything, you'd have to make grabbing content *really expensive*. So expensive that it's no longer commerically viable (while at the same time still accessible to human users). That's the only way out that I can see. ____________________ [1]: https://movq.de/noai.html [2]: https://anticapitalist.software/ 2026-06-28--more-gplv3 2026-06-28T09:53:04+00:00 tag:movq.de,2026-06-28:phlog/more-gplv3 2026-06-28 -- Using more GPLv3+ =============================== (I wrote about this in January already.) I'm slowly migrating my projects to GPLv3+. Every time I touch some- thing, I also change the license -- assuming that I'm the only copyright holder. This applies to everything: Normal programs and libraries alike. I am aware that this means that users of such a library will have to be GPLv3 as well. That is intentional. It won't stop AI crawlers. But I'm not yet ready to take my stuff of- fline. Well, at least not all of it. I have stopped publishing new projects (like my TUI framework "movwin" or my hex editor "bine" or sev- eral Vim plugins) and some smaller projects have indeed been moved to private repos. 2026-02-10--the-luminous-dead 2026-02-13T16:27:31+00:00 tag:movq.de,2026-02-10:phlog/the-luminous-dead 2026-02-10 -- "The Luminous Dead" (Caitlin Starling) ==================================================== (Some spoilers ahead.) Oof! This was advertised as SciFi horror, which isn't wrong: The main charac- ter explores a deep cave on an alien planet, but for various reasons -- one of the main ones being that she must not give off any scent and must remain mostly quiet, i.e. her voice must not be audible -- she's in a suit with lots of sensors and gadgets and a *feeding tube*. Without the connection to the main computer, she'd be completely lost. But then there's also the woman on the other end of the comm line, who's supposed to tell her where to go and what to do. Buuuuut is she trust- worthy? ;-) And why is it only *one* person in that control center in- stead of a whole team? What I didn't expect was that it would turn into a romance. I usually don't like that at all, but in this book, it worked really well for me. They're (almost) never in the same room, they only talk over the comm line and a video feed, which makes this thing way more interesting. The weird power imbalance (the suit can be controlled remotely) is eventu- ally overcome to some degree and actual trust is established -- I think. It's certainly not the most healthy relationship, shall we say. But it was great to read. The book comes with a map and that is sorely needed. I mean, it's all caves, so it would be hard to not get lost without a map. Towards the end, there's a lot of paranoia and panic, running low on re- sources. There are quite a few moments of "oh my god, no, please don't". Then again, there aren't that many "scary entities" in the caves (or are there?), so I'd rather classify this as "terror" instead of "horror", if that makes sense. It was one of the best stories I read these past few months, if not the best. I'll certainly have a look at Caitlin Starling's other books. This is a nice interview (only watch this after reading the book): https://www.youtube.com/watch?v=lui6zspzHts 2026-01-11--thinking-about-software-licenses-again 2026-01-11T18:10:25+00:00 tag:movq.de,2026-01-11:phlog/thinking-about-software-licenses-again 2026-01-11 -- Thinking about software licenses again ==================================================== I have used very permissive licenses in the past. I just wanted to make it easy for fellow "hackers" to use my code. If you see an MIT-licensed project, you know that you don't have to worry about anything. I *had* considered "Bad Actors". But back then, a Bad Actor was some company that takes my code and incorporates it into a proprietary prod- uct. Like, someone takes my window manager, modifies it, and uses it in an embedded device. Or someone uses my Gopher server and ... I don't know, what would you do with that in a commercial product? Okay, anyway, suppose that had happened. Then what? I would first have to know about this violation and then I would have to sue that company. This whole scenario was extremely unlikely. But the situation has changed last year. Now we have "AI" crawlers and I *know* that they are scraping my website like crazy. And now it is *evident* that they do not follow any license requirements (assuming I had used the GPL). They break the rules and everybody knows it (and most people don't care at all, because look how shiny it is). So here's what's bugging me: Since I use permissive licenses, I have no right to complain about this. I still won't sue anybody, but even from a "moral" standpoint, it's just my own fault. I allowed them to do this. I enabled "AI" companies. I was living under the assumption that I just host a small website that nobody really cares about. Maybe a few people, maybe sometimes parts of my code are useful to others. A small community. That's not true anymore. Huge companies make money by using my stuff. Granted, just a tiny fraction, but still. So, what can I do about it? - Take the website offline and move everything to Gopher or Gemini. - Switch my projects to GPLv3. I think that just moving everything to Gopher is probably just as naive as my original approach. So maybe switch to GPLv3? For the vast majority of my projects, I'm the sole copyright holder, so I can just do that. My gripe with the GPL is that it's so hard to understand. I literally can't read the original document and understand what it's about. It's a long legal document. No way. The only thing I can do is read third-party interpretations and then trust them. This isn't great. Sigh. This needs some more thinking. 2025-11-19--maybe-more-java 2025-11-21T18:06:19+00:00 tag:movq.de,2025-11-19:phlog/maybe-more-java 2025-11-19 -- Maybe more Java? ============================== For quite a while now, my "toolchest" has looked like this: - C or Assembler for very low-level tasks. - C with GTK for GUI programs. - Python for "normal" programs or larger scripts. - Shell scripts for glueing other tools together. Rust is very, very slowly creeping in as well, I sometimes use it for "systems programming" or other "low-level-ish" tasks. But I'll be hon- est, Rust is so hard to learn, I always shy away from it and don't use it often enough. Regarding GTK: I loved it during the GTK2 days but now I have to admit that I'm not that much of a fan anymore. It has become a pretty heavy toolkit by now. I've pretty much stopped using it in my own code and this now leaves a gap. I've dabbled a bit with PyQt6. And I've noticed that I'm slowly getting tired of Python's dynamic typing. I find it more comfortable to have a compiler that reliably tells me when types are wrong (or when there are typos), because frankly, I don't think I've ever really made use of dy- namic typing *at runtime* (except for parsing JSON files into dictionar- ies, I guess). My brain doesn't work that way. I want static typing. I want to get all kinds of stuff out of the way before the program even runs. That's just not the case with Python, you always have to test each and every code path just to catch typos (or more serious errors). Some people might say that this is good practice anyway, but I'm not con- vinced -- how many projects really do have 100% test coverage? (I didn't find tools like mypy or Python's type hints to be helpful.) Long story short, I used IBM Java 1.0.1 on OS/2 Warp 4 for Advent of Code in 2024 and that was a surprisingly nice experience. That's an an- cient version, of course. I wasn't new to Java, but I stopped using it while Java 6 was still the most recent version. Java is bad in these regards: - Does not produce native binaries. You always need a JVM. - Memory management is annoying. Having to decide up-front how much RAM my program is going to use is a bit silly. - The code can get rather "wide" because many class names are very long. But on the other hand, these properties are great: - Compiled language, static typing. - Memory safety is not an issue. - Produces small "binaries". - The offical documentation is really good and a pleasure to work with. - It can do proper multithreading (unlike Python). I've already used multithreading extensively during my Java 6 days and it was nice to work with. - OpenJDK is Free Software now. - It appears to be very stable with few surprises and shenanigans, old stuff still works, probably because it's used so much (?) in enter- prise software. And JVM ramp-up times are not an issue anymore. This used to be annoy- ing, but that's a thing of the past. A "java Hello" takes 33 ms on my box. I'm currently toying around with making a little file manager in Java that uses Swing for a GUI. Swing is *not that bad*. I'll also try to catch up with new features that landed in Java. And then we'll see. Maybe I'll get fed up quickly. Maybe not.