100÷. I used to work for a bank and the lending team didn’t even know how to calculate loan repayments. They just deferred to what the core banking system did.

The core banking system was written in a proprietary language in the 70’s and machine translated into another (slightly newer) proprietary language in the 90’s. At the time I wouldnt be surprised if management was patting themselves on the back for a modernisation job well done. Just get the computer to do the conversion, right? The sales guys of the new platform assured us they could migrate everything automatically and we always trust a sales guy!

Of course the machine translation is like reading machine code so very difficult to understand / follow / change. The developers working on it were in maintenance mode and everyone was afraid to touch it incase some calculation broke.

The point is that it’s exactly what you described - the users were trained to push buttons and trust the system output without actually knowing what they were doing and if it was correct.

Pretty sure the bank recently got fined for compliance breaches as well. It’s not because anyone there was bad, they just had no idea how anything was meant to work

France is a safe haven for foss media codecs because its law does not consider software patentable

TIL there is a country that sees reason about software patents

If I was a bot author intent on causing misery I’d just use the user agent from the latest version of Firefox/Chrome/Edge that legitimate users would use.

It’s just a string controlled by the client at the end of the day and I’m surprised the GPT and OpenAI bots announce themselves in it. Associating meaning on the server side is always going to be problematic if the client can control the value

even Valve told Ubuntu users to use the Flatpak for Steam instead of the Snap

Hahaha really? That’s awesome. I wonder if Canonical will ever take the hint that nobody wants Snap when better, more open alternatives exist

Yeah, package manager is a big one. Many of us got burned by rpm’s early on and just avoided all rpm-based distros since then.

Of course as you say that hasn’t been a problem for over 10 years but the scars haven’t gone away.

I’d only recommend Ubuntu to someone if I knew they knew some else using Ubuntu (so I could tell them to hassle that person instead of me when they have problems).

Otherwise, I’d absolutely recommend Fedora, because it’s actually up to date unlike Debian. I use it myself because it tends to have the best of what the open source community has to offer while not needing constant tweaking

And then managers go “why does shadow IT exist?”

…you have my condolences

Yep, being familiar with the data model is 98% of the effort.

The remaining 2% is the query

It’s more like android apps from early versions of Android before the permissions became user-managable.

It won’t prompt you to give the application access to certain permissions, all the permissions are predefined in the manifest by whoever published the application to flathub. When you run the application you just hope it won’t cause too much havoc (you can of course verify the permissions before running it, but I guarantee most people won’t)

Flatpak supports sandboxing but due to how most desktop applications want access to your home folder, network etc many apps simply disable it.

Regardless of the level of sandboxing applied to the app, Flatpak is a great way for a developer to package once run anywhere. Prior to Flatpak, if you wanted to support multiple distros, you had to build a package for each distro or hope somebody working on that distro would do it for you.

Inb4 AppImage was here first. And if you mention Snap then GTFO

I disagree unless the tests are reasonably high level.

Half the time the thing you’re testing is so poorly defined that the only way to tighten that definition is to iterate.

In this sense, you’re wasting time writing tests until you’ve iterated enough to have something worth testing.

At that point, a couple of regression tests offer the biggest bang for buck so you can sanity check things are still working when you move on to another function and forget all about this one

Wow, and here I was trying to set breakpoints using the devtools debugger and faffing around with sourcemaps.

Wish I knew about this 10 years ago!

Tbh im incredulous that explicit sync wasnt a thing from day 1.

Like what kind of sane API have you ever used that didn’t allow you to buffer / queue up operations and then flush them all at once?

I’d just like to interject for a moment. What you’re refering to as Linux, is in fact, GNU/LInux, or as I’ve recently taken to calling it, GNU plus Linux. Linux is not an operating system unto itself, but rather another free component of a fully functioning GNU system made useful by the GNU corelibs, shell utilities and vital system components comprising a full OS as defined by POSIX.

If I had a dollar for the number of BS CVE’s submitted by security hopefuls trying to pad their resumes…

Only the ones that work for the BBC

Err, no? At what point did I claim to be an expert?

It doesn’t take a genius to realise that serving 100-record chunks of a billion record dataset using limit 100 offset 582760200 is never gonna perform well

Or that converting indexed time columns to strings and doing string comparisons on them makes every query perform an entire table scan, which is obvious if you actually take the time to look at the query plan (spoiler: they don’t)

“Why can’t the system handle more than 2 queries per second? This database sucks”

Devs who don’t understand how SQL or relational databases work write absolute abortions of queries.

9 times out of 10 - yes it is absolutely the devs. I say that as the dev who gets tasked with analysing why these shitty queries from our low budget outsourced labour are so slow