both for me!

sadly, sometimes I have to write golang too

I get paid to write (mostly) Java and (some) Kotlin

quite a few people do, but most oscillate between clojure and other langs as there aren't that many clojure gigs

Both

I write mostly Clojure and Kotlin

99% clojure for a living

Cool. I don't think I'll be able to convince others at my company to start writing in Clojure, but moving some things to Kotlin is nice.

I got Clojure introduced by writing several emails to the CEO, CTO, showing the benefits and doing a standalone application all in Clojure, which is processing over 250,000 trips per day.

Did a few presentations etc..

Both =)…

Don’t forget clojure is just a java library… As long as people are ok with writing xml to “script” stuff, why not add something else that’s a “smaller” xml…

@U0JUM502E we can't use new libraries without approval (in theory), but we do have a tools directory in one main repository, that's basically a free-for-all - there's even some Haskell in there

Could you add something to tools? I’ve heard of people doing similar things =)… You can whip up some pretty nifty tools in clj quickly ;)…

Yes, at some point. We have ambitious deadlines at the moment, and some people have been furloughed, so there's not much time for anything like that at the moment

Makes sense =)…

99% clojure for a living here...

what's the 1% @U0524B4UW ? CQL or Terraform (shudder)

Cobol

no terraform anymore... js, SQL, CQL, java, obj-c, python, ruby probably in that order of frequency, highest first

although i had to revise the order once i thought about it a bit

I am just lurking, hoping that one day I might be able to use Clojure on my job :white_frowning_face:

When we introduced Clojure at work (nearly a decade ago now), there was sort of a three-pronged attack: 1) we had a problematic process that folks had already tried to solve in several different tech: I tried Scala first, when I joined, and although it "worked" it had some limitations (and didn't fit the mindset of the team) so I rewrote it in Clojure and, although it was slower at first, it was more reliable and it better fitted the team's mindset (dynamic language). 2) we rewrote some low-level routines in Clojure to improve their performance and/or their Unicode handling. 3) we started writing some standalone tests in Clojure (replacing some brittle old Selenium tests with more flexible WebDriver-based tests).

1) was what really convinced management, once I tuned it a bit (making parts of it concurrent was a breeze with Clojure). 3) was what convinced the team I think. 2) was just icing on the cake -- but it paved the way for a bottom-up rewrite of our entire stack over time.

Anything to get rid of Selenium 😂

Ironically we replaced the legacy app that was tested that way with several new Clojure apps that are tested in a different way so we no longer have the Clojure-based WebDriver tests either 🙂

We broke 100k lines of Clojure this past week (22.5k lines of tests, the rest is production code).

you made me go and look @U04V70XH6!

gulp: 130kloc + 56kloc tests

Impressive percentage of tests! I've had a number of people comment that our test:code ratio seems pretty high at around 1:3 but yours is closer to 1:2!

ah, but my way of counting lines of code was very primitive - and tests tend to have a lot of literal data-structures and not so many function points, so they aren't really very dense. probably a better counting metric would return a lower test ratio

i did this:

find . -name \*.clj -o -name \*.cljs -o -name \*.cljc | grep /\*/src/ | xargs cat | grep "^.*\S.*$" | grep -v "^\s*;.*$" | wc

and this for tests:

find . -name \*.clj -o -name \*.cljs -o -name \*.cljc | grep /\*/test/ | xargs cat | grep "^.*\S.*$" | grep -v "^\s*;.*$" | wc

We don't bother to exclude comment lines or whitespace lines. I might think about updating our little shell script for that at some point. We don't tend to have a lot of comment lines (we mostly put that sort of thing in docstrings). If I mentally remove one line for each var and each function, that takes about 4k lines off our total.

huh, our 130kloc only grows to 134kloc with comments included, although it's 151kloc with whitespace too

there are lots of docstrings in the yapster codebase which probably read as strings rather than comments?

I think in my current project we're probably closer to 1:1 code to tests

I’ve no idea — my best guess is if everyone was doing butterfly you’d get less distancing because you take up more width in a lane.

so the cynic in me says the fact they’re saying this just means they want to cram as many people as possible into the pool. Ignoring the reality that in most public baths a max of 1 person is typically doing butterfly at a time. It just lets them say they’ve implemented regulations that make it safe. In reality it’s just covid theatre.

Covid Theatre. Precisamundo.