Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
![](/static/devrant/img/pipeless-devrant-banner-white.png)
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
Search - "complete rewrite"
-
I'm developing my "game engine" for over 2 years. 9 complete rewrite and 3 language change (c#->java->c++) but I love doing it. It's an amazing experience :D15
-
Let's make something in angular sounds like a great idea.
6 months later
Let's port everything from angular to react as angular 2 is a complete rewrite.
Another 6 months later.
I think i like the simplicity of vuejs.
Lets try that now. 😂😂😂1 -
The last person who might have taken offense at this recently quit, so time for a consequence-free rant. I just want to say...
Fuck absolutely every single one of my teammates who quit this year. Fuck your shitty, undocumented spaghetti code from hell that the rest of us will have to rewrite because it's utterly broken and functions mostly on prayer and luck. Fuck the 1000+ git repos we'll have to rename so we can even begin to tell them apart. Fuck your complete lack of any sort of processes or procedures or standards. Fuck the person who hated tickets and decided we could just have hundreds of people ask us for help on Slack whenever they need it. Fuck the people who quit because we got a new manager who told us we need to support the applications we build. Fuck the person who said "I'm leaving because I want to move forwards instead of backwards" as if fixing bugs in the code YOU WROTE TWO WEEKS AGO is really moving backwards. Fuck the two people who designed their own separate pipelines and then used both without bothering to debate and pick the better one (spoiler: both are completely undocumented and broken as hell).
I hope your various new employers figure out that your strategy of covering shit with gold paint doesn't change the smell.
Now the rest of us have to fix it all, and we're probably going to start by demolishing most of it so we can rebuild it from scratch.12 -
Hey Root, remember that super high-priority ticket that we ignored for five months before demanding you rewrite it a specific way in one day?
Yeah, the new approach we made you use broke the expected usecases, and now the page is completely useless to the support team and they're freaking out. Drop everything you're doing and go fix it! Code-complete for this release is tonight! -- This right after "impacting our business flow" while being collapsed on the fucking floor.
Jesus FUCKING christ, what the fuck is wrong with these people?
If I dropped the ball on a high-priority ticket for two weeks, I'd get fired, let alone for five fucking months.
If I was a manager and demanded a one-day rewrite I can only imagine the amount of chewing out I'd receive, especially on something high-priority.
And let's not forget product ownership: imagine if I screwed up feature planning for someone so badly I made them break a support tool in production. I'd never hear the end of it.
Fucking double standards.
And while I'm at it. Some of the code I've seen in this codebase is awful. Uncommented spaghetti, or an unreadable mess with single-letter variables, super-tightly coupled modules so updates are nearly impossible, typos in freaking constants added across sixty+ files, obviously-incorrect comments, ... . I'll have to start posting snippets to show them off. But could I get away with any of it? ha. Hell no. My code must be absolutely perfect. I hear about any and every flaw, doesn't matter how minor, and nothing can go out until everything is just so.
Hell, I even hear about flaws in other peoples' code during my code reviews. Why? Because I should have fixed it, that's why. But if I do, I get yelled at for "muddying the waters."
Just. JESUS FUCKING CHRIST.
It's like playing a shell game where no matter which shell I pick (or point to their goddamn sleeve where they're clearly hiding it), I get insulted for being so consistently useless, and god damn, how can I never find the fucking pea or follow the damned rules? I'm so terrible and this is why "nobody trusts me." Fuck you.
I'll tell you why I can't find your damned pea: IT'S RATTLING INSIDE YOUR FUCKING HEADS, you ASSHOLE FUCKING IMBECILES.
That's right: one pea among the lot of them.
goddamn I am fucking pissed off.rant drop everything and rewrite your rewrite oopsie someone else made a mistakey double standards shell game root can do no right root swears oh my8 -
As a consultant, you get tasked with a variety of stuff. Last few weeks been struggling to maintain an old C++ application that was written by a complete tool of an a$$hole with zero knowledge on how to write maintainable and production quality code. It would hardly run without a crash. First it was a challenge I had to accept, but as I stabilized the code and just fell over even more traps, I had to admit defeat and review my approach.
Rewrite is something I would choose last, but this one ticked all the marks worthy of a rewrite. So, the customer is a very friendly researcher and gladly spent 15 hours with me explaining all the math and concepts - just a delight for a programmer to have such a customer. Two days in, with a DDD approach - a functional, more precise, faster and stable application.
Sometimes there is no rant to share, it's rare to have that perfect communication with a customer that is so dedicated that he spends so much time teaching you his speciality and actually understand your approach. DDD was really a lifesaver here, by using it's key concepts and ubiquitous language. The program is essentially 8000 lines of math, but wrapping it up with value objects and strong domain models made me understand his domain and him mine. It also allowed me to parallelize the computations, giving me a huge performance boost. Textbook approach, there will not be many like this!4 -
Why is it that pretty much zero package & framework maintainers understand semantic versioning?
1. If you do a complete rewrite of your package, but the resulting API is identical, you don't need to bump to the next major version. As a user, I'm thankful for your increased performance or cleaner internal code, but it doesn't really affect my update process.
2. If your package required some-framework 6.0.0, and now ALSO supports some-framework 7.0.0 but is still compatible with 6.0.0, you don't need to bump to the next major version. As a user, I can now upgrade the framework, and know that the package will keep working, but otherwise it doesn't really affect me.
3. Following your versioning along with the framework/language version is super annoying, especially if your library really doesn't need to differentiate between framework versions because it's not actually utilizing new framework functionality.
4. On the other hand, if you stop supporting a certain language, framework or shared library version, or change the public methods, exceptions, fields, etc, you MUST bump to a new major version.
Yet everyone gets this wrong.
For example, many of Laravel's underlying subpackages (for collections, filesystem, database, config, http, mail, etc) do not change their code in a breaking way, or do not even change at all between major framework versions.
Yet they follow along with the major framework version.
Now if someone makes a library "laravel-elasticsearch" which uses the support libraries and collections from laravel, they need to update their package to move along with the versions as well, and often they choose to number their library along with the framework in turn.
This means that to update the framework, you also need to update over 9000 dependencies.
FOR NO FUCKING REASON. THE ONLY CHANGE IN THOSE FUCKING DEPENDENCIES IS TO UPDATE COMPOSER.JSON TO BE COMPATIBLE WITH THE FUCKING FRAMEWORK.
Meanwhile, Laravel itself breaks repeatedly on minor/patch version updates, because breaking changes slip through their review process.
Ugh.3 -
Me and my manager throughout 2020
January:
Me: So umm, we can release the new app version
Manager: No we promised client X app first go build that
Me: umm, ok.
February:
Me: so the app is done, but client hasn't setup area L so there is no data there
Manager: ok, I'll have them setup area L soon ™️
March:
Manager: area L is too much work to setup, use workaround L thats way better
Me: ok ...
April:
Manager: client is nitpicking on design and layout please make this mess even greater
Me: ok, anything else?
Manager: yeah also start on app for client Z!
Me: and our app update?
Manager: later son! Risk tooo muchos!
May:
Me: the mess for client X is done, and first version for client Z is also ready for test
Manager: ok good work, here is a new set of things to mess up
Me: but... Seriously, wtf?!
Manager: clients want quality
Me: ah ok, not nitpicking, cool
June:
Manager: client X went MIA, but client Z will send you a weekly list of things they don't understand and want to change
Me: ah great, truly worth postponing my February holiday to release nothing
July:
Manager: so, how we doing on all them changes
Me: well, I am a loyal custodian with alot of pleasure in my work!
Manager: ah ok good!
Me: any news from client X??
Manager: who
Me: mkay ... n.v.m
August:
Me: can we release yet?
Manager: change, we can!!!
Me: are you Obama?
Manager: ambitions
Me: fuck you pay me
September:
Me: I am confident we can now release all 3 apps as promised mid september
Manager: great!! Good work
Also manager: you know that immensely complex area within the app? That needs a complete rewrite because we have bad ux there!!!
Me: ok... To which requirements?
Manager: good ux, we must have standards
Me: but the layout of page R id generic as page F so then we need to align there as well
Manager: go! Do!
Me: ok I'll come up with my own requirements then
Manager: we also need documentation
Me: really!!!! How clever of you to fire colleagues T & P and we now have zero workforce for that
Manager: things will get better someday
Me: ah, great! Put it on my calendar
October:
Me: I need a sabbatical biatch
Manager: a what?4 -
I feel a bit ashamed posting this, compared to some of the amazing things you guys have built.
Coolest thing I have built was my first app:
https://play.google.com/store/apps/...
Story:
It was back around new years 2014-2015. I bought a charango and started playing some gigs. I carried around a book with chords. I thought it was a bit annoying to have to take it with me. Looked for an app and there wasn't any (today there are 2-3 other). So I decided to make an app.
Bare in mind that I had just a bit of experience with C from university. No OOP. So I went on youtube and started watching some tutorials while I developed it. Learned by trying. Trial and error.
After around 2-3 months of working on it every day after class until going to sleep, it was ready.
I decided to put it on play store for other people to use. Turns out there was a need. I got 10,000 downloads in less than half a year (it is quite a niche, so unexpected). Since then it has stayed around 6000 installs on active devices.
It is my biggest personal project success.
Since then, I have continued making apps in my free time, getting better and more professional. But none has come even close to that ones popularity. My plan is that to mark the 5th anniversary, I am working on a v2.0 (complete rewrite) with new features and instruments.
Sorry about tl;dr5 -
"There's a problem with the app, can't you guys do something about it? It's wasting my time, aren't you gonna fix it already?"
Has been secretly working on a complete ground up rewrite which solves all the problems and reveals it to surprise and relief.2 -
Project in college, many moons ago.
Team is building a robot for a project. Nothing too crazy, it does some simple tasks like walk along a path and shit.
3 weeks for the project. 3 team members.
The largest graded part of the project is the ability to follow a path based on vision.
The 3rd member INSISTS on doing that part, he says “I want to prove to the professor that I am the smartest in the class so he helps me get a work term.”
Of course, my other partner and I see this as the complete selfishness of a child who will never be employed anywhere worth talking about anyways. He is a big asshole about it and we end up giving in.
## Week 1
We get our parts done (working together the way a team would) without his help.
He struggles, hits walls, complains. You know, dumbass grown child stuff...
## Week 2
We offer to help since we are done. He refuses. The teacher sees all of this and doesn’t like it at all.
After class the 2 of us go to the teacher and let him in on the details. The guy insisted, he is struggling and will not take help etc.
Teacher goes and talks to him and tells him it is a team project for a reason and that we should be helping. He says yes.
Then he misses the rest of the classes that week and send an email saying...
“Since everyone decided to keep interrupting me and breaking my train of thought, I could not get anything done in class. Therefore I will be staying home to finish the project from there.”
And to top it off, he didn’t even take home the robot’s connectors he needed to do the damn thing. Haha.
## Week 3
We know he wasn’t going to get it done, so we approached the teacher. We make it clear that we have done all we can and that we are not ok with losing marks because of this.
Since we are both good students that he likes, he decides to give us an option.
You can take a 50% on his part even if he doesn’t get it done (for trying to help) or we can do it ourselves and he won’t get the marks if he doesn’t finish.
## Night before
We say fuck it and do the thing.
In fact, since we were learning Java at the time we decided to do it in Java. Our other prof sees us playing with robots and gets excited, he stays with us and suggest improvements.
In the end we rewrite all 3 robot functionalities in Java and hand in the project the next day.
## The day of
Partner 3 comes into class and says this...
“That walking path part is impossible, I didn’t get it done, but I bet nobody else did either. So at least we will get a 60% on the other 2 parts!” (With a big shit eating grin)
Prof calls our group up. We walk up and the prof looks at the 3rd guy and says.
“Since you have decided to do your part alone, we will have you present your part alone at the end of the groups”
He tries to say something but the prof cuts him off and tells him to sit down.
We show all of our code and the robot does everything perfectly.
Groups go by, now it’s that guys turn.
He says that the walking part was impossible but seems to realize right away that he just saw EVERY other group get it working.
The teacher ask him to stay after class.
## Result
We got a 98 (prof said he was hoping we would have done in VB like asked but he liked the result a lot).
Other guy gets a 5% for his non-working spaghetti code on 0s on the other 2 sections. He blames us, of course.
Bonus Content:
That same asshat above once said this to me...
“I don’t indent my code so that if I work for a company and no one else can understand the code then I am unfireable!”
Yes, he wrote all code like this...
const Example = () => {
Stuff
More stuff
For() {
Stuff
If() {
Stuff
}
}
}
Fuck that guy🖕🏽3 -
Jeff Dean Facts (Source: God)
Jeff Dean once failed a Turing test when he correctly identified the 203rd Fibonacci number in less than a second
Jeff Dean compiles and runs his code before submitting, but only to check for compiler and CPU bugs
Unsatisfied with constant time, Jeff Dean created the world's first O(1/n) algorithm
When Jeff Dean designs software, he first codes the binary and then writes the source as documentation
Compilers don't warn Jeff Dean. Jeff Dean warns compilers
Jeff Dean wrote an O(n^2) algorithm once. It was for the Traveling Salesman Problem
Jeff Dean's watch displays seconds since January 1st, 1970.
gcc -O4 sends your code to Jeff Dean for a complete rewrite -
Sales team came in earlier this week promising a client an application with several major modifications in less then 4 months. Guesstimating about 6 months to complete all the modifications.
Manager's answer: "Let's rewrite the application". Top it off he wants to use a obscure language that he only knows.10 -
In only I were 1.15 times faster or had better planning (why didn’t I use the Saturday Sunday at the end of the first week 🤦🏼♂️), things would’ve happened differently. I think I’m becoming stupid and my tolerance levels are going down too.
So this happened a while back ..
I was given a code base which didn’t have any changes in the last two years and I was asked to add a feature to this. This was my first task in this new group I was part of. I had two weeks to do this starting on a Monday.
Partway through implementation I realised that the code base is a pile of shit and I wasn’t doing myself or anyone else any favours by shitting on it.
It’s Wednesday. I’ve dealt with many other codebases before but the urge to rewrite this particular one was just unlike anything else. And so I started changing code and before I realised, I modified almost all the important files.
I got sick of this mixed up code and started a rewrite from scratch. It was Friday and I finally had just the basic mechanics of the whole thing working. Now I needed to add all the functionalities and also my new feature.
It should be noted that at no point did I tell any of the superiors I was doing this fearing what they might say and also fearing going back to adding shit to shit.
By the end of the second week, the rewrite was complete and I only had the new feature to add. The rewrite was significantly smaller, compartmentalised and well commented because I did the bloody commenting (where it was not obvious from the code). So on Friday, I was asked about the progress and I told them that it needed some more work and that I need a couple more days. And I got shit for it. I was told it was a mistake giving this task to me and that I am not competent enough. One of the superiors told the other superior about perhaps giving me something more suited to my level. To be fair to them, they were expecting the work in the two weeks to be for the new feature.
And in two days’ time, on Monday (I worked on Saturday and half of Sunday), I finished the whole thing and gave it to them. New feature was working. And I still did not tell them what I did. The tool worked fine so they had no idea what happened because this project had no version control and I pointed them to a new directory with the new code with a first commit.3 -
So we had this legacy Objective-C codebase for a mobile app that was actually pretty good: I'd inherited the codebase and spent the past several years gradually improving it and I was actually quite proud of the work I put into it. So of course management decides to scrap it (with NO consultation from the engineers) and outsource a complete rewrite of the app in C# for Windows Universal.
Let me tell you. That code was without a doubt and without exaggeration the *worst* code I've seen in my close to 30 years of experience as a developer. I mean they broke every rule in the book, I'm talking rookie mistakes. Copypasta everywhere, no consistent separation of concerns, and yet way too many layers. Unnecessary layers. Layers for the sake of layers. There was en entire abstraction layer complete with a replicated version of every single data class *just* to map properties in pascal case to the same property in camel case. Adding a new field to a payload in the API amounted to hours of work and about eight different files that needed to be modified. It was a complete nightmare. This was supposed to be a thin client, yet it had a complete client-side Sqlite database with its own custom schema (oh and of course a layer for that!) completely unrelated to the serverside schema, just for kicks. The project was broken up into about eight or nine different subprojects, each having their own specific dependencies on various of the other subprojects in such a tightly-knit way that it made gradual refactoring almost impossible. This architecture was so impressively bad, it was actually self-preserving!
Suffice it to say it was a complete nightmare, and was one of the main reasons I ended up leaving that company. So just sayin', legacy code isn't always bad. :) -
Why Gmail. Why the fuck do your search parameters, especially your date filters, not work anywhere near as expected.
You make me have to query and test, query and test, just, randomly fucking guessing because, fuck it, right?
With a good 10 second refresh time. I love twiddling my thumbs and pulling my hair out.
after:2018/11/1 should produce emails from Nov 1st onward.
Not, TODAY ONLY, if no other parameters are
specified.
If there's a from: parameter, now we want to do after Nov 1st, right?
And also, don't show me how to sort in reverse order, either. Not without a complete rewrite of my class there, which clearly I'm too lazy to do right now.
Fuck the Gmail Api, responsible for weeks of wasted dev time... or more aptly put, "fuck devs using our gmail api" says the maniacal, sociopath devils that created it
fuckers.1 -
After reading some rants abut stupid project managers I remembered this situation that happened to me a decade ago.
One of the tasks was to move some html component to different place on the page. The whole page was a mix trs and tds and to achieve that I had to rewrite the whole page structure. I estimated around half a day to complete that task. It was my first job and I was not great back then, but still it was reasonable amount for this task.
Now lets introduce my PM : the guy was a complete tool. He was a former hardware store manager ( ͡° ͜ʖ ͡°) and had no idea what we were doing.
He started ranting how on earth such simple task can takes so much time. I started explaining myself, but he wasn't listening. Instead he started sharing his screen, he made a screenshot of the page, pasted it to the ms paint, cut the component, and moved it to desired place. Then he said : It took me like 10 sec to complete the task and I have no experience, maybe I will replace you?
I was speechless. I had no words and I just kept silence.
Then he said he would reassign this task to X, because he is competent.
X spend more then 4 hours and I heard no apologies.6 -
A few years ago I worked at company specialized in Magento(eCommerce) and Magento was changing their licensing model. At the time they had 3 Versions. Community(free), Pro and Enterprise.
They decided to ditch Pro and either make all migrate to Enterprise(with a discount) or go community which wasn’t really compatible. So some shops were in need of a more or less complete rewrite.
My hdd crashed literally the day before but hey no big deal everything is 99% done and on staging. So I had a Trainee at that Time and thought the last few crappy things could be done in pair programming so he can learn a few things.
But fuck him! That motherfucker! He managed to WIPE the staging server and no that was at a time without gut and no SVN. That dipshit just deleted 2 months of work because he thought it was a good idea to SYNC his empty project to the staging system.
Oh god I nearly stabbed him. He did that shit out of his own mind even though I told him a dozen times what would happen... we had to do the whole thing again with me sitting next to him watching every stroke he made.
Guess he learned something while inward silently raging the next weeks.1 -
Complete disaster. As a C++ dev I was assigned to maintain some Python applications (some of them acting as proxies, wtf) just because the original authors left the team. It's slow as hell and it's not even a product - it's a helper tool. Cannot rewrite because nobody will give a green light for that. Why? Why?!4
-
Any other IT company is like:
* Task -> Designer -> Markup coder -> Backend -> Finish
Our IT company:
Act I: "Art of setting up contact with idiots".
------
Items:
*Cave scripts (aka "typical task")
Designer: -- "DAFUQ?"
Customer: *gives another interpretation*
Designer: -- "Erm... really? White text on white background?"
Customer: -- "Make a decision by yourself. I was expecting much more independence from you. You are an expert after all."
Designer: -- "Well. I'm making decision by myself. The text will be placed *here* and will be gray-colored, because *bla-bla-bla*"
Customer: -- "I disagree."
Designer: *1 hour of silence later* -- "Well...k."
Act II: "Design meets ar(u)tist"
----
Items:
*Something, that was drawn by dumb kid while smashing his own head against desk. (PSD layout)
* Salt (to pour it on open wounds)
Designer: -- "I'm seeing this task *this way*"
Markup: -- "And how do u think i should get this done? Have you even seen what you made?? This is bullshit!"
Designer: -- "It's not bullshit! It's a sci-fi themed layout!"
Markup: -- "With gameplay elements and graphics from Alien Shooter??"
Designer: -- "Well, I don't care." *brings new edits and changes*
Markup: -- "????"
Designer: *smug face* -- "!!!"
Act III (7 days later, 9 hours till deadline): "Short story about boy, who was trying to hang himself, but instead fell out from window."
----
Items:
*Markup, smelling like it went through hell and back (x1)
* Markup coder with fried butt (x1)
Backend: -- "What. Is. THAT?"
Markup: -- "It's a work we should complete in 9 hours."
Backend: -- "WE?? I know u mean me, but that's a nightmare. What the f*ck were you doing all this time?"
Markup: -- "Well..." *finds out that he was only watching films and sleeping* "I was making this thing up..."
Backend: -- "You mean "f*cking" *this* thing "up"?"
Markup: -- "Not without it"
(*3 hours of edits and changes of color from white to white later*)
Backend: -- "Well, let's do this."
*Picks PHP and tries to bundle it up with MongoDB. After some time tries to rewrite everything to JS and starts shouting something like "F***CK" and looking for window to walk through. Figures out that he is on first floor. And that he is too lazy to go upstairs*
Act IV (3 days after deadline): "Pain and misery":
-----
Items:
*Something covered with insul(t)ating tape. (Final product)
Customer: -- "Really?"
Team: -- "Kinda."
Customer: -- "Well, thanks for your work anyway. It feels like it's going to disassemble right in my hands but it just works. Oh, also, you didnt made this in time, so your payment will be over9000 times lower. That's all"
Backend, on fluids: -- "Well...yeah..."
Markup: -- "Don't look at me like that. I really was doing my job."
Designer, with twitching eye: -- "Huh, I see. You worked so hard that we have nothing to eat now. Thanks for that."
Backend: ...1 -
Started a new web dev job, how long until I can tell my new boss that the codebase sucks and needs a complete rewrite?2
-
Take a computer, any computer. Remove all the useful shit that makes the operating system actually work and rewrite all the prompts in simple English. Just for fun, give it the complete inability to perform everyday tasks that a child could use an Etch-a-Sketch for. Put all the useful settings/options in the background and make updates look like they're going to install up until the last minute until they error with "Unknown" as the only reason. Now interface it with your cobbled-together backend system and REQUIRE it to communicate with that system for EVERY task.
Then paint it silver.
There. Now you have a Mac.
Fucking toys!13 -
We had to finish a student project, our project leader had done some work on the project before and his algorithm to detect an event was complete bullcrap.
I took it in my hands to rewrite the algorithm with the risk that it doesn‘t work and i hadn‘t done any work i could show, which would result in a bad grade.
Luckily my algorithm worked better in orders of magnitude and we managed to deliver on time -
As you grow older, both professinally as a dev and as a team player, you realise that a complete rewrite is rarely the better answer to the problem at hand.
With that being said, I'm rewriting the glorified-mass-of-infernal-human-feces-with-corn-bits-masquerading-as-mere-shit out of a production service right now. Wish me luck.2 -
When you give a basic touch of modern design to a README and critize their replies they end the conversation with
'locked and limited conversation to collaborators'
'We appreciate the effort'
Sure doesn't look like it.
'X is highly specialized software'
Like most other software? And?
'The docs are fairly out of date, and need a complete rewrite, not this kind of graphical adjustment, so it would do more harm than good to present information of how to run this application in a secondary page along with random outdated info.'
So you are too lazy to update them, probably won't for a long time and have a problem with updating the outdated information's design despite that not actually changing the situation.
Disregarding the fact that the 'graphical adjustment' work even if you update the content.
Got it, right.36 -
12 Stages of Software Development:
1. Analysis.
2. Development
3. Realization the whole analysis is complete bullshit and has nothing to with reality.
4. Denial about failing deadlines.
6. "Acceleration": adding more people to the project, bringing out big corner cutting machine.
7. Learning that massive amount of new features needs to be added, while the deadline is two weeks away.
8. Putting some random crap in production, riddled with horrid bugs and security flaws, to technically not miss the deadline.
9. Get the mess almost working long after the deadline has passed.
10. Maintain this steaming pile of crap for a year.
11. Start planning for full system rewrite that "Makes Everything Better".
12. Goto 12 -
After completing urgent projects that rewrite a lot of "overtime" or even personal projects that rewrite like a few weeks to complete.
Old web dev days pre-BS... All the CSS needed to and the cross browser compat. The divs don't align... Can't get elements to center... Or it does, just not in browser X...1 -
I swear the longer I use AI for auto complete in my IDE the worse it gets
today it can't even get the syntax of basic method calls right
last few weeks it was failing to rewrite code I already had in another part of the codebase and was just hallucinating confused garbage instead -- it knew what I meant but it was just not there right
when I first got it I could write comments and then autocomplete the code but that stopped working at some point and I just thought maybe I was asking things that were too complex for it, but now I'm thinking all these things degrade over time and can't not degrade for some reason
they keep claiming they don't learn but if they degrade they must be doing some sort of feedback system
I remember back when they did IBM AI and such and that stuff degraded as well, then AI fell out of fashion for a bunch of years4 -
So I recently finished a rewrite of a website that processes donations for nonprofits. Once it was complete, I would migrate all the data from the old system to the new system. This involved iterating through every transaction in the database and making a cURL request to the new system's API. A rough calculation yielded 16 hours of migration time.
The first hour or two of the migration (where it was creating users) was fine, no issues. But once it got to the transaction part, the API server would start using more and more RAM. Eventually (30 minutes), it would start doing OOMs and the such. For a while, I just assumed the issue was a lack of RAM so I upgraded the server to 16 GB of RAM.
Running the script again, it would approach the 7 GiB mark and be maxing out all 8 CPUs. At this point, I assumed there was a memory leak somewhere and the garbage collector was doing it's best to free up anything it could find. I scanned my code time and time again, but there was no place I was storing any strong references to anything!
At this point, I just sort of gave up. Every 30 minutes, I would restart the server to fix the RAM and CPU issue. And all was fine. But then there was this one time where I tried to kill it, but I go the error: "fork failed: resource temporarily unavailable". Up until this point, I believed this was simply a lack of memory...but none of my SWAP was in use! And I had 4 GiB of cached stuff!
Now this made me really confused. So I did one search on the Internet and apparently this can be caused by many things: a lack of file descriptors or even too many threads. So I did some digging, and apparently my app was using over 31 thousands threads!!!!! WTF!
I did some more digging, and as it turns out, I never called close() on my network objects. Thus leaving ~30 new "worker" threads per iteration of the migration script. Thanks Java, if only finalize() was utilized properly.1 -
Alright, it's been a while since I expressed my thoughts/feelings but here is what I'm dealing with.
Ever since I was a kid I've played games and even ended up enjoying the testing of new beta games more than actually playing games. The first games I played were atomic bomberman and worms. I was 4 at the time and lived in Denmark. By the age of 6 I moved to The Netherlands and have dealt with 8 years of being bullied for a reason I do not know. So as you can imagine I've dealt with a serious depression for a while and have always felt out of place.
Later after a few failed attempts of following an education I got into development. This was after I wasn't accepted into an education of game design. The course I follow now describes itself as application development but all we're doing there is building websites and not learning a proper way to keep code clean.
In the second year of the three year course we had to follow our first internship. This was the first positive thing I've had with school in my entire life. I ended up working for a company that had a game which tested your skill, the game was used by recruiters for bigger companies to pre select the right people for interviews. I had a look at the code of the game and it was a mess, after a couple of meetings further I managed to get them far enough that I could start working on a complete rewrite of their game.
So far it's been a rough road to becoming a game dev but I most certainly hope to own a studio one day. Now I only need to manage until I've got there3 -
Almost any internally tool developed by a non-dev who has read a book 'learning to program in 21 days' and now thinks he can code. Usually it is developer in excel and as years pass complete departments depend on it until the moment a consultant is hired to completely rewrite it without any specs.
-
This was a few years ago in my 2nd year of college. my very first foray into web dev for a team project. long story short, I wrote over 16,000 lines of code. teammate #2 wrote barely 1,000 lines. teammate #3 wrote around 250 lines, around 200 of which I had to rewrite anyway because it was such complete trash. and yet, still, it was this class that showed me I wanted to go into web dev. LOL3
-
One time I got a Skype interview and the interviewer asked me to complete all the coding questions (rewrite in actual code) and email the answers to them within 10mins.
But when I open the question sheet, I found that all questions contain pseudo answer, so I ended up rewriting them on specific coding language, which was easy.
After I finished all the questions and sent the answer back to the interviewer and she told me this test wasn't testing my skill level on that specific coding language but honesty.1 -
I have to rewrite a good chunk of logic because it is too hard for any one of 4 people to complete a 5 minute job within 7 days.
I hate users. -
Not a horror. I'm rewriting services.
It started as a help request. I was asked to help with completing a service dealing with push notifications which was a research prototype. It was suggested to keep core part of it, but it was so awful that I just removed all files and wrote the service from scratch.
The second service had been developed for more than a year by a junior and then by our manager who wanted to complete it as fast as possible, without taking care of code quality. Then I was asked to take over the project and after some time I agreed with one condition: I'll have 1 month on takeover. But when I looked at the code, it became clear that it's much faster and better to rewrite everything except API and database than to takeover existing code.
The third service dealing with file exchange was working, but the junior who wrote it advised to rewrite it because it was a very simple service. So, I initiated rewriting, designed a new API and reviewed the final result.
And now I'm dealing with the fourth one. It was developed in my team but not under control. Now, when I "inherited" this complicated project, I decided to rewrite it because it should be simple, but it doesn't. It features reflection, layers inside layers, strange namespaces, strange solution structure. And that's after months of refactorings and improvements. So, wish me luck because I want to keep part of the infrastructure, but I don't know if it's possible. -
Deep learning
I thought it would be a great course, learn some of the stuff that I always read about but couldn't understand jackshit, and maybe profit form it somehow.
I'm in my last assignment, they want us to pick some SNLI paper and implement, ok, so I find this one with the least amount of params because I thought hey this seems promising.
And boy what a ride it was, I implemented it using PyTorch, the results are way off, I read the paper again and rewrite some parts, still nothing, I get 79%, it's supposed to be 85%, and no matter how I try, nothing.
10 GitHub repos later, 40 hours of complete meltdown,
20 throwaway Google accounts using colab because we don't have GPUs in our uni and using AWS is not feasible.
Same shit, I'm at loss, the world is a lie, and I fell for it...
Fuck.2 -
OpenSource is fun they said. I being a bored teen thought, ah, another chance to experiment. Discover something new. Now I am into piracy, movies, music, software. If I can get it for free I ain't paying for it. So I went on to GitHub to see what exciting new Repos I could contribute to. I hate already implemented plenty of algorithms in GO for GitHub.com/TheAlgorithms so I was looking something more practical, more beneficial to society. Then I saw it, the perfect repo, not too complex and not amateur. SpotDL/spotify-downloader for downloading songs from Spotify, a grey area coz it's technically piracy. Well not from Spotify, we fetch the info from the Spotify API and search for the songs on YouTubeMusic. They were just about to release v3, a complete rewrite of the codebase stressing code readability and stuff. I spend about a day studying the codebase, trying to findout just where I could make my contribution. I can see outright that there's a huge problem with implementation.
First of all the script spawns 4 processes for downloading songs though you might be downloading only one song. Which means for everytime you run the script you have to wait for 4 other processes to be spawned before any downloading can happen. Sure this is faster when you are downloading more than like 4 songs, but it's actually slower when downloading a single song. But I ignored that coz I assumed that most users download playlists and albums. Anyway we talked with the like lead developer and he was all like, make those PRs anytime you feel like. So I made a really minor first contribution.
I introduced download from Spotify URI functionality, modified like 10 lines of code. I was half expecting that the PR would be merged within hours at most 24 hours coz of how minor of a contribution it was, 5 days in it was pending. So I tagged the lead Dev and he was all appreciative of the PR, calling it real 'clean code' and stuff. 3 more days, the PR is still not merged. I have now stacked 4 more commits to the same PR, I tag the dev and he's like he's waiting to see if my 'feature' will get atleast 10 upvotes so that it can be merged, he links an issue. I go to the issue and my feature is not there, So 11 days after I made my PR I have to write a comment explaining the 'feature' introduced in my PR and then wait for 10 upvotes.
I was like f**k this, I'll just develop on my fork if you want the features on my fork, you will make your own PR! I am so done with OpenSource, development is slow. I have no idea how you guys do it. I can't handle development where I don't have write access.6