Was overwhelmed by a large problem.
Broke it up into small problems.
Now overwhelmed by large number of small problems.
@[email protected] "I followed your advice and turned my monolith into this jenga tower. What was the next step?"
Was overwhelmed by a large problem.
Broke it up into small problems.
Now overwhelmed by large number of small problems.
@[email protected] "I followed your advice and turned my monolith into this jenga tower. What was the next step?"
I guess I'm probably more pessimistic about this than both of you. Maintainers aren't always responsive, releases take time to prepare, patches aren't evenly propagated, and even when all fixes are applied companies often wait to update - if at all.
Anyone with a credit card will soon be able to order the exploit vending machine. Projects and orgs that have their shit together will be alright. Many however will not.
@[email protected] @[email protected] @[email protected] Not to be too dark, but hasn't this been true for a while? I'm sure I could find places on the net to buy (or pay people to find) exploits already, and probably roughly as fast. There are a lot of projects/servers or there with less than stellar security, after all. It almost feels like the bigger change is how much it is now public knowledge you can pay for exploits, rather than whether or not you could already.
„I went for therapy
I needed a remedy
She said, "Rou, there is no other solution"
And prescribed me global revolution“
@[email protected] see, you posted that tagged as no context but... gestures out the window
Turns out that my Boox Note Air 3C actually does a surprisingly good job of running #krita. Slightly laggy displaying the brush strokes but records them accurately, and the e-ink display and roughly A5 form factor means I can draw outside in bright sunlight and on the go.
The e-ink display does make judging colour saturation a little difficult though!
@mavnn Sorry, afraid not!
@[email protected] a shame, but I suspected that would be the case. Supporting that kind of ui slickness across the two different stacks would be a huge pile of work. It looks gorgeous, by the way.
@dstan thank you!! ☺️
@[email protected] @[email protected] Do you have any plans for an android version?
Worrying eats up so much energy and I find that part of it very specifically annoying
@[email protected] I would tell the choir to join in on singing the truth of this, but they're all too tired
@HighlandLawyer @passenger @graydon Yup: I suspect it goes back even further than that, too.
@[email protected] @[email protected] @[email protected] @[email protected] Thing is, I can't help suspecting that each time these people are so pleased with themselves for having come up with their very own, very unique, cunning plan.
Y'all are missing out on a major investment opportunity. @[email protected] is already moving to secure her cut of the vegan human rights (and gluten) free alternative brain industry.
I mean, that's got to be worth paying at least for to six times the cost of non-certified brain alternative provision, right? At least based off how much I get charged for gluten free food...
@mavnn i’ll have my agent talk to your agent !
@mavnn will you write my linkedin posts for me? that’s talent
@[email protected] No I strongly suspect that would rapidly cause me permanent damage. But for a small, small, fee of 1 billion dollars I will produce something I like to call a "vegan human rights free brain alternative" that will turn anything you desire into LinkedIn posts with only minor side effects to people who mostly aren't me (such as the collapse of copyright law, psychosis, etc, etc)
im in bed chilling during work hours. do i feel guilty? not at all! i’ve found the perfect excuse. I’m Background Processing!
@[email protected] "I was employing somnambulant value provision from my horizontal economic person desk"
@mavnn I power @openbenches using my ActivityBot.
I have a small Python script that looks for new RSS entries every 20 minutes and then posts them that way.
@[email protected] I finally got around to doing something about this, and wanted to say thanks for the discussion and pointing me to activitybot; it lead to a much simpler solution than I'd originally thought of, even if it ended up going in a different direction to activitybot (I'm federating rss by running an ap client rather than an ap server)
Fediverss
Read this post in its full formatted glory at https://blog.mavnn.eu/2026/04/10/fediverss_release.html
You might have noticed that this blog has suddenly acquired a comments section at the end of the articles. I've tried some work arounds in the past - in a previous version of the blog using an external service, or manually linking to fediverse posts made about the blog post - it never sat quite right.
For this version of the blog, I didn't want to force people to create a log in for "my blog" or run a bunch of dedicated comment receiving infrastructure, but I also didn't the comments to be controlled by a third party service with all the archival and privacy concerns that implied.
So I've now built fediverss. This is a program designed to sit on the same server as a static blog automate the process of keep that blog on "the fediverse" via its rss feed. It is very much pre-release quality at the moment, but I'd very much like to polish it up so that it can be deployed easily by non-coders and in environments where running it on the same server as the blog isn't practical. That shouldn't be impossible, especially given that all the dependencies are already managed in Nix, so much this space for future developments.
And, you know, leave a comment via the fediverse if this seems interesting!
@blog In case you're browsing over to the blog because you're curious what the comments look like... they look like this.
bonfire.mavnn.eu/pub/objects...
So this is very much a work in progress (see me having to quote the post to add the hashtags...), but meet fediverss - a simple F# program that puts your static blog rss in the fediverse. #activitypub #fsharp #blog
Fediverss
Read this post in its full formatted glory at https://blog.mavnn.eu/2026/04/10/fediverss_release.html
You might have noticed that this blog has suddenly acquired a comments section at the end of the articles. I've tried some work arounds in the past - in a previous version of the blog using an external service, or manually linking to fediverse posts made about the blog post - it never sat quite right.
For this version of the blog, I didn't want to force people to create a log in for "my blog" or run a bunch of dedicated comment receiving infrastructure, but I also didn't the comments to be controlled by a third party service with all the archival and privacy concerns that implied.
So I've now built fediverss. This is a program designed to sit on the same server as a static blog automate the process of keep that blog on "the fediverse" via its rss feed. It is very much pre-release quality at the moment, but I'd very much like to polish it up so that it can be deployed easily by non-coders and in environments where running it on the same server as the blog isn't practical. That shouldn't be impossible, especially given that all the dependencies are already managed in Nix, so much this space for future developments.
And, you know, leave a comment via the fediverse if this seems interesting!
@mavnn One long-ago employer released a brochure that had a mis-spelling right on the cover. Someone at COMDEX visiting our booth spotted it. To be fair, the font was a modern, stretched-out one that made it hard to spot that "multimedia" was spelled "MULITMEDIA."
But, yikes!
Happy for that security guard.
@[email protected] ah yes, a close relative of the famous glance behind you at the 4 metre high projection of your slides at a conference while talking and spot the spelling mistake moment.
@[email protected] I... both do, and very much do not, want to know where they get the rest of their discourse.
@[email protected] I worked in desktop publishing for a while in a large bank in the UK, producing things like product leftlets and terms and conditions sheets (exciting, I know).
The bank and a printing company jointly paid one contract night security guard a significant amount of money as a thank you after he happened to spot at the beginning of his shift that one of the pages of a 70+ page full colour brouchure was upside down, an error missed by the people at both the bank and the print house. All because he got bored enough to flick through the test print left lying around before the overnight run started.
@[email protected] The bank had around 2,000 branches and we were due to ship multiple copies to all of them, to give a sense of scale. Full colour printing in 1999 wasn't cheap, either.
I was a hardware & software QA person for about 20 years, making sure problems didn't get out the door.
I feel professional pride in reporting this problem so that camera company can fix things.
Boy, I imagine some people are in deep shit over this fuckup. A really bad oversight.
@[email protected] I worked in desktop publishing for a while in a large bank in the UK, producing things like product leftlets and terms and conditions sheets (exciting, I know).
The bank and a printing company jointly paid one contract night security guard a significant amount of money as a thank you after he happened to spot at the beginning of his shift that one of the pages of a 70+ page full colour brouchure was upside down, an error missed by the people at both the bank and the print house. All because he got bored enough to flick through the test print left lying around before the overnight run started.
@[email protected] @[email protected] I'm not a fan in general for a lot of other reasons, but the fact that TypeScript types only exist at compile time (and you can tell the compiler to ignore them) does make things like strongly typed IDs easier to deal with. Generated code can just throw the types away, while user facing code gets type safety (for the value of safety TS allows at the best of times) on ID types without run time impact. Haskell allows alias types that have similar properties.
It's not really the kind of thing the dotnet type system is set up to deal with gracefully though.
@[email protected] @[email protected] Case in point: I use strong IDs in my TypeScript and Haskell projects, but I don't in my F# project that uses Marten. And it isn't because of Marten's support for them or not; it's the amount of hassle it causes in other places and the annoying nagging feeling of taking a run time overhead for code that is less pleasant to work with.
@simoncropp Only if there was a VB-esque concept in the language itself as the types being automatically convertible.
Honestly, I doubt it.
@[email protected] @[email protected] I'm not a fan in general for a lot of other reasons, but the fact that TypeScript types only exist at compile time (and you can tell the compiler to ignore them) does make things like strongly typed IDs easier to deal with. Generated code can just throw the types away, while user facing code gets type safety (for the value of safety TS allows at the best of times) on ID types without run time impact. Haskell allows alias types that have similar properties.
It's not really the kind of thing the dotnet type system is set up to deal with gracefully though.
News and community around mavnn.eu projects.