Junior dev: I’m bummed by how little of the code I write will ever run in production

Senior dev: I’m horrified by how much of the code I’ve written is still running in production

@norootcause *Especially that code designed as an emergency hot fix that we'll fix properly Very Soon Now.

@xinit

I got a call from my then employer, who had a call from a company that I used to work for, late one Friday night.

They'd bought some software from my new employer and said there was a problem with it.

Nope. No problem there. The problem was with software I had written years before and it was the only way they could get someone who might be able to fix it.

It was supposed to be a short-term bridge between two systems, good for three months until they replaced one of them. It was now 3 years.

They'd just upgraded one, changed API. Fixed in 2 hours. Billed for 20.

@norootcause

Chalk mark. $5. Knowing where to put chalk mark. $95. @EricLawton @xinit @norootcause
@xinit @norootcause
After I realized that, I explicitly added my name and the date in a comment next to temporary hotfixes. There's at least one from 2017 still around, which was re-hotfixed in 2020 according an added comment from a colleague 😆

@norootcause I once got a call from an international bank that I’d written some code for, to solve a specific and important issue, about 5 years before.

I happened to have the code on my laptop, so I looked for the problematic code they described via eMail.

Couldn’t find it. I wrote back, confused, that the code I wrote didn’t even contain the line they said was erroring out.

They wrote back explaining that it wasn’t the code I’d written, it was the code REPLACING the code I’d written.

@norootcause happiest months are when the LOC goes down.
@norootcause retired dev: I am horrified when I get asked for advice on code written by me and last changed by me 15 years ago - it doesn't need modification or maintenance (it is still working perfectly) - but then again, when I was a junior dev, I worked on modules which had last changed before I was born.
@norootcause principal dev: thank goodness someone stopped me from shipping more code.
@whack @norootcause This is so true, it hurts
@norootcause code my dad wrote in the early 90s is still in production...
@norootcause Cobol dev: Oh, my sweet, sweet summer children.

@norootcause

Vibe coders: I don't know why everyone is so horrified my code is running in production.

@norootcause My Dad's Fortran code written in the 70s (he retired 25 years ago) was only replaced a few years ago (by my wife and her team!). He was horrified when we told him. "That should have been replaced a GENERATION ago!" - Dad
@buddhawilliams @norootcause That's a "new" definition of inheritance in programming.
@buddhawilliams @norootcause That's the real legacy project here.

@buddhawilliams @norootcause I'm curious about this project, mainly the motivations for the rewrite. I do a lot of legacy Fortran recovery and I'm always on the lookout for recovery and replacement stories.

Of the codes I've recovered, there's only been one I would call bad. The modeling was too coarse and there were clear signs of serious truncation and roundoff error; it wasnt clear what the purpose of the code was at the time it was written (late-60s). The underlying engineering wasn't very good so no amount of modernization was going to fix what was really broken. Most of the other codes were some degree of inconvenient, twisted around platform and language limitations that no longer exist. Generally the codes are still useful if limited and their main fault is they've been neglected for 30 years or more.

@arclight @norootcause it wasn’t my job so don’t know a lot about it. But it was a rostering system that had a huge number of rules for assigning rosters within hospitals I believe. Every time they added a new rule (‘don’t put person x on same shift as person y’) etc the processing was taking exponentially longer so it needed to be looked at. The original code was built by external contractors so little institutional knowledge kept so when they decided it needed cleaning up it made sense, to them at least, that they rewrite it rather than fix the code.
At least that is my understanding from a third party.
@buddhawilliams @norootcause Yeah, that sounds like a bad application of Fortran. But if it's the 70s, F66 may have been the only decent language available. Sounds like rewriting was the right choice for a lot of reasons.
@arclight @buddhawilliams I have always loved how scientific programmers refer to software as “codes”
@norootcause @buddhawilliams You should hear us keep referring to input files as "decks" :D
@arclight @buddhawilliams At least tell me they aren’t on punch cards anymore!

@norootcause "It's an emergency there's a pandemic starting now. I'll quickly ducktape something to show what we can do to analyze the virus, and surely the pro will eventually move in with a better solution“

Five years later and a major national virus surveillance platform still runs on what's basically an overgrown descendant of the scripts I use to backup my smartphone.

@norootcause I recently started a new job. The previous guy had retired after being in post for over 30 years…

I have inherited Perl scripts he inherited.

These scripts are both essential and *going to be replaced*

@norootcause also senior dev: "Who wrote this code that somehow got into production??? Oh no, it was me"

@norootcause
@trevdev

By that I am a junior-midlevel-senior dev:

I hate that none of my code runs in production.
I know that it would be better if my code would run in production rather than the code of others.
And I am terrified by the fact that soon my code will run in production.

All at once.

@norootcause maint programmer... i'm bummed by how much code you wrote is still in production.
@norootcause Every time I return to a tool I worked on previously...
@norootcause there is a non zero chance that code I wrote is on every single nvme firmware ever made by SanDisk, and that terrifies me