It really comes down to what you’re used to. If you use Windows tools then you already know many of the workarounds for Windows and you don’t know the tools that haven’t been ported there.

For example, you know not to use Python directly, but that you have to install anaconda instead, or whatever the current problems with Python development on Windows are.

The big obvious thing that you can’t get away from is that you have to do things differently if you have develop for two different OSs with a view to deploying on Linux.

In particular support for shell scripts is crap on Windows. I could learn powershell or there’s workarounds using WSL and a bunch of other stuff that I don’t need to care about, but I’d rather not bother.

I mean coding is difficult enough as it is, I wouldn’t choose to use an OS that makes it even harder.

I use Linux because it makes my life easier. It has better support for development. Some of the other stuff is maybe not as easy or polished, but the support for dev tools and the ease of deploying to from local machines to servers that are also running Linux makes up for it.

If I wanted more effort I’d still be using Windows. It would force me to work on cross platform development and deployment. The idea that there’s value in making things unnecessarily hard is just weird. I want Linux to be as simple as possible to use, so I can spend that effort on things that actually matter.

Lots of people still split latex documents into one section per file, because subversion used file locks and we only knew how one person could edit a file at a time.

deleted by creator

Honestly, if you’re sharing office files you’re probably using office 365. This means everything is a web app first and therefore Linux compatible.

I tried using the desktop version of word on a Mac last week, and the latency was so bad on a shared document that I had to switch to the web app anyway.

Basically, if you just want to use Linux you’ll be fine. If instead you don’t want to use Microsoft, you’ll probably have lots of problems.

Microsoft have been brutally effective in getting their tentacles into academic institutes, and you’ll find that everything from email to logging into internal sites relies on an office 365 account.

Pi is predictable and deterministic.

Computer programs exist that can tell you what the next digit is. That means it’s deterministic, and running the program will give you a prediction for each digit (within the memory constraints of your computer).

The fact that it’s deterministic is exactly why pi is interesting. If it was random it would typically be much easier to prove properties about it’s digits.

No. 1011001110001111… (One 1, one 0, two 1s, two zeros…) Doesn’t contain repeating patterns. It also doesn’t contain any patterns with ‘2’ in it.

But pi is believed to be normal. https://en.m.wikipedia.org/wiki/Normal_number

So it should contain all finite patterns an infinite number of times.

Yeah, but a combination of this approach, and adding all compiled file types including .pyc to .gitignore would fix it.

It’s very true on a Mac. Almost every time you click the green button, it jumps to full screen and then you can’t drag another window on top of it.

It’s a pain in the arse because my workflow is to have a reading screen with documents and emails on, and a work screen with whatever I’m actually doing. But if outlook is full screen, you can’t drag any other windows on top of it.

Don’t know why the first guy was saying this is a Windows thing though. I only run onto it on macs.

Dev jobs and data scientists often get a lot of leeway.

Very big tech companies tend to be more open to it. When I was at AWS their threat model was basically to treat every end user device as untrusted, which then meant that they didn’t rely on keeping laptops locked down for security.

Yeah I thought you’d ask this. Basically they’ll never do this, just because their attitude is “fuck you I’m a bank”.

Beyond this, there’s a big difference between source code and having a working system.

For very long running systems their state depends very heavily on how they were maintained, little bits of informal design decisions that get components working together, and the order stuff was loaded in, and what other services were up and running when you booted up.

None of this magic is captured by source code, and it can make even setting up a replacement server, even as part of the same infrastructure really hard.

Of course banks are moving to more modern dev methods that encourages turnkey deployment, but the fact that they still rely on a bunch of COBOL code tells you there’s a lot of very old system running in “do not touch” mode

It wouldn’t matter much.

Most of what a bank does isn’t on your phone, but server side.

In fact most bank apps could be replaced with an internal web browser that is pointing at their website, and a password manager, with no loss in functionality or change in security.

And if you’d like to review the client side code the bank is using you can just open dev mode in your browser, right now.

This is one of those things where you have to read your employment contract.

If your main job involves generating intellectual property (code, or research or something else) and you live in the US there’s probably a clause saying you can’t do this without permission. Otherwise, it’s probably fine.

However, if you’re in the US, they can still fire you whenever they feel like, so if you think this might seriously piss your company off and you need your job, you might want to ask permission first.

Garmin does that too if you want to buy something from a company that still exists.

Because any detector has to be based on machine learning you can open source all code providing you keep model weights and training data private.

But there’s a fundamental question here, that comes from Lemmy being federated. How can you give csam detecting code/binaries to every instance owner without trolls getting access to it?

Some instances will be run by trolls, and blackbox access is enough to create adversarial examples that will bypass the model, you don’t need source code.

That sounds like a bug.

Software can theoretically recover location by searching.

Like my TV hasn’t moved in the last 3 years, if any app searching for Bluetooth sees it they can know where I am.

You don’t need to what?

Location services have to be turned on in Android whenever you’re searching for Bluetooth devices.

This isn’t because they’re necessarily using your location but because scanning Bluetooth/wifi devices could be used to localise you.

As far as I can tell, Garmin does not use location on the phone. Weather updates come through the phone’s internet connection but only work based on the watch location.