https://kotlinlang.org logo
Channels
Powered by
#arrow
  • r

    raulraja

    09/21/2017, 10:00 AM
    Here is an easy one if anyone wants to get started contributing to Kategory: https://github.com/kategory/kategory/issues/300
  • r

    raulraja

    09/21/2017, 11:25 AM
  • s

    simon.vergauwen

    09/21/2017, 11:39 AM
    Doesn’t the whole monads don’t compose and this is why we need monad transformers and this is the concept deserve a separate doc?
  • s

    simon.vergauwen

    09/21/2017, 11:40 AM
    I had that feeling when I was reading through OptionT and even more so when also reading EitherT because both docs start the same.
  • r

    raulraja

    09/21/2017, 1:15 PM
    The patterns section on monad transformers and the monad docs will probably expand on that and refer to the transformers examples. The issue is that we can easily fall into theoretical concepts that can confuse people, so we have to be careful and balanced between pragmatism and turning the docs into explanations of concepts from CT.
  • r

    raulraja

    09/21/2017, 1:16 PM
    Currently there are links there to a blog where it explains it. Both the OptionT and EitherT are clones of each other since the concept is essentially the same.
  • r

    raulraja

    09/21/2017, 1:16 PM
    Also feel free to expand on whatever you think it may be missing from those docs.
  • s

    simon.vergauwen

    09/21/2017, 1:30 PM
    The issue is that we can easily fall into theoretical concepts that can confuse people
  • s

    simon.vergauwen

    09/21/2017, 1:32 PM
    That is true. I need to make some time to go through all of it.
  • r

    raulraja

    09/21/2017, 1:36 PM
    What we need in the docs though is a section for reference links to good sources that expand on each of the concepts
  • r

    raulraja

    09/21/2017, 1:38 PM
  • s

    simon.vergauwen

    09/21/2017, 1:39 PM
    I just read through the advanced scala with cats book by Underscore and I think that would be an amazing resource for Kategory.
  • s

    simon.vergauwen

    09/21/2017, 1:40 PM
    I was thinking of doing some of their case studies in Kategory to show case Kategory and why it is useful.
  • r

    raulraja

    09/21/2017, 1:43 PM
    We just need to make sure we adhere correctly to the commons license, I can talk to Dave and mention it, he is a very nice guy
  • s

    simon.vergauwen

    09/21/2017, 1:48 PM
    I only briefly checked out their license seemed like using their use-case example would be fine when giving them the appropriate credit of course.
  • r

    raulraja

    09/21/2017, 1:50 PM
    yeah, I contacted him privately also to make sure there are no concerns.
  • r

    raulraja

    09/21/2017, 1:51 PM
    Him and Noel Welsh will come to LambdaWorld too
  • s

    simon.vergauwen

    09/21/2017, 1:59 PM
    The Underscore books were my favorite books so far. They are very well written.
  • s

    simon.vergauwen

    09/21/2017, 1:59 PM
    Quite some people I am excited about to see in person @ LamdaWorld.
  • r

    raulraja

    09/21/2017, 2:00 PM
    yeah it's gonna be awesome
  • r

    raulraja

    09/21/2017, 2:01 PM
    In case someone is out of the loop http://www.lambda.world/
  • s

    simon.vergauwen

    09/21/2017, 5:10 PM
    What is considered better 
    inline fun <A, reified B> Fold<A, B>.fold(M: Monoid<B> = monoid(), a: A): B = foldMap(M, a, ::identity)
    or 
    inline fun <A, reified B> Fold<A, B>.fold(a: A, M: Monoid<B> = monoid()): B = foldMap(M, a, ::identity)
  • s

    simon.vergauwen

    09/21/2017, 5:11 PM
    Because I cannot call the first method 
    someFold.fold(list)
    I am forced to use named arguments.
  • r

    raulraja

    09/21/2017, 5:21 PM
    we don't have yet a good strategy for that, so whatever is more flexible and we gotta make sure is consistent across all modules
  • s

    simon.vergauwen

    09/21/2017, 5:31 PM
    The best is the “hardest” which is mixing. i.e. 
    find(FT: Foldable<F> = foldable(), noinline f: (A) -> Boolean): Option<A>
    is best with the default value as first param so you can use 
    list.find { it % 2 == 0 }
    but when you’re just supplying value like the above mentioned signature it would be most convenient to have the default value as last param.
  • r

    raulraja

    09/21/2017, 5:33 PM
    we really need to get the implicits stuff working because having 
    inline reified
    also prevents from those calls being called in non reified contexts unless they can be inferred in the function arguments
  • r

    raulraja

    09/21/2017, 5:35 PM
    For example right now you can't call 
    monad<F>()
    inside a class unless you tell it what 
    F
    is even though 
    F
    may be a type argument to the class.
  • r

    raulraja

    09/22/2017, 11:43 AM
  • s

    simon.vergauwen

    09/22/2017, 3:17 PM
    Is it just me or does IntelliJ freak out for everyone when working on Kategory once in a while?
  • p

    pakoito

    09/22/2017, 3:35 PM
    it's better now than what it was before
1...929394...416Latest