Linus from LTT asks Linus if he’d ever heard of software developers being terminated based on how many lines of code they’d written .
Linus Torvalds responds “Anyone who thinks that’s a valid metric is too stupid to work at a tech company…”
It’s clear Torvalds doesn’t know who this is about when questioned.
Linus hints to him it’s about Musk.
“Apparently I was spot on [about Elon Musk being such and individual who is too stupid to work at a tech company].”
I know a CSO of a Biotech who is too stupid to be working in any biomedical research. Guy has a 30 year track record of failures, never worked on a successful drug and he just keeps getting promoted… he’s literally stupid.
LOC is a terrible metric. The worst programmer I ever had work for me had the highest LOC of anyone on the team, and his code was crap that barely worked.
A few posts up was a meme about Arch linux updates having a net negative file size :)
I’ve never had to code professionally, but even on my personal projects, I don’t want a single extra line in the program that doesn’t need to be there and I should be able to understand the purpose of every line years later.
My eyes glaze over whenever I look at corporate code because there are so many moving parts at that scale all from different qualities of programming.
I don’t know if this is a practical thought, but I really wish we could get away from every project being monstrously sized. I prefer small packaged ideas similar to terminal commands. Just because it has a GUI doesn’t mean you need to design every piece of software as if I’m going to spend a day in it. Just give me small, purpose-built tools I can understand and then stop eternally developing and adding features.
To add to this, it seems that every company now either makes one piece of software or 36 different softwares. If they make one piece of software, they endlessly pack it with features people don’t want and if they’re the latter, every piece of software is a hastily-cobbled-together half idea and they just move onto another piece of software. Is there really not a middle ground here?
That’s the result of product managers, project leads etc constantly thinking users need stuff, maybe trying to beat their competition, etc. I have watched a few products get bloated with the aim of beating their competition not providing user value.
The really disgusting part is that actually works (if you’re primarily selling to other corporations). Most of the most popular pieces of corporate software have the common trait that they do tons of stuff really poorly and nothing well. They get picked by the bean counters because the bean counters don’t care that it’s a fucking trash fire of a UI, they’re just looking at the list of other software they can remove because this new software does the same job significantly worse. That or they’re just mesmerized by the giant fucking bullet point list of “features”.
softwares
This is still not a word, my dude.
Hopefully this isn’t falling on deaf ears, but language is intimately more beautiful when you bend the rules: https://youtu.be/J7E-aoXLZGY
I kinda agree with your first point, but AI assistance is so incredibly powerful that it’s foolish not to use it, unless you’re working on some really important logic. And even then, having an AI skim for common mistakes, inaccuracies or inefficiencies is still very valuable.
And what you’re describing is really “Unix philosophy” and I strongly agree with that. Make a piece of software that does its one thing really well, and have it communicate with a simple API (POSIX).
In Unix/Linux you generally just “pipe” one program’s output into another program’s input, and can chain them virtually infinitely.
When I was stuck with that, my rebellion was to widely announce all my merges with negative line of code. Let them try to challenge that publicly.
Of course my current gig is new features generating positive lines of code but the new stupid metric is how much did the ai add. So far I’m losing that battle. Making me more efficient? No, so far ai is doubling the amount of time I’m stuck code reviewing junior developers
My favorite part of Junior devs is that if you tell them in a code review comment to never do that again, they usually won’t.
My least favorite part of AI is that it is convincing the jr devs to ignore me, leading to a lot of pain for them when they get sent to the doghouse for writing production destroying garbage.
And my least favorite side effect of AI is that thanks to all the garbage ai-driven devs churning like a boiling sea, companies aren’t building bases of competent jr devs that will eventually be senior devs anymore, because the good ones are getting lost in the noise.
AI generated commenting? Idk I’ve never coded anything beyond modifying powershell scripts
It could probably do a decent job generating those scripts, given adequate prompting and a few cycles of feedback from you. But it’s almost never a final result. It’s still on you to know what it’s doing and whether it meets requirements, whether it’s sufficiently performant and scalable, whether it’s resilient and flexible. Most importantly it’s up to you to ensure good quality that future you can read and maintain.
I love folklore.org. A long time ago when I was a lowly junior engineer, I read that story about Burrelll Smith and the mustache. So I grew a beard and got a promotion and raise the next week. Had a beard ever since.
Want to make tons of lines of code from pretty normal code?
Just unroll your fixed size for loops (i.e. convert them into multiple copies of their contents, one after the other as many times as that loop would loop).
You can actually automate it and in fact some compilers will do that when generating assembly for some microprocessor architectures (if the loops aren’t crazy big) because it increases performance in those (because the JMP instruction at the end of the loop is quite expensive).
Lines of code should be on the balance sheet as liabilities, not assets
So, code golf?
Almost as good as a sabotage. Lines of code metric is no different to someone talking a lot and doing little work, very likely counter productive and a big problem. Leave the good quality stuff to open source while big tech take it and do their best to “embrace, extend and extinguish”.
deleted by creator
