Fork me on GitHub
#off-topic
<
2019-04-09
>
mg00:04:03

@john seems like a cool idea, although might have some of the same issues as pgp around how do you actually establish trust with people and pragmatically exchange keys

john03:04:31

Well, fb and Twitter often have personal friends. Especially fb. You could auto share keys locally or just trust the dht to rendezvous, like we do with block chain

john03:04:50

The strongest trust is initially sharing keys face to face

john03:04:15

@michael.gaare but large p2p networks can provide a pretty trust worthy rendezvous if you send some unique id out of band, like #blockchain eagle-screams

mg03:04:02

same is true with pgp keys, but it proved to be one of the big stumbling blocks to any significant adoption from what I understand

Sasho06:04:42

@john What if one doesn’t have a lot of friends and especially not a lot of computer savvy friends 😞 Your project will leave out the loners 😄 Joke aside, I think nowadays this is a substantial portion of the population. Also, isn’t the benefit of blockchain namely this - one doesn’t need to “trust” the other participants?

👍 4
ben09:04:42

IIRC, the Idris type-checker already allows you to check your programs for totality

ben09:04:15

I found working through “Type-driven development with Idris” very compelling and it really felt like the “future of programming” to me

👍 4
evocatus10:04:36

Any Emacs users? What packages do you use apart from (obviously) CIDER? (I would like to try Emacs to do Clojure development)

mpenet11:04:51

joker + flycheck is a nice addition

mpenet11:04:18

then the usual suspects: paredit / yas etc

evocatus11:04:14

I heard of some clojure-mode package

borkdude11:04:11

@gr.evocatus I have been using joker for a long time. Since a few weeks I’m using my minimal linter as well. https://github.com/borkdude/clj-kondo/ https://github.com/borkdude/flycheck-clj-kondo

jaihindhreddy11:04:33

One way to communicate with globally precise names is to use namespaced keywords and EDN (or Transit). Are there other data interchange formats that have the same thing?

borkdude11:04:59

@jaihindhreddy JSON supports “foo/bar” as field names as well

jaihindhreddy11:04:30

Something that distinguishes b/w names and strings?

borkdude11:04:40

Yeah, then transit is better

👍 4
jaihindhreddy11:04:55

In the key position, it's fine but when I want to send a bunch of names in an array somewhere

jaihindhreddy11:04:03

The distinction matters

borkdude11:04:28

I have the same problem with symbols as keys

jaihindhreddy11:04:06

Hmm... I never quite understood why two different reified name types are needed

jaihindhreddy11:04:18

Can't symbols do everything keywords do

borkdude11:04:21

symbols are mostly used for vars. keywords for data

👍 4
borkdude11:04:52

but yeah, good question. I’d like to know more about this too

jaihindhreddy11:04:57

Yeah, but as far as communication b/w programs is concerned, except for the inconvenience of quoting symbols everywhere

jaihindhreddy11:04:11

I can't see a fundamental advantage to having keywords separately

jaihindhreddy11:04:57

Thought symbols were interned strings too. They can have metadata too. Nice to know. Thanks!

henrik12:04:41

Transit does compress repeated use of keywords quite nicely. Maybe it does the same with symbols, I'm not sure.

👍 4
mpenet11:04:48

@borkdude is clj-kondo doing anything joker doesn't support yet?

mpenet11:04:12

I am trying to evaluate if a switch is worth it or not

borkdude11:04:33

yeah, that’s the point of clj-kondo. as a joker user I wanted some features that weren’t in joker yet, and I decided to try it myself.

borkdude11:04:42

so I’m using them both

borkdude11:04:43

it’s fairly alpha and the code is hacky, but it already provides useful feedback in my editor

borkdude11:04:13

like arity errors across namespace boundaries

borkdude11:04:41

and linting inside comments, where I usually experiment when working in a namespace

mpenet11:04:02

> like arity errors across namespace boundaries sounds cool

borkdude11:04:34

yeah. I already detected one in an open source project: https://github.com/Commonfare-net/just-auth/issues/29

borkdude11:04:24

about symbols/keywords: I’ve wondered if it matters for performance when using a giant map with symbol keys or keyword keys. probably not?

borkdude11:04:46

@mpenet here it’s documented how you can enable both joker and clj-kondo: https://github.com/borkdude/flycheck-clj-kondo

rboyd14:04:22

haha kondo 5 stars for the name

borkdude14:04:52

thanks 😄

rboyd14:04:54

btw do you know of a way to increase tip font size?

rboyd14:04:41

I guess yes. I was using flycheck-joker yesterday and it came up

borkdude14:04:35

no idea actually. I’m finding some references to flycheck-pos-tip. not sure how that all works

rboyd14:04:31

weird right? I saw https://github.com/flycheck/flycheck-pos-tip/issues/20 but I'm on osx and not sure it applies. I got into the pos-tip .el yest and didn't find any font size references either

rboyd14:04:42

oh well, I'll just squint 😀

borkdude14:04:30

try in #emacs

borkdude14:04:48

actually I asked myself now, since I’m curious

mpenet11:04:16

what kind of overhead does it add? how fast is it compared to joker atm?

mpenet11:04:38

joker adds quite a bit it seems but the benefits outweight the costs imho

borkdude11:04:09

what overhead does joker add for you?

borkdude11:04:01

btw, it’s very fast. linting a single file takes only 10-40 ms. you can see how much time linting a file took in the last line of the output

4
mihaelkonjevic14:04:29

Hi everyone, one of our mobile apps is a Webby finalist. It’s written in Clojurescript with Reagent and Keechma, and it would help us a lot if you would vote for us on this link - https://vote.webbyawards.com/PublicVoting#/2019/apps-mobile-and-voice/apps-mobile-sites-general/lifestyle . The app is named Loosid . The interesting part is that it was built by 1.5 frontend devs over 6 months, during which we survived multiple redesigns and pivots 🙂 . I’ll probably do a writeup about the general architecture, and how we used our stack to our advantage. If you want to find out more about Keechma you can ping me over at the #keechma channel

👏 12
Daniel Hines15:04:25

Re: Tau.alpha - that looks really cool! What are people doing to manage concurrency in NodeJs CLJS?

Chase16:04:49

2019 Stack Overflow survey findings are out. Kind of interesting. Still has Clojure developers as most paid (and experienced after F# which is also a functional language, right?). Also well loved of course. https://insights.stackoverflow.com/survey/2019

💰 4
Chase16:04:11

One trend I don't quite understand is that clojure ranks quite low on the "Most Wanted" category which doesn't seem to be true for the other "Most Loved" languages. I hope it doesn't indicate a lack of people wanting to try and use clojure. We gotta get them to try it before we can get them to love it!

henrik04:04:37

I'm just sceptical about StackOverflow getting a representative chunk of the Clojure community visiting them with regards to any variable or dimension.

💯 4
emccue05:04:32

It makes sense really. Take python for example: if someone is locked in a statically typed Java-like language and aren't flexible or a weakly dynamically typed language like PHP they can see the immediate benefit of switching for their work

emccue05:04:21

With clojure, I think people need to either use it or get slipped some Kool aid before they start to see where it would be a good fit

emccue05:04:42

Doesn't bother me all that much, but maybe it should

Alex Miller (Clojure team)16:04:57

FOP (fear of parens)

parens 20
omartell12:04:14

My reaction was more like FOMO after seeing Clojure 😂

manutter5117:04:37

I’d file that under They Don’t Know What They’re Missing

👍 4
hipster coder17:04:42

Is anyone experiencing headaches from using their tower desktops?

hipster coder17:04:22

I turned mine on. And I have a migraine, bad today. I am suspecting that it is off gassing from the plastics heating up.

jthibaudeau18:04:24

Looking for some opinions for a UI testing question, you have the option of going with one data-test-id with multiple values or multiple data-test-* keys each with single values which option do you choose?

jaide18:04:09

We need to get Rich one of these. https://media.giphy.com/media/3o6ZsW5lgayQl1uX6w/giphy.gif Just imagine the quality hammock time he’d get!

28
rich 20