On Sun, Nov 13, 2016 at 6:43 PM, Patrick Pelletier <code@funwithsoftware.org> wrote:
On 11/13/16 1:29 AM, Michael Snoyman wrote:
By the way, regarding your comments about resourcet and FP Complete: I'm honestly offended at this tone. I've clearly attempted quite often and quite sincerely to get such cooperation to happen, and have been rebuffed. Even if you're unaware of that, the implied accusation assumes a lot which isn't true. I honestly considered ignoring this thread entirely based on this tone, I'm tired of dealing with it.

I'm very sorry!  I did not mean to offend.  I take Simon PJ seriously in his call to have a productive, respectful discourse in the Haskell community.

To be clear, I did not object to resourcet linking to Stackage; what I objected to is that the link to stackage is in place of having any sort of description of resourcet, rather than in addition to it.  To me, that feels like it is making Hackage less useful (by denying it the package description) rather than making the most of Hackage, even though it isn't perfect.  This makes me a little grouchy, and perhaps this grouchiness seeped into my comments, for which I apologize.

Also, Hackage does have the ability to upload docs very conveniently, via "cabal upload --documentation", so this can be used to work around the fact that "Hackage documentation generation is not reliable."

We need to all work together in good faith to make the Haskell ecosystem better.  I am doing my best to do that, even if I am not perfect.  I have already volunteered privately to Gershom B to work on the Hackage documentation generation code.  I would like to make Hackage docs more reliable.  This may or may not involve integrating with the S3 docs.  That is a technical decision, and no slight should be implied, whatever the outcome.

Once again, I'm sorry if I have not succeeded in upholding Simon PJ's standards for the Haskell community.  I genuinely appreciate the contributions you and FP Complete have made to the Haskell ecosystem.

--Patrick


Thank you for that, offense _un_taken :). Just to let you (and others) know where I'm coming from on this:

I regularly get complaints of "why don't you do X," where X is a significant amount of extra work. Writing up dual descriptions in both README.md and cabal description fields is a prime example, and something I argued very hard for on Hackage. I'm disappointed with the way the description is displayed; I'd have much preferred that the README.md files I write would have been used. If you look on resourcet on Hackage, for example, there's a much more thorough description of the package at the bottom. I think this was a major mistake, but there's not much I can do about it.

So my general request: instead of saying "I dislike X" and imply that the author (me, in this case) should take on some unspecified work to change what they're doing, try assuming some good faith. I put a lot of effort into getting display of my packages to be better on Hackage, and eventually gave up. Each comment about how I didn't do enough work is offensive, and pretty much puts another nail in the coffin of me wanting to be involved in Hackage ever again.

Michael