Programming as a career

what's the point of writing tickets if they contain 1 useful bit of information, and 5 bits of misinformation?

channel_05 to 08 is actually channel_01 to 04

"pTO value should be adjusted appropriately." yeah, thanks for that.

the example file i'm to base everything on is for some test system that bears no relation to the target (actually, either of the two different targets, which is a pain in itself)

oh, and it needs doing for the end of the week.

is that a knock on qa?

i had an uncomfortable interaction once with someone because they thought that if you "keep it simple, it will be simple"

i told them i had one simple request for them to do on their tickets. the information they provided wasn't completely wrong, but half of it was based on a misunderstanding of how the code worked on the backend

was a task ticket. config needed generating and deploying so that things can flow through the system but, y'know, if the inputs i'm given are wrong AND the outputs i'm given are wrong then it's going to be all the harder.

(usually this configuration is 600 urls spread over 80 files. and there are ~30 sets of this. (i spent a bunch of time sorting out templates and scripts to generate this, which made helped) but these need a different (much smaller, thankfully) set of templates)

oh that sounds brutal. nothing like wasting time doing so much work for nothing

you know you're in trouble when the reference documents look like this

https://dashif.org/docs/DASH-IF-IOP-v4.2-clean.htm

wow. good luck, man. a related gripe i have about doing things that take a while is that the next person/middle management comes in and says how simple the solution was, because they're hardly aware of all the work it took to get there. everything seems simpler in retrospect.

tickets for system experts should be outcomes and not specific values imo

the problem is no one writing or (not in your case) reading tickets is an expert so it’s a shitshow

https://dashif.org/docs/DASH-IF-IOP-v4.2-clean.htm

LOL! we had a vendor come back to us today saying "the file you generated has a lowercase 'y' but it should be uppercase" ..and i flashed back to the discussion i had with my developers a month ago about how bizarre it was that it would be 'y' and 'N' but the docs clearly show a lowercase 'y'...

and then you're here* until 9:15 making sure all your ducks are in a row and the next morning the ducks prior to your ducks are nowhere to be seen, not a quack.

until, about 6:30 tonight when they'll start squawking, probably.

(* here being just sat on the sofa so no real hardship. and i did take an hour out to eat something and watch grand designs australia)

> the next person/middle management comes in and says how simple the solution was,

yeah, this. sometimes a ticket is only a 3 word change, but working out what those 3 words were, ruling out all the other words, and where they need to go can take a week. i do wonder how we managed before the internet.

career: team lead pushing to get us all promotions. but that entails writing 300 words each on things like "Flexibility" and "Influencing" and demonstrating same and that writing seems so far away from what i actually do, and i'm so terrible at it, that i really can't be bothered.

Buckle up and do it, it’s not like you’ll be held accountable to it, or pay someone to do it

18 months we've been doing this project (originally scheduled for 6 months, but hey...). yesterday something happened which broke everything (in the TEST environment) and we've no idea what is was or how to fix it. things are arriving at the first component and instead of being saved to s3 and passed onto the second component they are just stopping. no logging of errors. nothing.

also quite poor that it took 24 hours for us to notice it had broken.

^ someone updated the db to be Khaki with a capital K and our code is case sensitive. the input urls had khaki with a small k and didn't match.

that is so classic <3

we'd still be looking for the problem if it was a trailing space, say

(which is why to this day every time i log something i put square brackets around it)

one tiny button, 42 live aws instances redeployed... always a nervous time

go on holiday for a week (i say holiday but it was the same sofa, different laptop) and they rewrite all your code...

worst is this. i think i saw it first in perl but this is ruby, moving the condition after the action... my eyes see the call to error, and don't see the condition because who puts the condition on the end?

// Object must not have a thing
error("Something is wrong with {object}") unless thing.nil?

(also in real life the message is longer so the condition is much further over to the right)

(it used to say

// Object must not have a thing
if ! thing.nil?
  error("Something is wrong with {object}")
end

rubocop prefers the former for conditions with single lines. i think rubocop is mostly bad and the way we do whatever it says even if it's patently worse annoys me.

happy programmers' day btw, 256th day of the year...

(Gah! code mangled because bbcode complained like 6 times in a row. doesn't seem to like hashes or single quotes or something)

ha. that is weird. but i've never used ruby.

since it's programmer's day, i guess it's only good and proper that i make publicly aware my biggest pet peeve: not just lack of documentation but wrong documentation!

i'm going over this c# codebase and it's lightly sprinkled with comments here and there and some of it references outdated or blocks of code found elsewhere, lol

happy programmers' day!

I've been doing Ruby for 5-6 years now and the modifier syntax is actually pretty easy for me to read at this point.

Also, Rubocop hates this syntax:

if !... 

tbh, i kinda hate that syntax, it's very easy to overlook the ! especially if there's no space following

post-conditions make more sense if the action is the default rather than the exception

do_something() unless it_is_the_weekend

i might have mentioned this before but in a previous job someone wrapped virtually every line of code in a method called log_error()

log_error(perfectly_good_call());
log_error(no_problems_here());
log_error(this_works_fine());

it would only log if the call returned an error but it confused the hell out of me.

this is one of the things i like about (python) decorators. they make that kind of wrapping more easily ignored by everyone else.

it's a weird thing to have baked into your linter, but i guess whatever is making ruby simpler must butt heads with testing for false, as it can be counterintuitive?

Well I think that there are two reasons:

1. Sometimes people will write this:

if !...
else
end

2. Ruby has unless (which I personally hate), and the out-of-the-box linter favors using it to

if !

https://www.youtube.com/watch?v=rge17TciHfU

my friend sent me a screenshot of some code he had to pick up for maintenance at work and the first line was

String number;

off to a good start

I actually spend a lot of my time working on a system that was built from the ground up in Java by someone who must've been entirely self-taught, because the formatting and code style she used was unlike anything I've seen. She did wind up making her own int, String, and boolean objects and I feel like I have to answer once a week "what is an EsInt?"

xp

haha, at least it wasn't String n;

oh cos it's a string. this is why i'm not a programmer too.

there are a million reasons why you’d put something numeric in a string but don’t name it number ffs

I enjoyed this

https://www.youtube.com/watch?v=baY3SaIhfl0

The QA tester is doing a great job of QA testing there!

ha. it's either like that star trek video or a machine/robot spitting out line after line of detailed information super fast

i'm currently dealing with an insane amount of misspellings everywhere from strings/variables to text/messages we output to the user. when i train our users i actually need to refer to some of the text and, for the sake of simplicity, i need to use the really bad grammar. it's horrible!

can't run that ruby script that worked fine last week because someone has removed the gems it uses from the repository (ok, probably security related, but still a forced update, and it's never only one gem)

can't install aws stacks because someone has changed the script to use poetry to control dependencies now and not whatever we used before. and bumped the required versions of *both* ruby and python to versions released less than a month ago.

hours installing those, and the required pips and gems over a slow 3g connection (still not finished). and all i wanted to do was add *one* value to an aws stack.

i consider myself somewhat of an expert at "getting things to build" and i've had so many problems trying to get python shit to work that i don't even bother anymore

somewhat related: https://drewdevault.com/2021/11/16/Cash-for-leftpad.html

https://sean-parent.stlab.cc/2019/01/04/iota.html

so, thirty lines of java, which can throw exceptions in 5 different places within it. i did have try-catch around each one but it was ridiculous (and half as long again) so instead i put one try catch around the entire lot and it was much neater.

only i got moaned at in the code review that the block was too long.

the internet is no help. stack overflow suggests putting each bit of code that can throw exceptions into a different method which gains you NOTHING as far as i can see - the method will still throw the exception, still needs handling in the caller. and needs overhead with passing parameters.

also, "don't log and rethrow" was another comment. i guess that's so it doesn't get logged in the child and again in the parent. but when unit testing you don't see the parent's logging. so i'm leaving that, at least until the project is mor than 20% done.

i wonder if any of those companies that offer me jobs on linked-in that start at twice my current salary have such code reviews?

(nah, my imposter syndrome is already bad enough)

Will never understand religious adherence to coding guidelines. Also I hate the term "code smell".

I'm surprised exceptions are still handled the same way 22 years after I learned them. such an ugly mess.

just put `throws Exception` on every method and let the spring's global handler deal with it

You guys should look at Mike Nesmith's FB post from April 3. Made me LLOL. #OneInstructionPerClockCycle

> career: team lead pushing to get us all promotions. but that entails writing 300 words each on things like "Flexibility" and "Influencing" ... i really can't be bothered.

reader: didn't do this, didn't get a promotion, but did get the equiv payrise! (and paid in jan, but backdated to sept!). was the last thing the previous dept boss did before he left for sunnier climes, although it had been in the works for a while apparently. i think they are desperate to keep people.

doesn't stop the person above me keeping my things in review for ever over perfectly good design decisions. week's worth of code, 3 weeks in review and it's got so toxic that i can't bear to look at his latest suggestions.

(basically i encapsulated the logic in two static methods in another class with known inputs and outputs, keeping the parent class and *all the parent's tests* boilerplate. he wants the logic in the parent class where it means the parent class and the testing will be different for every future version of this code (it's a microservice, we'll reuse it). and the logic will go into predicate classes, which are generally 20 lines of code to call one if...)

people who insist on doing cosmetic refactors in the middle of fixing review comments. make the diff twice as confusing, drag the review process out a bit more.

github's new 3-panel diff is terrible, they've added the file tree as a leftmost column meaning there's even less width for the other two. everything is truncated or wrapped. i'm guessing their designers all have 4k screens.

i have two 27-inch monitors and i find those aren't enough. we all work from home but another dev i work with tells me he has a huge curved monitor and every time he screen-shares he has to change the resolution so things won't be so tiny on the other person's screen

on another note, i took a sick day today because the last two weeks have been so incredibly busy i hardly had time to eat

Fortunately you can collapse that file tree sidebar, and it'll remember that preference.

people with those massive curved monitors seem like they hate their computers whenever they're on hangouts

wait what's the 3 panel diff, I only have the file tree (which I actually welcome, find it really useful) and the file changes

my monitor is like 22" or something and that seems plenty. the one at work is ridiculous, like 32" or more and people do have trouble with those when sharing on zoom because nobody can read anything