[ghc-steering-committee] Recommendation for #378: support the design for dependent types

[Spiwack, Arnaud]

On Tue, May 18, 2021 at 8:37 PM Richard Eisenberg <rae at richarde.dev> wrote:

>
> > There remain a few individuals who appear to remain deeply unconvinced.
> However, these seem to be a small minority. The reasons they are not
> convinced appear to be around lack of understanding of the proposal/design
> and general worry about unintended consequences. I have tried to address
> both of these, but I do not believe my efforts have been fully successful.
>
>
> I stand by this claim -- though I can see how the language here might be
> off-putting. This section is meant to be a summary of the discussion, and
> this really was a part of the discussion. However, I have replaced this
> bullet with a narrower one, describing the particular worry about people
> being forced to use dependent types because of early library adoption.
>

You are claiming that they don't agree with the proposal because they don't
understand. It's a very strong claim to make. But you've changed the
sentence already, so I guess no point in arguing further.


>
>
> > Many industrial Haskellers came out of the woodwork to support this
> proposal.
>
> is rather misleading, since there seem to have been just as many
> industrial Haskellers which came out opposed to the proposal.
>
>
> I disagree here. The GitHub ticket has 412 positive emotions, and 3
> negative ones. The 412 might be an overcount, because I believe one
> individual can put multiple emoji on the same ticket, but there are at
> least 274 individuals in favor. It's true that I don't know how many of
> these are industrial users, but there were quite a few people who
> identified as industrial users in support. I actually don't know that any
> of the detractors identified themselves as industrial users, though I
> didn't read through all the comment trail to check this.
>
> We don't -- and shouldn't -- operate solely (or primarily) on the basis of
> popularity, but I think both the emoji expressed and the comment trail
> suggests that this feature is desired in industry more widely than it is
> opposed.
>

There is no doubt that the idea of Dependent Haskell is popular (Dependent
Haskell in general, and this proposal in particular), but I'm not sure that
this is what your sentence is claiming (I wouldn't object to this claim!).
If it's not misleading, this claim is at least very hard to verify. It does
make me uncomfortable (as for industrial contradictors I remember Matt
Parsons being one of them).

PS: I didn't mean to imply that there being few claims/design choices in
the proposal was a bad thing. I suppose, though, that like many such
high-level summaries of a thing one comments, it was meant mostly for my
own consumption, to ground my reasoning.

/Arnaud
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-steering-committee/attachments/20210519/38271e38/attachment.html>