[Nix-dev] Stackage Support Will Be Discontinued

[Thomas Tuegel]

Hi Daniel,

I'm going to reply to your remarks out of order, because there is a
clarification I need to make before I go on.

On Fri, Jun 10, 2016 at 8:05 AM, Daniel Peebles <pumpkingod at gmail.com> wrote:
> I see a few possible outcomes here:
>
> Peter is swayed by Anthony and Thomas's arguments, and decides to keep
> (parts of?) stackage support around.

To be clearer, I don't really have a side in this discussion. As a
Cabal developer first, I typically "eat my own dog-food," i.e. I
primarily use Cabal/cabal-install instead of Nix. The only notable
exception is `emacs2nix', where I have chosen to use haskellPackages
as a convenience to my primary body of contributors (Nix users).

Originally, I had no intention of involving myself in this discussion.
After Peter remarked on IRC that he was surprised at the lack of
complaint, Anthony and I discussed (on Twitter) some problems with
dropping Stackage support [1]. I encouraged him to write to nix-dev,
so I felt responsible when his complaints were summarily dismissed.
That is why I entered the conversation, to correct what I perceive to
be a misrepresentation of Anthony's complaints.

> Peter is putting all this work in for free and has almost single-handedly
> been responsible for drawing a sizable chunk of the Haskell community
> (including me) to the Nix world. That is awesome. He's also obviously free
> to pour his free time into whichever direction he thinks is best for the
> project (or not put any time in at all, obviously).

Peter's monumental effort should be lauded! I am completely on-board
with any maintainer saying,

"We are going to do it this way. Although that causes problems for
you, doing it another way causes greater problems for me."

I have been the guy who says that. I have said it in nice ways and I
have said it in unpleasant ways, I'm sorry to say.

I am not on-board with saying

"The problem you have is not a problem because..."

If a user says they have a problem, they have a problem. We are
dropping support for a convenient Haskell workflow. We can ask our
users, as Peter has done, to change their workflow. The user still has
a problem: they must re-tool their workflow. The problem has been
shuffled around a bit, but not eliminated. This may be the optimal
balance of inconvenience for everyone (users and maintainers alike).
It may not be. That is the discussion worth having.

Regards,
Thomas

[1]. https://twitter.com/a_cowley/status/740649945808658432