Thanks for the reminder! I probably knew that at one point, but with so many Free Software/Culture organizations lighting themselves on fire over the past year or so, they start to all blur together…

I forget whether it’s through the data page or somewhere else, but I gather that they’ve always made it pretty straightforward to grab the questions/answers/comments. For a while, web searches were a mess, because of all the fly-by-night outfits syndicating all conversations with SEO trickery to sell ad space.

It needs more traffic (and that’ll require better servers and more moderation with training to make sure that it doesn’t follow the same arc) to become a standard place to go, but I’ve become fond of Codidact. If I remember correctly, the founder left Stack during one of their bigger moderation blow-ups.

Two reasons, three if you include the footnote that already answers the reductive “substitutable service” idea. Thirty-plus years ago, Free Software developers often used proprietary compilers. That was reliance on proprietary products. This is relying on proprietary tools to generate code, unless we want to pretend that people are out there running A/B tests between hand-written and generated code. The tools also generate unmaintainable code while making the developers less capable, making the tools more necessary as teams use them, which also sounds like reliance.

Personally, I think that using LLMs is a terrible idea, especially for Free Software. It creates a reliance on proprietary software^[1]. Generated code has uncertain provenance, for the same reason that researchers can provoke LLMs to “auto-complete” famous books, undermining the license. LLMs rewrite instead of fixing, since the latter requires intelligence, so it makes less-maintainable code. And studies show that using models slows work down, rather than speeding it up.

I’ve only found two arguments in favor: The technology is (allegedly) here to stay. And sure, the studies say that it slows down work, creates bugs in most of its changes, makes everything harder to maintain, alienates the team, and erodes skill, but don’t worry, because I have a system.

I wouldn’t risk it on any project of mine, and especially not for anything released under a public license where it needs to promise people that they can use the code for any purpose.

It’s…a web-accessible log file, basically. Timestamp-and-status, line after line, and “clients” read it periodically to report updates. My now-defunct feed has a couple of comments (# ...) at the top for the URL and my “user ID,” but I don’t remember if they were requirements, conveniences, or just a cargo cult thing.

I used it for a few years. It was a nice idea, but my disinterest in following everybody’s feed meant that I was mostly just posting blindly, which then became just automated posting of blog post announcements, and then a breakage in the posting tool (which I built up all my infrastructure around) convinced me to give it up, since I wasn’t a particularly “good citizen.”

I tried to solve some of those problems with a desktop tool, but the desktop framework was never ready for prime-time and Node.js was probably the wrong choice in general. I know that there was a hosting service, but that also feels like the wrong way around.

It’s still a pretty good idea, though.

It seems interesting, and we have needed new thinking on licenses for a long while, but a few details mildly concern me.

First, while not universal, brevity improves readability, but often leads to loopholes. Mature licenses grow so long because lawyers have needed to repeatedly game out threats and how to resolve them before arguments break out.

Then, I don’t quite know how to phrase such a thing, but it’d be nice to bar Contributor License Agreements and other copyright assignments. Prior to LLMs laundering Free Software and helping novices flood everybody with low-quality patches, the biggest threat to Free Software was (and probably will be again when the AI bubble pops) was projects exercising the copyright holder’s authority to re-release without a license, taking the community’s work with them behind the paywall.

The LLM condition also seems concerning. I don’t like how the big companies have behaved either, but the problem is corporate exploitation, not the technology. If a teenager wanted to train their own model with a selection of projects, we don’t want to tell them to stop, and if a company wants to repeatedly scrape projects for a PowerPoint deck, that’s just as bad. Plus, training a neural network probably falls well inside the bounds of Fair Use, though publishing the output probably wouldn’t.

Otherwise, you might want to reach out to the folks working on Copyleft Next, which has a similar interest in building on the GPL, since Kuhn and Fontana have been at this for a long time.

I agree that creating is inherently political, because politics pervades creation whether we choose the politics or not, but that’s not a useful argument after somebody says “it doesn’t matter to me.” If you want to get into that shouting match, it’s your time to waste.

My point is that, behind the garbage philosophy, we also now know that it’s garbage technology, so all these people telling us about their utopian meritocracy where we just ignore bigotry are exposed as full of it. Cloudflare, Framework, and so forth, are not only OK with Great Replacement rhetoric, but also incapable of telling solid software from broken, and that’s a stronger indictment than just trying to drag the conversation back to the bigotry.

That’s probably a good choice, since it makes a perfect response to “politics shouldn’t matter, here, just merit.” Turns out…

I primarily use GNOME Authenticator, but after an inopportune crash, I now also run 2FAuth on my home server as a backup, and now just hope that I remember to do the export/import dance going forward.

I’m another conflicted person on this. I ran Tiny for years, so I never hated it. But it had so many updates that assumed that I’d know in advance to update something on the system (PHP libraries, database schema, etc.), and then putting the git repository behind Cloudflare led to a cycle of notifications that I needed an update and then waiting for Brigadoon to reemerge so that I could pull the latest source. And any time that I needed to look for a solution to a problem, reading through the forums made me regret the choice a tiny bit more.

It’s reasonable software, but I ended up moving to Fresh RSS on an in-house server, and that has gone better, but I hope that the Tiny community pulls together something better to keep the space diverse.

In my case (not necessarily your case, of course), the cheapest selling-point has become that I already have a browser open for almost everything else, so that’s one less thing to install and check in on. But it’s also easier to keep up to date reading when individual computers have problems and usually has a nicer API for scripting, if you need that sort of thing.

That’s close to how I think about it, yeah, but I’d push more in terms of the investment. Since Jekyll, Hugo, Svelte, Eleventy, and the rest just generate flat HTML to upload, there’s nothing wrong with using it for a single page. But you end up needing to learn the whole build-and-deploy process and all the layout quirks, which (especially if you’re starting from scratch) will take longer to get the page out. And like you point out, the more material you have, the better that investment looks.

But then, if you already know the system, there’s no new investment, so it becomes more of a toss-up whether to build things that way, since a page of Markdown is slightly faster to write than the equivalent HTML.

Personally, after churning through all the static site generator options, I landed on Jekyll, one of the first of them. It’s definitely not the sexiest solution, but it’s Markdown-in and HTML-out (my main page is still raw HTML/CSS from like twenty years ago, though), was the easiest for me to match the styling that I wanted from the base theme, and it’s been along for long enough that it’s mostly surprise-free.

That said, if you only want the equivalent of a business card, I might argue that setting up anything is probably overkill, all overhead for just a tiny bit of content. In that case, you can grab some modern-ish HTML boilerplate like this one, then use Pandoc to convert the Markdown (which you presumably already know if you’re messing with Hugo) to the HTML that goes between <body> and </body> in the boilerplate. Add CSS, and you’re done.

Oh, and actually, depending on how broadly you want just the “business card” idea, something like Littlelink might also fit your needs, where you hack out the links that you don’t care about and fill in destinations for the rest.

I developed this script for creating permanent/static archives of social media exports, so it’s not a full solution - not a web service, expects file inputs, uses a probably incomplete list of shorteners to avoid pulling real pages - but it along with the shorteners.txt file in the same repository, iterating to find a domain not on the list, might at least inspire a solution, if it’s not good for your specific cases.

I buy it.

As it turns out, a couple of months ago when a laptop crapped out at an inopportune time, I needed to retreat to a much older machine with barely enough memory to keep a browser running all day. As I tried to work out a recovery plan for the things that didn’t seem properly backed up (they were, just not where I expected them), I remembered that I had a couple of old Raspberry Pi units that I never did much with, and decided that could take the load off of the laptop if I tossed them in the corner.

So far, I have Code Server to substitute for Visual Studio Code, Cryptpad for Libre Office, Forgejo just because I really should have done that a long time ago, Fresh RSS for a rotating list of RSS readers since I dropped my Internet-accessible Tiny Tiny RSS installation, Inf Cloud and Radicale for a calendar/address book, Jellyfin that used to run on the then-in-use old laptop, Snappy Mail for Thunderbird and the bunch of heavy webpages from mail providers, YaCy because I’ve wanted to use it more for many years, and a few others.

Moving onto a more functional computer, I decided to keep the servers running, because the setup works about as well as the desktop setups that I’ve run for years, if I use a few pinned tabs. I’m sure that I’ll scream about it when something goes wrong, but it does the job…

Yeah, it’s on the local network, so I’ll need to mess around with aliases again. And they seem to think that it’s possible to set this up on a subfolder, with the APP_SUBDIRECTORY variable, but it doesn’t exactly give the impression of rigorous deployment testing, so you’re right that I should assume that part doesn’t work. Thanks!

I’ve been using different versions of SearX for a long while (sometimes on my server, sometimes through a provider like Disroot) as my standard search engine, since I’ve never had great luck with the big names, and it’s decent, but between upstream provider quota limits, and just the fact that it relies on corporate search APIs at all, sometimes the quality craters.

While I haven’t had the energy to run YaCy on my own, and public instances tend to not have a long life, I don’t have nearly as much experience with it, but when I have gotten to try it out, the search itself looked great, but generally didn’t have as broad or current an index. Long-term, though, it (and its protocol) is probably going to be the way to go, if only because a company can’t randomly tank it like they can with the meta-search systems or their own interfaces.

Looking at Presearch for the first time now, the search results look almost surprisingly good if poorly sorted, but the fact that I now know orders of magnitude more about their finances and their cryptocurrency token than what and how the thing actually searches makes me worry a bit about its future.

I believe that YouTube supports RSS. I haven’t used it in years, but gPodder allowed subscribing to channels.

Ah, yeah. From this post:

• Go to the YouTube channel page.
• Click more for the About box.
• Scroll down to click Share channel. Choose Copy channel ID.
• Get the feed from https://www.youtube.com/feeds/videos.xml?channel_id= plus that channel ID from the previous step.

From there, something (like a podcast client) needs to grab the video.

Otherwise, I’ve been using Tartube to download to my media server, which is not great but fine, except for needing to delete the lock file when it (or the computer) crashes, and the fact that the media server hasn’t the foggiest idea of how to organize the “episodes.”