hub.darcs.net :: ross -> transformers -> issue

concrete functor and monad transformers

#24Add DeclareT transformer

swagger2 package has Data.Swagger.Declare 1 module where a DeclareT monad transformer is defined. Since it is a general structure, it might make sense to add it to the transformers package.

  • At first glance this looked like a special case of StateT, but now I see it's a new species. I guess both lazy and strict versions would be required. The latter might be a replacement for the useless strict WriterT. The name "Declare" doesn't seem general enough, though.

  • Yes, this is a new species.

    I named it "Declare" because I used it to declare definitions, but I agree that name is not general enough. Here are some alternative names: Patch/Patcher, Update/Updater. We can leave "look" method as is and "declare" to "patch" or "update".

  • I was thinking it's something like an accumulator. On further reflection, I think it will have the same non-constant space behaviour as WriterT, so maybe there's no point to a strict version. It also uses twice as many monoid operations as WriterT or a StateT-based accumulator.

    Still, a new species is interesting. To fit with WriterT and StateT, we'd want to swap the order of the pair, and of the arguments to (<>). It would also be useful to have monad transformations from ReaderT (pair the result with mempty) and WriterT (ignore the input), and to StateT (add the input into the output).

  • Maybe Accum and AccumT then? What about "look" and "declare"? What should we name those?

    Indeed this version uses twice as many monoid operations. I should reflect that in the documentation though.

    I see that I need to swap pair to preserve consistency with WriterT and StateT. However it seems that I have arguments to mappend correct.

    I agree that it can be useful to go from ReaderT/WriterT and to StateT. Thanks for the suggestion!

    I have forked this repo and made this new species available as AccumT. Please see Control.Monad.Trans.Accum. I am new to Darcs (and Darcs Hub) so I am not sure what is the usual workflow here.

  • I've merged your patches, though we may think of better names later. You're right about the mappend order matching WriterT.

  • WriterT has extra listen and pass functions. While listen is not interesting for AccumT (since we have look), we probably need to add an analogue for pass (under a different name I guess).

    Should we ask on the mailing list and/or reddit about better names?

    I also think at least one use case should be added to the documentation (like for StateT).

  • Also censor (from WriterT) and local (from ReaderT) might make sense for AccumT.

  • see also #92.

  • darcsden 1.2 beta built Jun 23 201810:48:19
  • report problems to simon