On a more inane front,  does this give a path to either making $ less magical, or better user facing errors when folks use compose (.) style code instead and hit impredicativtity issues that $ magic would have handled ?<span></span><br><br>On Sunday, October 2, 2016, Ganesh Sittampalam <<a href="mailto:ganesh@earth.li">ganesh@earth.li</a>> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
  
    
  
  <div bgcolor="#FFFFFF" text="#000000">
    <div>Elsewhere in the thread, you said<br>
      <br>
      <blockquote type="cite">
        <p class="MsoNormal" style="margin-right:0cm;margin-bottom:12.0pt;margin-left:36.0pt"><span style="font-family:"Arial",sans-serif">1)
            ImpredicativeTypes enables types like `Maybe (forall a. a)`.
            Do those just disappear, or are they also enabled anyway? (I
            would guess the former.)<u></u><u></u></span></p>
        <span style="font-family:"Calibri",sans-serif">Yes,
          they’d disappear.</span></blockquote>
      <br>
      but here you're talking about 'xs :: [forall a . a->a]' being
      possible with VTA - is the idea that such types will be possible
      but only with both explicit signatures and VTA?<br>
      <br>
      On 30/09/2016 16:29, Simon Peyton Jones via ghc-devs wrote:<br>
    </div>
    <blockquote type="cite">
      
      
      
      <div>
        <p class="MsoNormal"><span style="font-family:"Calibri",sans-serif">Alejandro:
            excellent point. I mis-spoke before.  In my proposal we WILL
            allow types like (Tree (forall a. a->a)).<u></u><u></u></span></p>
        <p class="MsoNormal"><span style="font-family:"Calibri",sans-serif"><u></u> <u></u></span></p>
        <p class="MsoNormal"><span style="font-family:"Calibri",sans-serif">I’m
            trying to get round to writing a proposal (would someone
            else like to write it – it should be short), but the idea is
            this:<u></u><u></u></span></p>
        <p class="MsoNormal"><span style="font-family:"Calibri",sans-serif"><u></u> <u></u></span></p>
        <p class="MsoNormal" style="margin-left:18.0pt"><b><span style="font-family:"Calibri",sans-serif">When
              you have -XImpredicativeTypes<u></u><u></u></span></b></p>
        <p style="margin-left:54.0pt">
          <span style="font-family:Symbol"><span>·<span style="font:7.0pt "Times New Roman"">        
              </span></span></span><b><span style="font-family:"Calibri",sans-serif">You
              can write a polytype in a visible type argument; eg.  f
              @(forall a. a->a)<u></u><u></u></span></b></p>
        <p style="margin-left:54.0pt">
          <span style="font-family:Symbol"><span>·<span style="font:7.0pt "Times New Roman"">        
              </span></span></span><b><span style="font-family:"Calibri",sans-serif">You
              can write a polytype as an argument of a type in a
              signature  e.g.  f :: [forall a. a->a] -> Int<u></u><u></u></span></b></p>
        <p class="MsoNormal" style="margin-left:18.0pt"><b><span style="font-family:"Calibri",sans-serif"><u></u> <u></u></span></b></p>
        <p class="MsoNormal" style="margin-left:18.0pt"><b><span style="font-family:"Calibri",sans-serif">And
              that’s all.  A unification variable STILL CANNOT be
              unified with a polytype.  The only way you can call a
              polymorphic function at a polytype is to use Visible Type
              Application.<u></u><u></u></span></b></p>
        <p class="MsoNormal" style="margin-left:18.0pt"><b><span style="font-family:"Calibri",sans-serif"><u></u> <u></u></span></b></p>
        <p class="MsoNormal" style="margin-left:18.0pt"><b><span style="font-family:"Calibri",sans-serif">So
              using impredicative types might be tiresome.  E.g.<u></u><u></u></span></b></p>
        <p style="margin-left:54.0pt"><b><span>  type SID = forall a.
              a->a<u></u><u></u></span></b></p>
        <p style="margin-left:54.0pt"><b><span><u></u> <u></u></span></b></p>
        <p style="margin-left:54.0pt"><b><span>  xs :: [forall a.
              a->a]<u></u><u></u></span></b></p>
        <p style="margin-left:54.0pt"><b><span>  xs = (:) @SID id (
              (:) @SID id ([] @ SID))<u></u><u></u></span></b></p>
        <p class="MsoNormal" style="margin-left:18.0pt"><b><span style="font-family:"Calibri",sans-serif"><u></u> <u></u></span></b></p>
        <p class="MsoNormal" style="margin-left:18.0pt"><b><span style="font-family:"Calibri",sans-serif">In
              short, if you call a function at a polytype, you must use
              VTA.  Simple, easy, predictable; and doubtless annoying. 
              But possible</span></b><span style="font-family:"Calibri",sans-serif">.<u></u><u></u></span></p>
        <p class="MsoNormal"><span style="font-family:"Calibri",sans-serif"><u></u> <u></u></span></p>
        <p class="MsoNormal"><span style="font-family:"Calibri",sans-serif">Simon<u></u><u></u></span></p>
        <p class="MsoNormal"><span style="font-family:"Calibri",sans-serif"><u></u> <u></u></span></p>
        <div style="border:none;border-left:solid blue 1.5pt;padding:0cm 0cm 0cm 4.0pt">
          <div>
            <div style="border:none;border-top:solid #e1e1e1 1.0pt;padding:3.0pt 0cm 0cm 0cm">
              <p class="MsoNormal"><b><span style="font-size:11.0pt;font-family:"Calibri",sans-serif" lang="EN-US">From:</span></b><span style="font-size:11.0pt;font-family:"Calibri",sans-serif" lang="EN-US"> Alejandro Serrano Mena
                  [<a href="javascript:_e(%7B%7D,'cvml','trupill@gmail.com');" target="_blank">mailto:trupill@gmail.com</a>]
                  <br>
                  <b>Sent:</b> 26 September 2016 08:13<br>
                  <b>To:</b> Simon Peyton Jones
                  <a href="javascript:_e(%7B%7D,'cvml','simonpj@microsoft.com');" target="_blank"><simonpj@microsoft.com></a><br>
                  <b>Cc:</b> <a href="javascript:_e(%7B%7D,'cvml','ghc-users@haskell.org');" target="_blank">ghc-users@haskell.org</a>; <a href="javascript:_e(%7B%7D,'cvml','ghc-devs@haskell.org');" target="_blank">ghc-devs@haskell.org</a><br>
                  <b>Subject:</b> Re: Getting rid of
                  -XImpredicativeTypes<u></u><u></u></span></p>
            </div>
          </div>
          <p class="MsoNormal"><u></u> <u></u></p>
          <div>
            <div>
              <div>
                <div>
                  <p class="MsoNormal" style="margin-right:0cm;margin-bottom:12.0pt;margin-left:0cm">What
                    would be the story for the types of the arguments.
                    Would I be allowed to write the following?<u></u><u></u></p>
                </div>
                <p class="MsoNormal" style="margin-right:0cm;margin-bottom:12.0pt;margin-left:0cm">>
                  f (lst :: [forall a. a -> a]) = head @(forall a. a
                  -> a) lst 3<u></u><u></u></p>
              </div>
              <p class="MsoNormal" style="margin-right:0cm;margin-bottom:6.0pt;margin-left:0cm">Regards,<u></u><u></u></p>
            </div>
            <p class="MsoNormal" style="margin-right:0cm;margin-bottom:6.0pt;margin-left:0cm">Alejandro<u></u><u></u></p>
          </div>
          <div>
            <p class="MsoNormal" style="margin-right:0cm;margin-bottom:6.0pt;margin-left:0cm"><u></u> <u></u></p>
            <div>
              <p class="MsoNormal" style="margin-right:0cm;margin-bottom:6.0pt;margin-left:0cm">2016-09-25
                20:05 GMT+02:00 Simon Peyton Jones via ghc-devs <<a href="javascript:_e(%7B%7D,'cvml','ghc-devs@haskell.org');" target="_blank">ghc-devs@haskell.org</a>>:<u></u><u></u></p>
              <blockquote style="border:none;border-left:solid #cccccc 1.0pt;padding:0cm 0cm 0cm 6.0pt;margin-left:4.8pt;margin-right:0cm">
                <div>
                  <div>
                    <p class="MsoNormal">Friends<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p class="MsoNormal">GHC
                      has a flag -XImpredicativeTypes that makes a
                      half-hearted attempt to support impredicative
                      polymorphism.  But it is vestigial…. if it works,
                      it’s really a fluke.  We don’t really have a
                      systematic story here at all.<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p class="MsoNormal">I
                      propose, therefore, to remove it entirely.  That
                      is, if you use -XImpredicativeTypes, you’ll get a
                      warning that it does nothing (ie. complete no-op)
                      and you should remove it.<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p class="MsoNormal">Before
                      I pull the trigger, does anyone think they are
                      using it in a mission-critical way?<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p class="MsoNormal">Now
                      that we have Visible Type Application there is a
                      workaround: if you want to call a polymorphic
                      function at a polymorphic type, you can explicitly
                      apply it to that type.  For example:<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p>{-# LANGUAGE ImpredicativeTypes,
                      TypeApplications, RankNTypes #-}<u></u><u></u></p>
                    <p>module Vta where<u></u><u></u></p>
                    <p>  f x = id @(forall a. a->a) id @Int x<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p class="MsoNormal">You
                      can also leave out the @Int part of course.<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p class="MsoNormal">Currently
                      we have to use -XImpredicativeTypes to allow the
                      @(forall a. a->a).    Is that sensible?  Or
                      should we allow it regardless?   I rather think
                      the latter… if you have Visible Type Application
                      (i.e. -XTypeApplications) then applying to a
                      polytype is nothing special.   So I propose to
                      lift that restriction.<u></u><u></u></p>
                    <p class="MsoNormal"> <u></u><u></u></p>
                    <p class="MsoNormal">I
                      should go through the GHC Proposals Process for
                      this, but I’m on a plane, so I’m going to at least
                      start with an email.<u></u><u></u></p>
                    <p class="MsoNormal"><span style="color:#888888"> <u></u><u></u></span></p>
                    <p class="MsoNormal"><span style="color:#888888">Simon<u></u><u></u></span></p>
                  </div>
                </div>
                <p class="MsoNormal" style="margin-bottom:12.0pt"><br>
                  ______________________________<wbr>_________________<br>
                  ghc-devs mailing list<br>
                  <a href="javascript:_e(%7B%7D,'cvml','ghc-devs@haskell.org');" target="_blank">ghc-devs@haskell.org</a><br>
                  <a href="https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail.haskell.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fghc-devs&data=01%7C01%7Csimonpj%40microsoft.com%7Cd4eb1fd61d0148cea9f508d3e5dca6fe%7C72f988bf86f141af91ab2d7cd011db47%7C1&sdata=ZM3djztmpA09J6x1DmmV0LEeftsA1FhjPhjwLuG5w%2FE%3D&reserved=0" target="_blank">http://mail.haskell.org/cgi-<wbr>bin/mailman/listinfo/ghc-devs</a><u></u><u></u></p>
              </blockquote>
            </div>
            <p class="MsoNormal"><u></u> <u></u></p>
          </div>
        </div>
      </div>
      <br>
      <fieldset></fieldset>
      <br>
      <pre>______________________________<wbr>_________________
ghc-devs mailing list
<a href="javascript:_e(%7B%7D,'cvml','ghc-devs@haskell.org');" target="_blank">ghc-devs@haskell.org</a>
<a href="http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs" target="_blank">http://mail.haskell.org/cgi-<wbr>bin/mailman/listinfo/ghc-devs</a>
</pre>
    </blockquote>
    <p><br>
    </p>
  </div>

</blockquote>