Clojurians Log v2

This looks absolutely fantastic! Quite envious.

The actual ski conditions were pretty bad but yeah it’s gorgeous

Good reminder, I should go watch your talk finally!

It's been fun meeting you there.

Yes, and as far as I know, we’ll meet again in the not so distant future 🙂

yes!

Your opinions on what constitutes confusing code mirrors much of what I've seen at my workplace. It's never straightforward, though, each code base, maybe even each ns and/or each function has an individual balance to be struck in terms of making it readable, unsurprising as well as harmonious with its context. I think it really is an art that one can keep perfecting, so it's nice to see someone talking about this "humble" topic 🙂 We don't have time allotted for "innovation", sadly, but I can't help but to try and make nonsensical code make sense, if I get the chance. I find it quite enjoyable 🙂

> but I can’t help but to try and make nonsensical code make sense, if I get the chance. I find it quite enjoyable I think that’s what I find the most enjoyable to do. Also, I like to think about what made us write nonsensical code, is it sensical to the people who wrote it, and is there a deeper pattern to what is sensical and what is nonsensical.

Indeed. I find it very challenging to find a common ground in my very diverse set of colleagues. There are patterns that I enjoy and find delightfully terse (in the positive way 😄 ) that I think some my colleagues would find bothersome, and others would skate right through. There's also considerations like: In this particular business context or technical context, this particular way of doing things fits so neatly - but it's incongruent with the rest of our code base, for example by relying heavily on some opinionated library. Do we go for the perfect fit here, and eat the fact that people will have a bit of a learning curve the first time they go here? (And possibly a learning bump the second and third time, since they'll reading this rarely?)

We’re not using specter for the reasons you’re mentioning. We’re also very light on transducers because of this.

Hah, exactly some of the same topics I've thought about in regards to our code base. There is some usage of specter which I think works brilliantly in that context, but I'm pretty sure only 1 guy has a great handle on it - everybody else is just cargo culting (succesfully). Transducers is something I think about often, but I'm not sure introducing them would gain us as much as it would burden particular colleagues unnecessarily. We seem to get by with "normal" use of map, filter etc 🙂

I think if I were starting something from scratch, I’d base it around transducers. I think 🙂

It's interesting to think about. "What if our codebase, but in the style of x? What if our domain, but with heavy usage of library y?" I guess AI makes it easier to answer such questions, but the jury is still out on whether we're allowed to use it here.

I don't think we should shy away from tools like spectre because of unconventional syntax / readability if it can be used to great effect. Git is a good example of this: a terrible and confusing learning period, but something that i never would go back from.

We have a codebase at work that dates back, in places, to 2011. We were all learning Clojure then. So there's a big stylistic difference between our old code and our newer code -- but since that old code stills runs on the latest Clojure and still works, even when we have to update it, we tend to still do it in the old style... because the alternative would be to rewrite that old code into more modern style Clojure and that's just too much work for most small changes 🙂

@reefersleep I don’t think targeted use of transducers needs to be confusing even to those colleagues who are a bit scared of them. You can place inline comments on the side of the transformations explaining the steps to make it completely obvious that it’s basically the same order of operations as a threading macro. And everyone knows how to use into so creating the final collection using that function with the input coll and the transducer, can’t be all that confusing. It mostly maps to what they already know, I would think.

this discussion made me think of Reginald Braithwaite's old essay on transducers, written in terms of javascript (not important, except that it's written in a functional style). It explains really well the concept of transducers, building on (and explaining) reducers, using examples all the way. A good read and might be helpful for others struggling with the concept: https://raganwald.com/2017/04/30/transducers.html

Glad you enjoy it! It's a stand-out track from that album, imo. And since you liked that, you might like a stand-out track from a different album; Two Dots by Lusine. https://youtu.be/4iIvRXCV9lk?si=2RQ788JBezG8ewBJ I also really like the Pezzner remix, I go back and forth on that and the original being the best version. https://youtu.be/ot-G-GUAegc?si=wTWu1OoHYEjGW1lZ

Whether or not our tastes overlap, I’m always happy to see someone share what they love! 🫶

That reminds me it's been a while since I listened to MBR. Thanks!

What's MBR? 😮

https://www.youtube.com/watch?v=d5kKqa7hiak 🙂

Good stuff! And that reminds me that I managed to omit Synthwave and chiptuney stuff entirely from the list. Gotta make another sometime

This Scuba track is hot stuff, thanks for posting it