Fork me on GitHub
#off-topic
<
2017-08-14
>
seancorfield02:08:50

@andy.fingerhut Slack payments are centralized to the owner(s) so ther's no option for team members to pay.

qqq05:08:36

I keep hearing "modules are more generic than haskell typeclasses" -- is there any good tutorial explaining this?

mpenet07:08:03

It's arguable, they are 1st class (in OCaml). Still some people wish to have something more akin to typeclasses too, modular implicits could be just that

mpenet07:08:53

But I am not sure it's planned to be done at all, just research for now I think

noisesmith15:08:35

the word “implicit” tends to scare me, not just in general, but also after some horror stories about using scala

mpenet15:08:26

I like OCaml in its current form(the compiler/runtime, it s another story), not sure it'd be a very valuable addition.

roberto16:08:43

I thought scala’s implicit was bad, until I started missing it in kotlin 🙂

roberto16:08:21

the problem with it in scala is the abuse and not just the concept itself. Odersky has given a talk about it and how to properly use it (and when not to).

hmaurer16:08:31

@roberto all functions should have 0 arity and pass all arguments as implicit, right?

noisesmith16:08:46

@roberto the horror story involved using a lib that overused implicits - the lib was the natural one to use for the task, but it imposed this implicit tuple type explosion that rippled through the code

noisesmith16:08:18

I haven’t used scala myself so I could be getting part of this story wrong - but it sounded like a bad corner he painted himself into

roberto16:08:50

yeah, I have been bitten by that. My own observation is that implicits should not be used in libs, but in the application at the outer layer when “wiring” up the interactions

noisesmith16:08:12

OK, that sounds reasonable

roberto16:08:18

it is a good alternative to dependency injection

roberto16:08:28

the problem is that people abuse them everywhere and don’t stop and think for a bit

akiroz17:08:46

sounds awfully similar to some of the operator abuse in haskell, seen some pretty exotic species of tropical fish while diving through haskell code before ><>˚

ghadi17:08:40

Anyone know where to get a name normalization database or have a recommendation? something that turns bill -> william and liz -> elizabeth

noisesmith17:08:38

@ghadi have you seen the downloads from oracle?

noisesmith17:08:54

haven’t used them in anger

noisesmith17:08:38

oh, wait - where’s the download for that…

akiroz17:08:48

wow, never knew bill is short for william.... I'll never understand how english names work lol

noisesmith17:08:41

@akiroz in english, “peggy” is short for “margaret”

noisesmith17:08:54

(so is “daisy”)

ghadi17:08:20

ok now that I know what to search for I found many of these name normalization dbs

adambard18:08:12

@noisesmith @akiroz we could also all be citing Dick Hickey all the time

adambard21:08:05

Apparently "Dick" came from "Hick" as a nickname, so it could also be Hick Hickey

mobileink21:08:23

hah. of course "Dicky Hickey" is also an option.

mobileink21:08:30

or "Richard Hichard"?

deg19:08:57

Is the Eclipse license suitable for a doc-only project? I am putting together a document with Clojure-related info. It is only text and tables; no code. It will be open to multiple contributors; I don't plan to commercialize it; and I don't want to hassle over legal issues. I like the general feel of the Eclipse license for this, especially since it is so accepted in the Clojure community, but don't know if it is suitable for documents. Thoughts?

Alex Miller (Clojure team)00:08:51

Yes, EPL can also be used for docs. All the Eclipse projects use it to cover both. The Clojure web site uses it too.

deg11:08:58

Thanks. I missed this reply from you last week. Meanwhile, I've released the doc (https://goo.gl/ZZH8fm) with no explicit license, and world-writable, in the hope that crowd-sourcing would add more value. So far, the experiment seems to have worked nicely; no visible vandalism.

sundarj22:08:13

http://bugs.java.com/bugdatabase/view_bug.do?bug_id=6804517 am i reading this right, that TCO is "planned" for JDK 13?

seancorfield23:08:04

I wouldn't be surprised if "JDK 13" was just a mythical version bucket that they use for all sorts of "unplanned" stuff...

seancorfield23:08:28

After all, the blog post referenced in that Bug is ten years old!

sundarj23:08:35

hah, that does sound plausible!