This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2018-12-25
Channels
- # adventofcode (18)
- # announcements (10)
- # beginners (59)
- # calva (9)
- # cider (1)
- # clojure (43)
- # clojure-austin (1)
- # clojure-europe (6)
- # clojure-italy (2)
- # clojure-nl (1)
- # clojure-uk (3)
- # clojurescript (8)
- # cursive (3)
- # emacs (3)
- # fulcro (3)
- # hoplon (1)
- # klipse (1)
- # off-topic (24)
- # reagent (2)
- # reitit (1)
- # shadow-cljs (41)
- # specter (5)
- # vim (1)
Fourteen years ago my aunt asked me to look at her windows computer. I told her I didn’t know anything about Windows, but she’s overbearing and persisted, and I sheepishly agreed. I got her locked out of her computer, and when she got upset i reminded her I said I didn’t know anything about it
@denisgrebennicov you might be interested in that ^
I think if java gets ported to wasm and i2p then runs in wasm, things like nightweb will have more reach
how is the current state of clojure on android? Found https://github.com/clojure-android but doesn't seem to get updated anymore (from a quick glance)
website linked (http://clojure-android.info) seems unrelated as well
@victorbjelkholm429 I haven't tried it in many years
@lilactown If you want strong privacy controls over a distributed system, some combination of nightweb referenced above and syndie (https://en.wikipedia.org/wiki/Syndie) would be ideal. Scuttlebutt mentioned above sounds like a similar idea.
One problem with anonymity is that it's not safe to let user's ship raw javascript between each other, so the pages you let user's expose to each other should ideal be defined by decidable structured text
So, instead of javascript, perhaps there should be a kind of distributed "plugin store" with ratings, so that new html elements can be defined in javascript and shipped from the store, which extends the html-like structured text specification.
Maybe that's what Tim Berners-Lee's new Solid thing is about https://solid.inrupt.com/
But, for instance, a website developer should be able to use some "data-picker" tag, without breaking the decidability of the markup, by having users "install" the widgets just-in-time from the trusted, distributed language extension plugin store
When you look at excel in abstract its a bunch of map reduce. So unlike opsystem development or audio processing or opengl - big data is indeed excel, just more rows.
Even the process is the same. You construct a formula, you feed it data, you look at results and cry, you go back to fixing the formula :)
:thinking_face: closing the 9th parentheses on one line ... should probably refactor 😄
I have a question: 1. Thing A is hard, but has to be done once in a while. 2. Let's build a DSL D in JSON to do A. That way we can later build a self-service GUI to do A. 3. D is accreting expressivity and is slowly approaching Turing completeness. I feel like D is a bad idea, but can't quite articulate why. So I guess my question is when do DSLs make sense? and especially when not to build DSLs?
If A is so hard, why do you think a gui for it will ever pay off
Since building the gui will probably be even harder
i have had a couple of projects where we figured that A was hard, like even 30 dev hours per month, but once we tried to guess how big the GUI build would be we figured that even 1000 dev hours wasn't going to guarantee a result. so we rather went with helper tools that cut the 30 down to 5-8 and took a fraction of the gui cost
“when do DSLs make sense?” is a damn good question, and I don’t how to answer it. Hope somebody here knows it.
This is a really good exploration of that subject: https://www.youtube.com/watch?v=i1KVwoE3n28
@U8QTB156K Thanks a lot for this! The talk has been an excellent starting point to my research about this.
No problem! I'm curious now: can you share any more details about what you're building?