[spire] [algebra] [cats] Why typeclasses Eq / PartialOrder / Order are not declared as contravariant

[Tongfei Chen]

Why don't we do things like these?

trait Eq[-A]
trait PartialOrder[-A]
trait Order[-A]

I'd like to know the rationale behind the decision of making them invariant. Thanks!

[Ben Hutchison]

Im guessing its this issue: http://www.scala-lang.org/old/node/10664.html

--
You received this message because you are subscribed to the Google Groups "Typelevel Users & Development List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to typelevel+...@googlegroups.com.
To post to this group, send email to type...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/typelevel/3c695808-8e99-4ec4-a3d9-a2634b9060cb%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Miles Sabin]

FYI, I've just reopened SI-2509,

https://issues.scala-lang.org/browse/SI-2509?focusedCommentId=74526

I think we can deal with underlying issue properly.

Cheers,


Miles

> https://groups.google.com/d/msgid/typelevel/CAGJkGJ8WdHD3M2mTu7vUgkHC_xYq%3DxaF03Wq1iocXXWtZ85amg%40mail.gmail.com.

>
> For more options, visit https://groups.google.com/d/optout.

--
Miles Sabin
tel: +44 7813 944 528
skype: milessabin
gtalk: mi...@milessabin.com
http://milessabin.com/blog
http://twitter.com/milessabin

[Ben Hutchison]

Nice! The dotty commit [https://github.com/lampepfl/dotty/commit/89540268e] implies that Martin's no longer holds the objections raised in the 2012 thread [https://groups.google.com/forum/#!topic/scala-language/ZE83TvSWpT4], thats encouraging progress.

For me a fix would be all upside, but possibly there are existing codebases that, perhaps unwittingly, rely on the current behavior and break when its fixed. 

It seems like implementing the fix isn't a major obstacle, but will there be a willingness to evolve the language when there's a chance of collateral damage? 

-Ben

[Matthew Pocock]

Just saying that if they are invariant, you can always recover contravariance by deriving [B <:< A] Eq [A] => Eq [B]. I've sometimes found this manual control of narrowing and widening to be easier to control.

To view this discussion on the web visit https://groups.google.com/d/msgid/typelevel/CAGJkGJ_URYWKqsK4c2_Dyt%3Dp%3Dy01kWqTHDChjjmmW4PodYDi4w%40mail.gmail.com.

[Tongfei Chen]

Great!