|
USSMICHELLEBACHMAN posted:i've only been programming 'seriously' for two years now, but testing is by far the most confusing issue i've dealt with. full test coverage is only a concept for managers; think about errors in tests. to actually check everything, you'd need tests for your tests, which would need their own tests, and so on tests are hard for newish programmers because you can't do it well unless you understand what 'words' you are making. see the cool fish video https://www.youtube.com/watch?v=_ahvzDzKdB0. it's hard because it's inextricably bound to design
|
# ? Mar 15, 2014 14:27 |
|
|
# ? May 10, 2024 09:24 |
|
i really like unit tests for testing computation. it's not so useful for testing i/o software engineering class in college distinguished between an error, which is a program in a bad state, and a failure, which is misbehavior due to an error. regression tests are useful because they check for errors that we know cause failures. a test for an error that never occurs or never leads to a failure is just a waste of time
|
# ? Mar 15, 2014 16:12 |
|
i have to imagine every time i run across a test on a model which goes code:
you should ideally structure your code so that the IO bits can be tested against real services and then the rest can retrieve mocked data. MYSPB (Make Your poo poo Pure Brah).
|
# ? Mar 15, 2014 19:07 |
|
You may or may not ever see this in a java code base. at least they will have tests though.
|
# ? Mar 15, 2014 19:07 |
|
FamDav posted:i have to imagine every time i run across a test on a model which goes this would actually be a cool feature though, hilariously, your IDE would need to understand dependency injection in order to figure out how to find aThing
|
# ? Mar 15, 2014 19:13 |
|
with all this talk of testing: are there any good resources that any of you have found that deal with writing tests for software that really hasn't had any formal testing written into it before? or just for help with writing better tests in general? bonus points if there are ideas on dealing with (horrible horrible) languages without much in the way of testing frameworks
|
# ? Mar 15, 2014 20:22 |
|
The Leck posted:with all this talk of testing: are there any good resources that any of you have found that deal with writing tests for software that really hasn't had any formal testing written into it before? or just for help with writing better tests in general? bonus points if there are ideas on dealing with (horrible horrible) languages without much in the way of testing frameworks Here's a good starting point USSMICHELLEBACHMAN posted:writing a test every time i fix a bug is one of the few things that makes sense to me
|
# ? Mar 15, 2014 20:58 |
|
The Leck posted:with all this talk of testing: are there any good resources that any of you have found that deal with writing tests for software that really hasn't had any formal testing written into it before? Suspicious Dish (iirc) pointed me towards Working With Legacy Code a few months back and it was pretty good
|
# ? Mar 15, 2014 21:06 |
|
in my very professional opinion TDD is better for designing than it is for testing
|
# ? Mar 16, 2014 00:21 |
|
just having yourself in the mindset of TDD forces you to adhere better to SOLID principles. if you're just programming to metrics like coverage it's a lot easier to write a few tests that offer huge statement coverage and actual value but aren't Unit tests. do something between what meets the requirement and what soothes the autism.
|
# ? Mar 16, 2014 02:43 |
|
writing a bunch of unit tests to make sure the thing meets requirements isn't worth it when, upon seeing the thing, your boss realizes those weren't actually the requirements
|
# ? Mar 16, 2014 02:47 |
|
JewKiller 3000 posted:writing a bunch of unit tests to make sure the thing meets requirements isn't worth it when, upon seeing the thing, your boss realizes those weren't actually the requirements
|
# ? Mar 16, 2014 03:28 |
|
Stringent posted:I want to work in a company where management writes acceptance tests and I code to them for a ridiculous amount of money. fundamentally, as a software developer, what drives your salary is developing the requirements any loving idiot can write code. this has always been true. it will always be true i have said it a hundred times in this thread: the only people who care about "code quality" are coders. the only thing anyone else cares about is getting the requirements right
|
# ? Mar 16, 2014 04:10 |
|
develop requirements => ship requirements to india => be disappointed => develop new requirements
|
# ? Mar 16, 2014 04:12 |
|
Bloody posted:develop requirements => ship requirements to india => be disappointed => develop new requirements yep as it turns out, software developers are the ones who have to establish and fill requirements no number of business analysts or managers will remove the need for developers to talk to actual users. that is what you are paid to do. that is where your expertise pays off sitting in front of eclipse and writing code is the least important part of your job
|
# ? Mar 16, 2014 04:13 |
|
im glad i am not a ~software engineer~
|
# ? Mar 16, 2014 04:15 |
|
this is also why the smartest coders you know are unemployed and why a lot of lovely java journeymen have solid careers and work 40 hours a week and take home bonuses
|
# ? Mar 16, 2014 04:15 |
|
I'm the smartest programmer i know and i have a job, what now
|
# ? Mar 16, 2014 04:21 |
|
Kevin Mitnick P.E. posted:I'm the smartest programmer i know and i have a job, what now you are probably underpaid because you think being smart matters even a little bit if you're smart, apply that brainpower to figuring out what people need. worry less about whether code is good and more about whether the app will deliver for users
|
# ? Mar 16, 2014 04:28 |
|
being smart matters but it's not coding ability, it's your talent at bullshitting and office politics why work hard when you can fail upwards
|
# ? Mar 16, 2014 04:29 |
|
Notorious b.s.d. posted:you are probably underpaid because you think being smart matters even a little bit but delivering for users is what the code is for? I'm not sure what you're getting at. focusing on business value i guess. which i agree a lot of times writing code is not the most valuable thing for me to be doing. right now i'm doing a lot of tier 3/ops support cause i've got more experience than ops with diagnosing loving weird problems totes agree that i'm underpaid tho Forums Terrorist posted:being smart matters but it's not coding ability, it's your talent at bullshitting and office politics if i had excellent social skills i'd be in sales, not dev
|
# ? Mar 16, 2014 04:42 |
|
Brain Candy posted:full test coverage is only a concept for managers; think about errors in tests. to actually check everything, you'd need tests for your tests, which would need their own tests, and so on hhahahah he lists eric raymond in his list of names l a m o
|
# ? Mar 16, 2014 04:47 |
|
https://www.youtube.com/watch?v=csyL9EC0S0c lol its that one guy
|
# ? Mar 16, 2014 04:54 |
|
Kevin Mitnick P.E. posted:if i had excellent social skills i'd be in sales, not dev this is the attitude i am trying to combat sales people are good at sales. this usually implies that they are very friendly. it almost never implies strong "soft skills" Kevin Mitnick P.E. posted:but delivering for users is what the code is for? I'm not sure what you're getting at. focusing on business value i guess. which i agree a lot of times writing code is not the most valuable thing for me to be doing. right now i'm doing a lot of tier 3/ops support cause i've got more experience than ops with diagnosing loving weird problems fixing bugs only matters in proportion to how much it affects users working on an awesome n**2 to n*log(n) transition matters exactly not at all unless users are complaining or it enables users to do something new with the product talking to users is like 99% of the useful part of your work. if you can't talk to an end user about what you are about to do to the product, you are a great target for outsourcing to alabama/india
|
# ? Mar 16, 2014 04:57 |
|
the light is very distracting in that video
|
# ? Mar 16, 2014 04:57 |
|
Forums Terrorist posted:being smart matters but it's not coding ability, it's your talent at bullshitting and office politics usually I find people who complain about "bullshitting and office politics" have no loving idea what they are being paid to do people are basically really high-end monkeys. what you call "bullshitting" I call "social grooming" and what you might see as "office politics" i see as the loving purpose of the office beep boop i am a robot why are we not all telecommuting guys?
|
# ? Mar 16, 2014 04:59 |
|
i wasn't complaining mind
|
# ? Mar 16, 2014 05:04 |
i have a terrible programmer e/n. i've been feeling kinda lovely the last week or so for a number of reasons i'd like to do some oss stuff but it feels like any of the stuff i'd like to work on i'm too behind in terms of understanding. taking elixir for example, it's a neat language and there are a couple cool projects, but the few things there are really heavy on the http/server poo poo (new frameworks being built, see: phoenix) or stuff like database querying (see: ecto). i get by decently enough with the rails stuff i do now but it's ultimately totally different from building non-insignificant programs/libraries is still over my head, and by the time i could contribute anything of value it'll either have already been implemented or something. a lot of what i'm used to essentially abstracts what's going on in the background, which is great until you need to understand it. basically i look at projects like those and still think "where the gently caress do you start" and it's the one thing that ever makes me regret not having a traditional cs education (to an extent). at the same time, as long as i'm working and getting paid i guess i could continue to take the who gives a gently caress approach with regards to doing oss poo poo, but i feel like it puts me behind. it's a pseudo-imposter syndrome that bothers the gently caress out of me and i'm not sure what to do. maybe i really just need to sit and read books or code bases for a few hours on end but i don't know if i have the mental patience, or maybe i just need to do something that isn't a loving web app. in a way i really hate what i do now which is be part of a place that does consulting (even if the senior guys are doing more the initial estimate poo poo) because i end up working on a bunch of vastly different applications, and the differences between 3-4 year old libraries is mixing my brain up and i can never 100% commit anything to memory. we have a bunch of rails 2 apps that still have to be moved to rails 3, and i only really know so many of the rails 2-isms, or we have old loving versions of capybara that i can't remember what the differences are from the newer ones and it slows me down so often. basically it's a massive level of inconsistency because i don't get to pick what we work on nor the stack in some cases and i think it's hindering my growth, but at the same time it wears down my personal drive to do anything else. that was longer than i wanted it to be but i feel pretty lovely and don't particularly have anyone to talk to about it. i'm supposed to have an onsite interview this week for a scala/ruby job and the context of what i'd be doing sounds really cool in theory, but i honestly have no idea what's going to happen or if i'd end up really wanting the job if i was fortunate enough to be offered it.
|
|
# ? Mar 16, 2014 05:21 |
|
double sulk posted:do something that isn't a loving web app.
|
# ? Mar 16, 2014 05:27 |
|
double sulk posted:i have a terrible programmer e/n. i've been feeling kinda lovely the last week or so for a number of reasons i'd like to do some oss stuff but it feels like any of the stuff i'd like to work on i'm too behind in terms of understanding. taking elixir for example, it's a neat language and there are a couple cool projects, but the few things there are really heavy on the http/server poo poo (new frameworks being built, see: phoenix) or stuff like database querying (see: ecto). i get by decently enough with the rails stuff i do now but it's ultimately totally different from building non-insignificant programs/libraries is still over my head, and by the time i could contribute anything of value it'll either have already been implemented or something. a lot of what i'm used to essentially abstracts what's going on in the background, which is great until you need to understand it. we live in a time when the best job available is dominated by the mentally ill
|
# ? Mar 16, 2014 05:38 |
|
double sulk posted:basically it's a massive level of inconsistency because i don't get to pick what we work on nor the stack in some cases and i think it's hindering my growth, but at the same time it wears down my personal drive to do anything else. jesus i see why you are called sulk don't sweat it so much. go to work. don't be the smartest guy in the room. as long as the checks cash, it's all good. seriously it is that simple. as long as someone around you is better than you are, you're learning. as long as your checks cash, you're getting paid. win/win
|
# ? Mar 16, 2014 06:18 |
|
also pretty much all code is legacy code written for a version of the library/framework you know from like five years go. that is not a consulting-related phenomenon. it will happen to you at "real" jobs just as often count yourself lucky for learning to roll with the punches this early in your career
|
# ? Mar 16, 2014 06:20 |
|
ultimately you don't need to be a programming god to do your job. The hypothetical perfect but spergy software developer is still thoroughly outclassed (and liked more) by average joe coder who can talk like a normal human to other humans. If you're trying to get some kind of life satisfaction out of being a good coder, it'll be a horrible path full of the worst kind of people. Also you don't know where to start because you didn't put enough time in on using/reading the library. Consider the average worker coming into a company needs weeks to get acclimated to a code base even with other developers helping. Contribute to OSS is a huge time investment and not worthwhile for most normal people who don't live and breath code. My contributions mostly came from being frustrated at a lack of a feature or a lovely implementation, writing my own and doing a merge request.
|
# ? Mar 16, 2014 06:26 |
|
oss contributions are your last desperate hope when your resume doesn't qualify you for anything. not coincidentally, that is why the idea comes up so frequently in discussions w/ people trying to break into tech
|
# ? Mar 16, 2014 06:28 |
|
Notorious b.s.d. posted:seriously it is that simple. as long as someone around you is better than you are, you're learning. as long as your checks cash, you're getting paid. win/win Time for me to sulk, I've kinda been stuck in a rut for a couple of years where I feel like I'm the smartest guy in the room and have nothing to learn from the rest of my team. I know the only solution is to move on, but there have been other factors keeping here and it's best gig in Mexico town afaik (the rest involves selling my soul to Indian near shoring agencies). This is the last year I plan to spend in this rut, and in the mean time I'm constantly playing around with Coursera CS classes and project euler problems to keep learning new stuff.
|
# ? Mar 16, 2014 06:46 |
|
Notorious b.s.d. posted:talking to users is like 99% of the useful part of your work. if you can't talk to an end user about what you are about to do to the product, you are a great target for outsourcing to alabama/india the problem here is the bullshit you'd have to go through within the company to be able to talk to the users
|
# ? Mar 16, 2014 10:29 |
|
Plastic Snake posted:420+219+21.9+(2.19*4.20)-4.20 = ~666 actually its the difference between god, nature, and satan - evolution (god's computers: the apple macintosh, ourselves, computers squared - complictated or evil computers - think android linux, or windows XP or 8) fun fact lets put this in other numbers - 2^4 = 16 (carbon, u, the shitposter), 42 = H20 = water (sea, earth) 1+(9-2) = 7 = nitrogen (air, fire) stick them all together and u get N20 - laughing gas - and N20(9) (H20)2 C (420) = the weed number look at it backwards, and you get funny stuff (laughing (N20) at (2) things (C)) if the things are computers theyre 64 bits = 2^9 * 2 = 18/2 = 9 so the funny computer number is for reelz the 219 the pos is the union of 420 and 219, the true chill place to hang out with ur bros for more on this exciting new videogame blend of progamming math and weed, keep an eye on ur local app store. itll be here in hovr form
|
# ? Mar 16, 2014 11:40 |
|
double sulk posted:basically i look at projects like those and still think "where the gently caress do you start" and it's the one thing that ever makes me regret not having a traditional cs education (to an extent). they don't really teach you this in CS courses.
|
# ? Mar 16, 2014 11:43 |
|
quote:double sulk posted: lol sulk but ill bite learn SOLID/TDD/MVC arch and doing things RIGHT. if you want to program a computer, do it the low effort easy way whatever happens. theres loving shittons of FOSS poo poo on github these days so just be smart and steal their poo poo. throw up some vms and servers, get comfortable with a bit of linux janitoring, and use the right lang for the job - whether it's C, ASM, C# ,Python (but never PHP kill you're are self first) if you're using easy tools (langs, libs), computers (buy a fuckin mac), and arch (mvc/solid) then that will stop you loving up for the most part and also make it way easier also buy 3 screens i reccomend that its pretty baller
|
# ? Mar 16, 2014 11:47 |
|
|
# ? May 10, 2024 09:24 |
|
the only way to contribute to an open source project is to use it and fix poo poo as you need it, unfortunately
|
# ? Mar 16, 2014 11:52 |