Fix dot documentation by treeowl · Pull Request #32 · haskell/parallel

treeowl · 2018-06-04T13:47:58Z

I goofed up and made the strat1 `dot` strat1 = strat1 law too
strong. Fix that. Also, add laws for strat `dot` rseq and
strat `dot` rpar.

I goofed up and made the ``strat1 `dot` strat1 = strat1`` law too strong. Fix that. Also, add laws for ``rpar `dot` rseq`` and ``rseq `dot` rpar``.

simonmar · 2018-06-05T19:40:27Z

Control/Parallel/Strategies.hs

+--
+-- > strat `dot` r0 == strat
+-- > strat `dot` rseq == strat
+-- > strat `dot` rpar == strat


Is that true?

I believe so, yes. runEval m is a lifted value; it doesn't do anything until it's forced. Forcing runEval (r0 x) runs the usual return x IO computation, so that gives you x. Forcing runEval (rseq x) runs an IO computation that forces x to WHNF and returns the result (exactly the same as forcing x directly). Forcing runEval (rpar x) runs an IO computation that sparks a thread to evaluate x to WHNF, but immediately forces the result so the thread (almost certainly) fizzles.

r0 `dot` rseq == r0 r0 `dot` rpar == r0

Would be counterintuitive to my expectations, of the first description of dot, namely: "Compose two strategies sequentially". Or more explicitely:

"Composing the 'do nothing'-strategy with the 'evaluate to WHNF'-strategy sequentially, creates the 'do nothing'-strategy"

Am I the only, who thinks this sounds weird?

r0 `dot` rseq is not r0. But rseq `dot` r0 = rseq.

Last comment corrected.

dot is a composition operator in a semigroupoids sense: namely, it is associative. But it doesn't form a category, which makes intuition around it a little tricky. @Borgvall, by definition,

f `dot` g = \x -> f (runEval (g x))

Think about what runEval (r0 x), runEval (rpar x), and runEval (rseq x) do when forced. Ultimately the same thing (if you ignore sparks that fizzle). If you don't force them, they all do ... nothing.

I care about sparking, this is the main reason to use this library!

I'm still concerned about strat dot rpar == strat

We can construct contrived examples in which this doesn't apply, at least operationally (and operationally is after all the point of the library, so the equalities should be operational ones). e.g.

let strat n x = rseq (fib n) >> return x in strat 99999 `dot` rpar

If strat is very slow then the spark doesn't fizzle immediately and we get some useful parallelism. Of course it's hard to reliably define a strategy that's very slow (the example above could easily be optimised), but the fact that it's possible at all implies that we can't assume the spark fizzles immediately.

Control/Parallel/Strategies.hs

Borgvall · 2018-06-09T19:50:29Z

You are contradicting yourself. In the commit line 308 you say: "strat `dot` rseq == strat" for all strat, but if strat = r0 it is not. -- Diese Nachricht wurde von meinem Android Mobiltelefon mit GMX Mail gesendet.Am 09.06.18, 18:16, David Feuer <[email protected]> schrieb: @treeowl commented on this pull request. In Control/Parallel/Strategies.hs:

--

--- > strat2 `dot` strat1 == strat2 . withStrategy strat1 +-- 'r0', 'rseq', and 'rpar' are all one-sided identities of 'dot': +-- +-- > strat `dot` r0 == strat +-- > strat `dot` rseq == strat +-- > strat `dot` rpar == strat r0 `dot` rseq is not r0. But rseq `dot` r0 = r0. —You are receiving this because you commented.Reply to this email directly, view it on GitHub, or mute the thread. {"@context":"http://schema.org","@type":"EmailMessage","potentialAction":{"@type":"ViewAction","target":"https://github.com/haskell/parallel/pull/32#discussion_r194232298","url":"https://github.com/haskell/parallel/pull/32#discussion_r194232298","name":"View Pull Request"},"description":"View this Pull Request on GitHub","publisher":{"@type":"Organization","name":"GitHub","url":"https://github.com"}} {"api_version":"1.0","publisher":{"api_key":"05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":{"external_key":"github/haskell/parallel","title":"haskell/parallel","subtitle":"GitHub repository","main_image_url":"https://assets-cdn.github.com/images/email/message_cards/header.png","avatar_image_url":"https://assets-cdn.github.com/images/email/message_cards/avatar.png","action":{"name":"Open in GitHub","url":"https://github.com/haskell/parallel"}},"updates":{"snippets":[{"icon":"PERSON","message":"@treeowl commented on #32"}],"action":{"name":"View Pull Request","url":"#32 (comment)"}}} { "@type": "MessageCard", "@context": "http://schema.org/extensions", "hideOriginalBody": "false", "originator": "AF6C5A86-E920-430C-9C59-A73278B5EFEB", "title": "@treeowl commented on 32", "sections": [ { "text": "", "activityTitle": "**David Feuer**", "activityImage": "https://assets-cdn.github.com/images/email/message_cards/avatar.png", "activitySubtitle": "@treeowl", "facts": [ ] } ], "potentialAction": [ { "targets": [ { "os": "default", "uri": "#32 (comment)" } ], "@type": "OpenUri", "name": "View on GitHub" }, { "name": "Unsubscribe", "@type": "HttpPOST", "target": "https://api.github.com", "body": "{\n\"commandName\": \"MuteNotification\",\n\"threadId\": 342129584\n}" } ], "themeColor": "26292E" }

treeowl · 2018-06-09T20:03:20Z

I mistyped that comment. It's corrected on the web.

…

On Sat, Jun 9, 2018, 3:50 PM Johannes Röhl ***@***.***> wrote: You are contradicting yourself. In the commit line 308 you say: "strat `dot` rseq == strat" for all strat, but if strat = r0 it is not. -- Diese Nachricht wurde von meinem Android Mobiltelefon mit GMX Mail gesendet.Am 09.06.18, 18:16, David Feuer ***@***.***> schrieb: @treeowl commented on this pull request. In Control/Parallel/Strategies.hs: > -- --- > strat2 `dot` strat1 == strat2 . withStrategy strat1 +-- 'r0', 'rseq', and 'rpar' are all one-sided identities of 'dot': +-- +-- > strat `dot` r0 == strat +-- > strat `dot` rseq == strat +-- > strat `dot` rpar == strat r0 `dot` rseq is not r0. But rseq `dot` r0 = r0. —You are receiving this because you commented.Reply to this email directly, view it on GitHub, or mute the thread. ***@***.***":"http://schema.org ***@***.******@***.***":"ViewAction","target":" #32 (comment)","url":" #32 (comment)","name":"View Pull Request"},"description":"View this Pull Request on ***@***.***":"Organization","name":"GitHub","url":" https://github.com"}} {"api_version":"1.0","publisher":{"api_key":"05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":{"external_key":"github/haskell/parallel","title":"haskell/parallel","subtitle":"GitHub repository","main_image_url":" https://assets-cdn.github.com/images/email/message_cards/header.png ","avatar_image_url":" https://assets-cdn.github.com/images/email/message_cards/avatar.png","action":{"name":"Open in ***@***.*** commented on #32"}],"action":{"name":"View Pull Request","url":" #32 (comment)"}}} { ***@***.***": "MessageCard", ***@***.***": "http://schema.org/extensions", "hideOriginalBody": "false", "originator": "AF6C5A86-E920-430C-9C59-A73278B5EFEB", "title": ***@***.*** commented on 32", "sections": [ { "text": "", "activityTitle": "**David Feuer**", "activityImage": " https://assets-cdn.github.com/images/email/message_cards/avatar.png", "activitySubtitle": ***@***.***", "facts": [ ] } ], "potentialAction": [ { "targets": [ { "os": "default", "uri": "#32 (comment)" } ], ***@***.***": "OpenUri", "name": "View on GitHub" }, { "name": "Unsubscribe", ***@***.***": "HttpPOST", "target": "https://api.github.com", "body": "{\n\"commandName\": \"MuteNotification\",\n\"threadId\": 342129584\n}" } ], "themeColor": "26292E" } — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#32 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABzi_QAA8CwmZzijHI6DlZ91wjF05WO_ks5t7CcGgaJpZM4UZHuw> .

Borgvall · 2018-06-10T09:23:25Z

In the edited version of your comment you sai: "r0 dot rseq is not r0", but in line 308 you say:

-- strat `dot` rseq == strat [for all strat even r0]

So you are still contradicting yourself.

treeowl · 2018-06-10T11:57:43Z

Yes, I did make a mistake. r0 `dot` rseq is indeed r0 after all.

Kleidukos · 2024-01-19T15:52:01Z

@treeowl can I let you merge it?

treeowl · 2024-01-19T16:29:30Z

@treeowl can I let you merge it?

I don't understand the question. I don't have permission to merge things to master.

Borgvall · 2024-01-19T17:29:19Z

As this old merge request has been resurrected, I revisited it. My current point of view is, that the documentation is as weird as the implementation of dot, and the documentation changes of @treeowl simply expose this fact.

IMHO the dotcombinator should be deprecated. The reasonable way to "compose two strategies sequentially" is to use Control.Monad.>=>.

konsumlamm · 2024-01-20T01:50:51Z

I agree that r0 `dot` rseq == `r0 is counterintuitive. >=>/<=< indeed gives you a more intuitive (IMO) strategy composition, e.g. r0 <=< rseq == rseq.

We should definitely fix the documentation (although there's still an unresolved review), but I'm sympathetic to deprecating dot, unless someone has an argument why this strategy composition makes sense.

Fix dot documentation

9a57483

I goofed up and made the ``strat1 `dot` strat1 = strat1`` law too strong. Fix that. Also, add laws for ``rpar `dot` rseq`` and ``rseq `dot` rpar``.

treeowl force-pushed the fix-dot-doc branch from 5eba78e to 9a57483 Compare June 4, 2018 14:05

simonmar reviewed Jun 5, 2018

View reviewed changes

Simplify and extend

1e27bea

treeowl force-pushed the fix-dot-doc branch from 53365ce to 1e27bea Compare June 18, 2018 01:03

Kleidukos approved these changes Jan 19, 2024

View reviewed changes

konsumlamm mentioned this pull request Apr 23, 2025

Deprecate dot #75

Merged

konsumlamm closed this in #75 May 3, 2025

Conversation

treeowl commented Jun 4, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

simonmar Jun 5, 2018

Choose a reason for hiding this comment

Uh oh!

treeowl Jun 5, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Borgvall Jun 9, 2018

Choose a reason for hiding this comment

Uh oh!

treeowl Jun 9, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

treeowl Jun 9, 2018

Choose a reason for hiding this comment

Uh oh!

treeowl Jun 9, 2018

Choose a reason for hiding this comment

Uh oh!

Borgvall Jun 10, 2018

Choose a reason for hiding this comment

Uh oh!

simonmar Jun 11, 2018

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Borgvall commented Jun 9, 2018 via email

Uh oh!

treeowl commented Jun 9, 2018 via email

Uh oh!

Borgvall commented Jun 10, 2018

Uh oh!

treeowl commented Jun 10, 2018

Uh oh!

Kleidukos commented Jan 19, 2024

Uh oh!

treeowl commented Jan 19, 2024

Uh oh!

Borgvall commented Jan 19, 2024

Uh oh!

konsumlamm commented Jan 20, 2024

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

treeowl commented Jun 4, 2018 •

edited

Loading

treeowl Jun 5, 2018 •

edited

Loading

treeowl Jun 9, 2018 •

edited

Loading