Release cycle

[santiweight]

We already have a milestone for this, but I thought I'd find a spot to have text to discuss when we should release.

I just uploaded a couple of PRs that I think solve some of the current issues with the package:

• representation of Position (still up for debate?)
• test coverage - doesn't need to be amazing for first release
• readme
• Hand -> Hole

The remaining issue I see is 7-card evaluation. I think it would be fine to elide this addition for the first release? What do you think? We can just link to poker-fold as a WIP project that allows people to have something to look at. I believe in particular you said that you have a quite-accurate (though not tested) inefficient hand calculation method right?

[santiweight]

@tonyday567

[santiweight]

I guess I didn't explicitly say above - I think we are basically there for first release. Let's start getting others to help us out!!!

[tonyday567]

Let me wrangle the non-lookup card evaluator I have (it does both 5 and 7 card evaluations using [Card]) into a PR and let's see how it looks.

Taking a look at readme and properties ...

[santiweight]

Thank you for the great review and feedback @tonyday567

I think we should give ourselves a timeline for the release. I think it is not unreasonable to suggest for ourselves that we release on Thursday/Friday of this week (in 4/5 days). What do you think?

[tonyday567]

I pushed the module unhiding PR - I needed a baseline commit for poker-fold to catch up with where poker-base is at right now.

We're done on API changes for the release? Ok, if there's more to come, but just trying to firm things up and concentrate on polishing.

[santiweight]

@tonyday567 I feel good to go. I've got full test coverage on Poker.Cards and Poker.Game.

Technically Poker.Game has no tests yet, but only because the doc-tests are extensive and I'll add tests for Position once that design is actually finalised (I don't know whether or not it is yet).

Some notable wrinkles:

• Random + Arbitrary instances #33 Still missing Random instances
• Improve doctests and CI #27 Tests aren't enabled yet - I think we can activate everything now...
• Hand and ShapedHand #12 We still haven't introduced the specialised versions of Range. What do you think about adding this for first release? If you tell me what data structures to use I can implement this one.

[santiweight]

I've also gone through a lot of the docs - but it would be good to do one revision of docs prior to release - likely tomorrow...

[tonyday567]

ok! Will respond to each issue about in their threads, but I vote for freezing everything prior to release, except for docs, haddocks and bugs (which I don't know of any).

Will concentrate on getting CI right.

[tonyday567]

Are you planning a post on release?

[santiweight]

Okay - library frozen!

I don't know - I guess it would be good for us to include a bunch of examples from other libraries, say poker-game, which might be quite intuitive in terms of educational value?

Let me know what you think about a post. I'm not sure whether/what kind of introduction poker-base needs.

[santiweight]

One place that might be good for examples would be monetary quantities...

[tonyday567]

I'm happy for a low-key start, and was actually thinking a slow buildup is a lot better than starting with something splashy.

If we did go with a post, something like "Why haskell and poker coding are a good mix" might generate some coder interest. The monetary story is a good example of why, I could add in the mmap library thing (I used a library that hadnt been updated in 10 years and it worked perfectly etc), pattern synonyms giving us zero-cost abstraction and flexibility in future design.

[santiweight]

I like your comment - let's go with something like:

1. Release library
2. Next week write a little post about why Haskell is good for this stuff
3. Release another one of our libraries as a good exposition of what can be done

Something mildly related - it might be good to get an open source poker website up and running. poker-maison by therewillbecode is a good reference point, though I don't think he currently wants to merge his game logic with our ecosystem. Anyway, it might be a good way to get attention etc.

[santiweight]

Co-opting this thread to discuss the aftermath of release (such as issues as they arise and possible blog post etc.)