These are chat archives for got-lambda/expression

9th
Nov 2018
jolod
@jolod
Nov 09 2018 12:51
@eriksvedang At least matlab's functions are pure! :-) (Unless you explicitly try to work around that.)
Erik Svedäng
@eriksvedang
Nov 09 2018 13:39
I guess... the code inside the functions is usually very imperative though :/
jolod
@jolod
Nov 09 2018 13:42
Yeah, and the syntax is pretty verbose... :-) But at least is fairly easy to reason about and test.
Not that there is any good way to write tests. But you know, YOLO!, and you have the REPL...
jolod
@jolod
Nov 09 2018 13:47
@Zalastax I looked at a presentation of the new type stuff for TS in 2018. It's looks pretty cool, and e.g. the exhaustiveness check for switch statements with singleton string types is pretty nice. Why do you think that TS works, but core.typed for Clojure didn't really catch on?
Pierre Krafft
@Zalastax
Nov 09 2018 13:58
@jolod I can think of several reasons. My assumption is that it's cultural rather than technical. I think there are many type system lovers who are "forced" to use JS and they welcomed TS with arms wide open
jolod
@jolod
Nov 09 2018 14:10
Yeah, maybe. IMO (without knowing too much about JS deving), the problems that JS faces when scaling up are similar to those that Clojure faces. In particular refactoring the data model. Python has also moved in the direction of TS. I don't know why the aversion to any kind of type support in Clojure is so strong, especially since everyone was using Schema for any non-trivial project, and then got really excited when spec was announced. Clearly there is a demand for it.
The obvious explanation is that Clojure is much smaller than JS and Python and thus does not have projects of the same weight.
Jean-Louis Giordano
@Jell
Nov 09 2018 15:22
I think a lot of it is being tired of hearing how irresponsible dynamic typing is and that it should be forbidden by law :p It doesn't really drive your interest to explore when it doesn't match your reality. That's kind of how I believe the flippant attitude of Rich Hickey escalated, by him being told again and again how he's an idiot that doesn't understand the first thing about types :p (I too have received this kind of feedback)
many JavaScript & Python developer have no issue saying their language is poorly designed, so for them adding types on top is like "at least we get some safety there". Clojure was "well designed" according to its users, so that's another hurdle to overcome.
jolod
@jolod
Nov 09 2018 15:28
@Jell In what context have you gotten that kind of feedback?
Jean-Louis Giordano
@Jell
Nov 09 2018 15:31
generally when I try to communicate some of the trade-offs I see in my own workflow
I'm also often wrong because I don't know all the state of the art of static typing
but then people use that as an excuse to dismiss what I say instead of trying to increase our shared understanding
jolod
@jolod
Nov 09 2018 15:36
Communicate with whom?
Jean-Louis Giordano
@Jell
Nov 09 2018 15:36
to be clear: I'm not pointing fingers at people here :p
jolod
@jolod
Nov 09 2018 15:37
Yeah, I assumed as much :-)
Jean-Louis Giordano
@Jell
Nov 09 2018 15:37
a few years back, I've been told at a Haskell meeting: "if you don't understand Type Classes, then I'm not interested in talking to you."
at ICFP, I've been dismissed by quite a few people by showing interest in racket & language oriented programming
jolod
@jolod
Nov 09 2018 15:38
Could that be the "dual effect" (to speak like a true FP person), i.e. Haskell folks are tired of hearing how much the type system is in the way?
Jean-Louis Giordano
@Jell
Nov 09 2018 15:39
that's for sure
jolod
@jolod
Nov 09 2018 15:40
The little I've read, the more vocal anti types Clojure people are often dismissive of Haskell and "deep monad stacks" without actually having experience of it, so that feeds into the perception of "you don't know what you're talking about".
It's unfortunate for you, since you know what you're talking about.
Jean-Louis Giordano
@Jell
Nov 09 2018 15:41
it's a strong polarizing effect: you have a ton of shitty static & dynamic languages with cult following that bash on static and dynamic typing, and the well design static & dynamic languages take that shit and put it on the other well designed opposed :p
jolod
@jolod
Nov 09 2018 15:41
I guess you should just choose your tribe and get on with the chanting. :-)
Jean-Louis Giordano
@Jell
Nov 09 2018 15:41
yeah it's not my style, I'd rather be hated by everyone :D
probably a lost cause though
or rather, no being hated by everyone is actually an easy cause :p
Jean-Louis Giordano
@Jell
Nov 09 2018 15:50
I just received a copy of the "Practical TLA+" book today! Might do some reading this week-end :)
Marco Zocca
@ocramz
Nov 09 2018 16:10
guys
I've come to realize something, in the past few months. There are parts of of PL community that definitely feel like "cabals", "sects", "tribes"
and the terms you just used are a further confirmation
however I did grow very tired and frustrated with this state of things. What started as a search for better tools became some sort of crusade
Marco Zocca
@ocramz
Nov 09 2018 16:15
and I don't even necessarily mean a crusade between languages, or static vs dynamic, even what happens between different levels of expertise within one language, or the dumbest of all, between build tools
my personal course of action will be to engage less and less with pure PL communities, and to focus more on applications
jolod
@jolod
Nov 09 2018 22:40
@ocramz How do you like the title of this episode? https://twitter.com/DefnPodcast/status/1040820826428334080
"Cult member echo chamber FTW"
jolod
@jolod
Nov 09 2018 22:48
@ocramz What do you mean by "pure PL communities" btw? Does Clojure count as a pure PL community?