Yes but as I understood it the type inferer was taking your type annotation for granted, and apparently in fact it re-specifies it.<br>What you said was in reality:<br>> f :: forall. a -> a<br>
> f x = undefined :: forall a1. a1<br>And it turned it into:<br>f :: forall. a -> a<br>f x = undefined :: a<br>(here, 'a' would be <b>scoped</b>)<br><br>There's something I don't get...<br><br><div class="gmail_quote">
2012/1/4 Yucheng Zhang <span dir="ltr"><<a href="mailto:yczhang89@gmail.com">yczhang89@gmail.com</a>></span><br><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div class="im">On Wed, Jan 4, 2012 at 9:41 PM, Yves Parès <<a href="mailto:limestrael@gmail.com">limestrael@gmail.com</a>> wrote:<br>
> Would you try:<br>
><br>
> f :: a -> a<br>
> f x = undefined :: a<br>
><br>
> And tell me if it works? IMO it doesn't.<br>
<br>
</div>It works.<br>
<br>
As I understand, in this situation we are specializing the 'undefined<br>
:: forall a. a'<br>
to a more specific dependent type.<br>
</blockquote></div><br>