Probably two very useful operations, however giving these a good "operator name" might be fairly difficult. For now, I'll suggest something like !> and <!, but bikeshedding very welcome!
LukaJCB
All 6 comments
Why not remove *> on Applicative, just use the >> from Monad and use *> for Parallel?
etorreborre
on 6 Nov 2017
@LukaJCB I think !> and <! are fine … I'm looking at Parallel for the first time and I think it should be constrained to CommutativeApply … need to review how we got here.
@etorreborre changing the constraints and semantics of *> and <* (which are heavily used) would be very disruptive so I don't think we should do that. >> is over-constrained and deprecated, and I think it's too soon to introduce it with new semantics, if this is what you were suggesting (I wasn't sure).
tpolecat
on 6 Nov 2017
@tpolecat I was thinking that many people were using >> for sequencing actions. I haven't seen much code using *> instead. The net changes will be to go from >> and *> to *> and !> and I wonder if assigning just another meaning to *> would be a better move in the long run (and keep similar names between Haskell and Scala).
etorreborre
on 6 Nov 2017
@tpolecat @LukaJCB feel free to ignore my bikeshedding opinion, I will be fine as long as we have an operator for Parallel.
etorreborre
on 6 Nov 2017
>> is deprecated in favor of *> … it's an anachronism from the days before Applicative in Haskell but it's still recognizable so I think we should not have it in cats. I use <* and *> all the time but I don't know how common it is I guess … I assume lots of people are using them since they're so useful.
In any case <* and *> are also well-known from Haskell and scalaz and I think we should not change the sequential semantics. a *> b is not the same as b <* a (see State for instance) although it is for CommutativeApply and would be for Parallel with <! and !>.
tpolecat
on 6 Nov 2017
I'm biased because I have used >> a lot more before I started using *>. And I agree with your last sentence.
etorreborre
on 6 Nov 2017
Related issues
Atry
·
5Comments
davidabrahams
·
3Comments
non
·
3Comments
durban
·
4Comments
SimY4
·
4Comments
Most helpful comment
>>is deprecated in favor of*>… it's an anachronism from the days beforeApplicativein Haskell but it's still recognizable so I think we should not have it in cats. I use<*and*>all the time but I don't know how common it is I guess … I assume lots of people are using them since they're so useful.In any case
<*and*>are also well-known from Haskell and scalaz and I think we should not change the sequential semantics.a *> bis not the same asb <* a(seeStatefor instance) although it is forCommutativeApplyand would be forParallelwith<!and!>.