These are chat archives for typelevel/general

6th
Mar 2016
Merlin Göttlinger
@mgttlinger
Mar 06 2016 12:31
Hi, I need to do a project of about 200 work hours next semester and would like to contribute to some cool scala project if possible. So is there any project that needs about that amount of work on a continuous topic, which I could possibly talk some professor at my university into accepting as a project I could do?
Greg Pfeil
@sellout
Mar 06 2016 13:14
Sooooooo exhausted. Also apparently left my laptop charger at Cira. Whoops.
@mgttlinger Anything in particular that interests you? I think there are probably many projects with opportunities like that.
Merlin Göttlinger
@mgttlinger
Mar 06 2016 13:22
@sellout I am open to anything with scala and happy if things like implicits and higher kinded types are involved. @antonkulaga has suggested a few ideas already around banana-rdf and biological informatics topics which I am currently reading/listening/watching about. My plan is to collect ideas and then write mails to the professors to see which of the topics I actually can do.
Adelbert Chang
@adelbertc
Mar 06 2016 15:19
@mgttlinger typelevel/cats has lots of implicits and HKTs :-)
if youre interested in functional programming/logic/category theory
Merlin Göttlinger
@mgttlinger
Mar 06 2016 15:21
@adelbertc I am interested, but have hardly any background knowledge in these topics
I don't know how much of a limiting factor that would be
Adelbert Chang
@adelbertc
Mar 06 2016 15:22
@mgttlinger shouldn't be. you dont actually need to know much about category theory, just consider it a library with lots of useful types
we've got some issues open at https://github.com/typelevel/cats/issues which vary in degrees of difficulty
you can hop onto the cats gitter channel if youre interested, and we can help you from there :-)
Merlin Göttlinger
@mgttlinger
Mar 06 2016 15:23
@adelbertc I'm following that channel regularly.
Adelbert Chang
@adelbertc
Mar 06 2016 15:24
awesome possum
@mgttlinger one thing is our code coverage is currently 88% and some N >= 90% just looks so much nicer :-) you can check our code coverage so far here https://codecov.io/github/typelevel/cats?branch=master and see whats missing
maybe add tests for https://codecov.io/github/typelevel/cats/core/src/main/scala/cats/MonadState.scala?ref=b638e298ca544739602e76388c9d266c4bc4357d - you can check to see that passing in identity doesnt do anything
let me know if you need any help
Merlin Göttlinger
@mgttlinger
Mar 06 2016 15:28
@adelbertc So you would recommend writing tests for cats to increase the code coverage as a possible project of the given extent?
Adelbert Chang
@adelbertc
Mar 06 2016 15:29
it probably wouldnt give you 200 hours
but i think it would be a start on figuring out how cats works
if youre interested inc ommitting to cats
i gave a talk on contributing to cats a couple days back, but the talk probably wont be out for a couple weeks
Merlin Göttlinger
@mgttlinger
Mar 06 2016 15:32
@adelbertc I'll see if I can convince a professor that letting me work on cats is a valid project
Rob Norris
@tpolecat
Mar 06 2016 15:50
Has anyone thought about a macro or plugin or something that constructs unapply versions of methods with higher-kinded method types?
Adelbert Chang
@adelbertc
Mar 06 2016 15:53
:-)
i feel like at worst you could do the whole SBT codegen thing
since at that point youre just emitting Strings, dirty as that is
Greg Pfeil
@sellout
Mar 06 2016 15:53
That worst sounds pretty bad ;)
Adelbert Chang
@adelbertc
Mar 06 2016 15:54
shhhhhh
Rob Norris
@tpolecat
Mar 06 2016 15:56
We also need a diagonalization of the shapes so we can just generate instances for all the "small" shapes once and for all.
We need to nerdsnipe someone into doing it.
"Sounds impossible to me, I'll bet even @non couldn't do it."
Adelbert Chang
@adelbertc
Mar 06 2016 15:57
"Frankly I don't think this is possible in Scala.. I bet you could do it in Haskell."
Rob Norris
@tpolecat
Mar 06 2016 15:57
nice
Greg Pfeil
@sellout
Mar 06 2016 15:58
@adelbertc You don’t have to do it in Haskell ;)
Adelbert Chang
@adelbertc
Mar 06 2016 15:58
@sellout shh youre ruining the nerdsnipe
Greg Pfeil
@sellout
Mar 06 2016 15:59
@adelbertc No, now Miles will point out where you would have to do something similar in Haskell, and then show how some Scala feature abuse works around it for us.
Adelbert Chang
@adelbertc
Mar 06 2016 16:00
haha :+1:
Michael Pilquist
@mpilquist
Mar 06 2016 16:37
@tpolecat I have vaguely thought about it for typeclassic... no idea how to implement though.
som-snytt
@som-snytt
Mar 06 2016 17:10
Thank-you for "nerdsnipe". Now I will always wonder if typelevel is nerdsniping me. For example, "at that point you're just emitting Strings." I still have a sneaking suspicion that typelevel folks would really rather be stringlevel, just comparing the names of things. In fact, I think the whole syntactic orientation of names first, optionally inferred type second, says that really we program in terms of names, with types as a backstop. Possibly we think nominally, investing names with meanings they may or may not possess, the proof of which is derived secondarily (via types).
Rob Norris
@tpolecat
Mar 06 2016 17:37
"@mpilquist I am confident that you can figure it out."
this is another kind of sniping but i don't know what to call it.
it's my only manager skill
som-snytt
@som-snytt
Mar 06 2016 19:49
My wife just attended Google women techmakers and loved it. One of her anecdotes was, if your manager assigns you a task, don't say "I can't do that" b/c that means, "Boss, you're an idiot, I can't do that."
I don't know if that rule applies to when I want to say, "why the hell would you want me to do that?"