>working in highly complex and specific banking software
>get assigned the task of taking a 20 year old program and rewrite it with modern libraries and more specifically, security standards
>spend almost 2 weeks literally copy and pasting code and then fixing everything until the compiler stopped bitching
>enter phase 2 where I individually run parts of the program and look for 'errors in translation' by comparing results with the original program so that I confirm that everything is working as intended
>a really, really fucking long function that reads through a database of god knows how many accounts and transactions, sums a lot of stuff, classifies a lot of stuff, and from that returns a 2 page Crystal Report (updated to flexreport by me) report.
>notice that there is something that the original program is doing and mine isn't
>start dissecting the program step by step to see where exactly the discrepancy is
>there it hits me
>these are actual variables: >m_sum1,m_sum2,m_sum3,m_sum7,m_sum10,m_sum100,m_sum200,m_sum99,etc.
>the nombres don't even note an order, as I said it goes from 3 to 7 and from 7 to 10 and then to 80, 89, 99, whatever
>no apparent context whatsoever to even understand what is the difference between sum 1 and sum 2
>People tell me about the good old days when there were no undergrad CS majors who did this kind of bullshit
>People tell me about the good old days when only true professionals and not codemonkey studies (CS) majors wrote code
>mfw this is all bullshit
>mfw there were no good days
>mfw there won't be good days ever
>mfw it started like shit and will only get worse
>mfw I'm a mathematician trapped in the sins of retarded computer scientists of the past and present
You all bullshitted me.
>>53924367
Fuck your memes
I've seen some shit as well, at the last job I worked at most of the project's old legacy code was written by some retired self taught coder who thought that it was the shit to write 3000 line functions in VB6 that pass all parameters by setting big scope variables and calling void functions calling void functions that read and alter them, and then copying and pasting the whole trainwreck into C# and tweaking it here and there until the compiler stopped complaining
>>53924540
That sounds even worse. It really is a big problem when you need to fix a specific issue. Today the clock hit 4 before I finished fixing all of it so tomorrow I will go back there, sit down, and continue dying inside.
I cannot understand how someone couldn't have the time to simply name their variables for what they are instead of just what they do.
Was code ever good?
>>53924335
By the time he got to m_sum8 he wasn't sure if he had already used it, saved time and went to 10. Sparse ordering predicted by human immediate memory and limited by typing laziness
>>53924703
It's sort of like how they do street addresses in rural areas
Your don't know if there's going to be a variable between the two at some point in the future or not so you just go coarser with your numbers
>>53924335
Coding was never done "right", anon.
It's only up to you to be THAT guy.
The one who builds things in a logical manner
The one who actually annotates the crap he does
The one who writes documentation for his program
The one who makes the next age of Programmers smart like you.
>>53924335
Don't stress anon.
I work building experimental aircraft. You'd shit a brick if you saw some of the embedded programming and worry about flying again. You'd shit a bigger brick if you saw how the aircraft are assembled and put together.
>>53924335
>20 year old program
So 1996? (fuck I feel old)
>People tell me about the good old days when there were no undergrad CS majors who did this kind of bullshit
>People tell me about the good old days when only true professionals and not codemonkey studies (CS) majors wrote code
CS existed and sucked back then. 1968 and earlier were the age of good CS guys
>>53925219
Fuck. So the reason people in the big software companies get paid so much is because they have to pay for their psychologist and all the calls to the suicide hotline. I wonder if you get taxed for that shit.
>>53925344
That is reason #2 for why I didn't go for engineering.
When are we passing legislation that makes it illegal to write bad code?
>>53925349
>1968 and earlier were the age of good CS guys
I'm pretty confident in that we write better code than dinosaurs.
>>53925349
> pointing out 1996 was 20 years ago
Don't do this to me anon. I was old in 2010
>get done with big project at work
>time for me to change gears on work on the next big thing
>have to add a feature to a project no one has worked on in over year
>original 'engineer' who built the whole thing long gone
>all of the code is in 3 giant files
>random 'optimizations' that do nothing and make the whole thing more cyptic
>single character global variable names
>documentation doesn't describe how to even test, making it up as I go
>some things are completely wrong logically, but I'm worried fixing them will break the rest of the program
>comments are memes
Just when I was starting to like my job, this shit.
>>53925219
>because no one actually knows how any of this works
true dat. Trying to find how to do complex stuff in xaml is horrible.
>>53925219
Holy crap is this real
even if you know the existing code sucks, and even if you try and come in and "modernize", your refactoring will probably suck.
In the old days, people wrote their programs on paper and carefully checked it for mistakes before sending it to someone who has access to a computer. It took a ton of time and it wasn't feasible for them to write monkey code by brute forcing the compiler errors.
>>53927869
pfft
back in my day we didn't have compilers. if you screwed up your code it just wouldn't work and you'd need to review your punch cards by hand to find the problem.
>>53924335
And it may not have been a CS person who did it. I convert shit all the time that was coded by engineers. Also, is Flexreport any good? My business uses DevExpress but in your situation, I would have cranked out some RDLC thing (given winforms) that satisfied the user and to minimize deployment files.
My related recent:
>replacing old calculation algorithm for client but don't have original code
>tsql...should be relatively easy
>PM hands me nothing hardware vendor's docs to go by (white paper and such)
>code per docs
>90% done, PM gets old code (in pl/sql)
>horribly different from vendor docs calculations
>original programmer fucked up factors, inversions, units
>original programmer fucked up a couple of Taylor series'
>impressed with myself that I actually remember enough calculus to spot it
>impressed that I'm coding for a vendor writing CRUD software and actually using math
>finish code that works *exactly* per hardware vendor docs
>post code
>PM tells me client wants it like old program so a decade's worth of calculated records aren't all of a sudden "wrong"
>turn shit over to lackey so I can work on something else in my pile (I'm lead dev. so I can do this)
>>53924335
>being a spic
just fucking kill yourself
>>53927911
Keep /pol in /pol. At least he has a well paying job.
>>53927891
The componentone flexreport thing is very good and just easy as fuck.
Anyways, from what university are people coming out that they forget factors and how to work with units.
>>53927911
I was actually surprised when I re-read my post because I didn't even notice that I had written a word in spanish. I don't know how that happened, maybe my brain crashed for a second and when it restarted fucking windows reverted back to spanish keyboard or something,
>>53927471
>comments are memes
kek
>>53925344
>You'd shit a bigger brick if you saw how the aircraft are assembled and put together.
pls elaborate
>>53924335
>mfw it started like shit and will only get worse
That's not necessary.
Hell, you're whole job was making things better.
>>53934683
5/32 rivets
>>53935902
thats some poo in loo tier stuff
