Comparison to lattices

[phadej]

I see partial-order has some additional combinators, not present in http://hackage.haskell.org/package/lattices, but is there any other reasons to prefer one over another?

[mtesseract]

Hi,

I'm not familiar with the lattices package, but it seems to me that they are rather different. partial-order really focuses on that: defining a type class for partial orders & providing instances for several types. Since partial orders induce a notion of equality, my package also provides some convenience functions like elem or nub.

It doesn't introduce the notion of lattices.

Also, it provides instances and functions for e.g. testing for the subset-relation of lists out of the box:

Prelude> import qualified Data.PartialOrd as PO
Prelude PO> [3,1] PO.<= [1..10]
True
Prelude PO> 

Does this answer your question?

[phadej]

Partially,

lattices has Algebra.PartialOrd too. And there is instances for subset partial order for Set.

There isn't a partial order for lists, but if added it probably be subrange relation (i.e. isInfixOf)

[mtesseract]

Ok, as I understand it we can summarize: The packages have some overlap, but don't provide equivalent functionality. What do you think, shall we close the issue?

[andreasabel]

lattices has Algebra.PartialOrd too. And there is instances for subset partial order for Set.

Some comparison between the interfaces in lattices and partial-order is here:

• Interface provided by PartialOrd isn't practical/efficient haskellari/lattices#117

(Unfortunately, neither interface fits my needs: connectedness of nodes in a DAG is a partial order, but it is relative to a context (the DAG), a nodes in a particular DAG do not make a good Haskell type.)