|
Powerful Two-Hander posted:then every so often I go and read my own docs and am like "Heh that is pretty cool as a profession we've built feedback into our processes compilers tell us almost immediately if we've made syntax errors or type errors test tell us if our program works at all code reviews tell us if our peers can read what we wrote (and plangs and poo poo jobs are the places that don't have these things) but no one will read your documentation so it's a fundamentally hard and lonely thing to do
|
#
?
Feb 17, 2019 15:51
|
|
"
|
|
| # ? Dec 6, 2025 09:21 |
|
|
to clarify, even if they can read it they probably won't have the time and energy to give it the thought necessary to evaluate it beyond did this grammar for commits at least what i'm trying to do is make a story for why i did what i did, and it might be a summary for several days worth of work. i can't really evaluate it because i have all the context that went into it hot & fresh in memory. if i did it well, i'll have enough information that i can reload that context later but i can't really know until later ofc there are indications but no dopamine wired response to a green orb
|
|
#
?
Feb 17, 2019 16:04
|
|
|
mystes posted:Using VBA? Why would someone do this in Excel? In my first job we did a lot of VBA stuff but one of our customers was a charity, their computers were ancient and not all the shops had Internet, this would have been 2003/4. But they did all have Excel. So we did a management system for them to track thier clients, the central system would produce Excel spreadsheets with embedded macros in them for the current period, they'd post them out to the offices in floppies. Every month they'd post them back in and they'd load them all back into the central system. It worked surprisingly well.
|
|
#
?
Feb 17, 2019 16:04
|
|
|
that was how i first learned to program, i was doing poo poo in excel and then got frustrated by its limitations (for example the nested formula limit that I think is gone) and started migrating the more complex stuff into vba procs, then using them from excel. it's a slippery slope!
|
|
#
?
Feb 17, 2019 16:13
|
|
|
Brain Candy posted:as a profession we've built feedback into our processes there is one huge exception to this rule which is apis. please document your loving apis. no, the unit tests you wrote to verify that your implementation of the api is correct do not count as documentation. no, your vague explanations of some specific use case on some random mailing list three years ago do not count as documentation. no, a wiki page with “coming soon!” does not count as documentation. no, the output of your language’s automatic api documentation tool does not count as documentation, if you have not in fact written any documentation so it’s just a plain list of functions and parameter types. write some goddamn documentation if you want people who do not work in the same room as you to use your poo poo without cursing your name and your offspring even unto the seventh generation.
|
|
#
?
Feb 17, 2019 16:17
|
|
|
Powerful Two-Hander posted:
sup buddy, the feeling of "thanks me a year ago for {writing docs, using strong types, making the right extension point}" is great
|
|
#
?
Feb 17, 2019 16:32
|
|
|
Schadenboner posted:Could do "you" as a semi-vocalized contraction? "Y'make"? this is what I had in my head but failed to get on the page call it a poorly documented couplet
|
|
#
?
Feb 17, 2019 17:40
|
|
|
Soricidus posted:there is one huge exception to this rule which is apis. please document your loving apis. a number of teams I deal with suffer from this. I also have to endure teams who only do documentation on a Splunk site where api examples are randomly incorrect, so you have wait days until their team comes back with “oh yeah we removed the need to enclose that in quotes recently whoops” and then not update the page
|
|
#
?
Feb 17, 2019 17:57
|
|
|
apis must be documented.
|
|
#
?
Feb 17, 2019 18:03
|
|
|
Soricidus posted:there is a fundamental law of software that states that the only documentation that ever turns out to be important is the documentation that didn't get written Working examples covering your obvious cases are a) often the most useful documentation b) used (but I guess not read so much as copy pasted)
|
|
#
?
Feb 17, 2019 18:12
|
|
|
just post your poo poo to stackexchange nbd
|
|
#
?
Feb 17, 2019 18:12
|
|
|
gently caress every tool that is configured using a source controlled dotfile I blame .gitignore for starting this trend
|
|
#
?
Feb 17, 2019 18:21
|
|
|
Sapozhnik posted:gently caress every tool that is configured using a source controlled dotfile why's that? some things shouldn't be source controlled and are (i hate finding .idea checked into a project) but things like .gitignore seem fine to me.
|
|
#
?
Feb 17, 2019 18:50
|
|
|
Sapozhnik posted:gently caress every tool that is configured using a source controlled dotfile checking your CI configuration into source control owns actually
|
|
#
?
Feb 17, 2019 19:03
|
|
|
DONT THREAD ON ME posted:why's that? some things shouldn't be source controlled and are (i hate finding .idea checked into a project) but things like .gitignore seem fine to me. it's me, 85th commit "Removes ancillary IDE files"
|
|
#
?
Feb 17, 2019 19:14
|
|
|
speaking of commit messages, i still have no idea if they should be 1) i, developer, "fixed the bug" 2) applying this commit "fixes the bug" 3) commit, "fix the bug"
|
|
#
?
Feb 17, 2019 19:15
|
|
|
cinci zoo sniper posted:speaking of commit messages, i still have no idea if they should be 2 imo
|
|
#
?
Feb 17, 2019 19:18
|
|
|
cinci zoo sniper posted:it's me, 85th commit "Removes ancillary IDE files" required dev computer setup should include a ~/.gitignore_global
|
|
#
?
Feb 17, 2019 19:20
|
|
|
|
|
#
?
Feb 17, 2019 19:26
|
|
|
Soricidus posted:
Ugh, too many times I've stumbled upon paragraphs of "goal and design" of the api, and no actual coding examples.
|
|
#
?
Feb 17, 2019 19:26
|
|
|
cinci zoo sniper posted:speaking of commit messages, i still have no idea if they should be 3 is the most common in style guides
|
|
#
?
Feb 17, 2019 19:29
|
|
|
3/imperative feels the most natural to me when reading through anything With 1/past tense, we already know it happened in the past, time is linear 2/Present tense also reads weird because it's not constantly happening, it happened at a specific point in time Using the imperative also asserts dominance over the computer, therefore delaying a skynet scenario
|
|
#
?
Feb 17, 2019 19:36
|
|
|
mod saas posted:3/imperative feels the most natural to me when reading through anything yeah this. 3 feels weirdest to write but reads way better imo and commit messages are read more often than they're written.
|
|
#
?
Feb 17, 2019 19:54
|
|
|
cinci zoo sniper posted:speaking of commit messages, i still have no idea if they should be
|
|
#
?
Feb 17, 2019 20:17
|
|
|
Fatty Crabcakes posted:5) "minor refactor, fix bugs"
|
|
#
?
Feb 17, 2019 20:25
|
|
|
one of my hobby projects has "commit whatever" as a prominently displayed commit message and when i started my first computer toucher job i got pulled aside to check that i knew that that wouldn't be acceptable in a workplace env lol
|
|
#
?
Feb 17, 2019 20:26
|
|
|
cinci zoo sniper posted:speaking of commit messages, i still have no idea if they should be I do basically 2: applying this commit will "fix the ID10-T error in frontend"
|
|
#
?
Feb 17, 2019 20:38
|
|
|
vodkat posted:Think I've got to the bottom of this, it wasn't this code that was the issue but the way in which I was applying the function using groupby, which due to a duplicate index or something (which I didn't bother trying to figuring out) it was sometimes passing more than one row to this function. I think I've fixed this by stopping trying to be clever with groupby and apply and using the much slower df.iterrows() instead. my first guess was that you were `apply`ing and not specifying the axis so pandas was getting confused about if you wanted to apply to the dataframe or the row or what sorry for not posting it but i was really drunk at the time and just assumed i was wrnog
|
|
#
?
Feb 17, 2019 21:20
|
|
|
urgh embedded programming is so bad, thousands of unspoken assumptions that crash your program if you so much as blink at them wrong. how does anybody get anything done like this
|
|
#
?
Feb 17, 2019 21:24
|
|
|
gonadic io posted:urgh embedded programming is so bad, thousands of unspoken assumptions that crash your program if you so much as blink at them wrong. how does anybody get anything done like this By using embedded Linux instead OP.
|
|
#
?
Feb 17, 2019 21:29
|
|
|
ratbert90 posted:By using embedded Linux instead OP. you've honestly convinced me. i was only using a samd mini for a nice and small form factor but i should be deving on my pi. i wonder what it's like using the rpi serial ports from rust? probably fine honestly and sure as gently caress better than i'm doing now
|
|
#
?
Feb 17, 2019 21:46
|
|
|
gonadic io posted:you've honestly convinced me. i was only using a samd mini for a nice and small form factor but i should be deving on my pi. i wonder what it's like using the rpi serial ports from rust? probably fine honestly and sure as gently caress better than i'm doing now Yeah, with things like the OrangePI being so cheap and small, there’s really very little reason to go full embedded these days.
|
|
#
?
Feb 17, 2019 21:59
|
|
|
ratbert90 posted:Yeah, with things like the OrangePI being so cheap and small, there’s really very little reason to go full embedded these days. and yet you see it all the time in a lot of places usually because its how a particular shop has always been run
|
|
#
?
Feb 17, 2019 22:55
|
|
|
cinci zoo sniper posted:speaking of commit messages, i still have no idea if they should be a functioning human being can understand all three equally well as they mean the same thing and as a result does not care in the slightest
|
|
#
?
Feb 17, 2019 23:16
|
|
|
Boiled Water posted:and yet you see it all the time in a lot of places both have their place, if you need a tight loop with lots of DMA use and rapid timing and digital/analog work going full embedded works better Linux is mainly good when you need to do complex logic, work with graphics, run a server or do other stuff that needs libraries for a reasonable implementation the device I (soon-to-be-used-to) work on has 3 different complexity levels of embedded socs/mcus in it, with embedded Linux at the top e: one example where a low-level mcu is absolutely necessary is analysing NDIR signal, or waveforming ultrasound; or really anywhere where you regularly need sub-microsecond precision. also its cheaper (it's like $0.50 per piece for a typical STM32F0) than a full Linux soc, which matters if you are not doing hobbyist things. oh and uses up orders of magnitude less power Private Speech fucked around with this message at 00:12 on Feb 18, 2019 |
|
#
?
Feb 17, 2019 23:47
|
|
|
NihilCredo posted:a functioning human being can understand all three equally well as they mean the same thing and as a result does not care in the slightest most attempts to apply grammatical rules to commit messages are intended to stop dipshits from submitting "asdf" and those people are already unhelpable
|
|
#
?
Feb 18, 2019 00:58
|
|
|
back in the early 90s i worked in a recording studio where a guy insisted on labeling his disks "dope poo poo" and "dope poo poo 2" and "dope poo poo 3" and i couldn't convince him why it was a terrible idea until he realized it on his own the hard way several months later
|
|
#
?
Feb 18, 2019 01:01
|
|
|
yeah, i wanna go back to that one dope beat we recorded last february...uh...help me out here...i think it was somewhere between dope poo poo 4 and dope poo poo 76
|
|
#
?
Feb 18, 2019 01:03
|
|
|
bisect still would’ve worked ok
|
|
#
?
Feb 18, 2019 03:35
|
|
|
|
| # ? Dec 6, 2025 09:21 |
|
|
hrmm i have a list of dicts where a certain key contains a list. i want a merged list of all those lists, but my list comprehension comprehension is lacking... halp? code:code:Carthag Tuek fucked around with this message at 13:05 on Feb 18, 2019 |
|
#
?
Feb 18, 2019 13:01
|
|
