Fork me on GitHub

@mario.cordova.862 I tend to do more code-related work outside office hours if, for whatever reason, I’m doing less of it at work (depends on the stage of the project etc). But I’ve pretty much always “programmed for fun” since I was a kid (in the 70's).


If you can’t face it, outside office hours, then it’s probably worth figuring out a way to carve out 2-4 hours of your week at work to do the sort of code-related work folks might otherwise do outside office hours.


After all, it should be part of your self-development to learn new stuff — new languages, new libraries, new techniques — and work should budget that into your “training” time and therefore should allow you at least a few concentrated hours a week doing “other stuff” that could benefit your job.


Where I work, we’re encouraged to contribute to open source libraries — on company time if our work projects use those libraries — and we’re generally encouraged to spend some time on “learning” and there aren’t many restrictions placed on that.


I’ll often spend an afternoon watching videos, reading blogs, poking at various technologies or open source Clojure libraries, as a “break” from regular work, once a week. But then we’re a small company and we’re also expected to be reachable in an emergency outside normal office hours, where reasonable.

Mario C.00:01:49

Company pays for pluralsight so at least there is that. I just have to make some time in the weekends. Not sure they would be too happy with working on other things during company time


Sounds like you should pitch for “training time” during office hours…


Does anyone have recommendations for error handling with Kafka streams apps please? It seems a lot of people publish failed messages to an error topic, but this strikes me as reinventing syslog and complecting streaming with logging. Am I missing something here?


@jayzawrotny if you’re hooked up on fun, there are only two things that will tear you away: something more fun, or something that makes it impossible to continue.


You could make an appointment, or a date, just after your period of work stops.


Or you could set your laptop to automatic shutdown at the end of work time. Sure you can boot it up again, but it will break you out of the trance.


JVM just became slow after some update in my linux machine(maybe a kernel update?!). Both JVM{8,11}-{enterprise,community}. GraalVM isn't affected. Anyone else noticed this? I'm on archlinux.


Also using Arch, haven't noticed any performance hiccups lately. Using both 8 and 11 too


(not enterprise though)


Yes I did notice this too!


i.e. on JDK 8 (didn't test graal)


I switched back to JDK 10 and it seems to be running fast there


I will try JVM10


Yeah. JVM10 still fast. JVM10 is no longer updated. Maybe it's some security update on JVM8 and 11. I will see the commits


I suspect about this 8199552 but I need to login to see this issue


I moved from 11.0.2 to 11.0.1 and it's faster again. I will dig more about it.

Eccentric J17:01:30

@lady3janepl That’s very helpful, thank you. I started leaning towards that conclusion myself and even began prototyping a tool that will run a terminal command, close it after 20 min, and bring up a git file selection and commit window. That way any running server processes will be shutdown. Unfortunately this is also another fun side project I’m going to be sprinting through in a trance-like state 😭


Are you on a mac?


doesn’t shut it down but puts an overlay on your screen

Eccentric J17:01:45

That looks pretty nice. I’ve been using and which also does an overlay effect. is the prototype. I’m thinking of making a bigger thing where people can have a .sideproject EDN file in their projects and maybe run sideproject start then it prompts for a goal, has a start command, a stop command, session length, break length, etc… then populate a git commit message based on the goal. The intention is to help people make their ideal workflow the path of least resistance.


I don’t know mate, anything that requires me to remember to type a command gets lost in noise

Eccentric J17:01:33

That’s true. I think it’s about the same cognitive load as remembering to run a timer app and start it before each project. The difference is it starts my project like lein run or tmux for a project and closes it to force me out of it when the timer stops.


well, whatever works for you 🙂


(For people who were following the org mode question yesterday: I found which looks like it’ll do about what I want: semi detailed time tracking without me having to remember to turn it on/off)

Nick Stares18:01:48

what low-level/systems language appeals most to the sensibilities of Clojurians?

Daniel Hines18:01:46

There seem to be a lot Clojurians also into Rust.


And Clojurians looking into Forth/Factor


how's the rust story for embedded/arm looking these days? ~2-3 years ago it looked like there was some good work going on there

Daniel Hines20:01:44

What is a signal graph in terms of programming? It comes up now and then, e.g. point 5 of the famous


My friend does some embedded development in rust


I've been told the biggest blocker is just finding libraries that work in nostd mode when you want to use a library


Other than that custom allocators are useful and there are stripped down, nostd, versions of rusts stdio

Daniel Hines13:01:50

Thanks for that gold mine! I’ll have to read into this more, but how is Dataflow programming different than FRP?


AFIACT, FRP is kind of a particular kind of dataflow programming


the “signal graph” concept is that you can imagine your program as functions that depend on particular “signal” values (reactive atoms, for instance). when a change is pushed to one of those signals, it propagates across your application; any functions that depend on it’s value will re-run, which may cause other functions to re-run, etc.


you may also have signals that themselves are compositions of other signals


thinking of it this way, you can draw a “signal graph” which represents these signals and how they depend on each other

Daniel Hines17:01:02

That's really sweet! Thanks for this! My googling had taken me down the path of information theory, but I didn't think packet loss/etc. was really a concern way up here at the application level, so I couldn't find the connection.


not sure if I’ve got everything right yet. I’m researching exactly dataflow/FRP and how it relates to reagent & re-frame for a post I’m writing, just got side tracked by some other research

Daniel Hines17:01:16

Where can I expect to find this post when it's done?


I’ll shoot you a link 😄