Talk:Y combinator: Difference between revisions
Content added Content deleted
GordonBGood (talk | contribs) (→Temporary reversion of a recent Haskell edit: response after reversion...) |
|||
| Line 308: | Line 308: | ||
:: Understood – many thanks for the explanations – will you fix the non-compiling part and revert the rest in the way seems best to you ? [[User:Hout|Hout]] ([[User talk:Hout|talk]]) 22:31, 10 October 2018 (UTC) |
:: Understood – many thanks for the explanations – will you fix the non-compiling part and revert the rest in the way seems best to you ? [[User:Hout|Hout]] ([[User talk:Hout|talk]]) 22:31, 10 October 2018 (UTC) |
||
::: Good edits and explanations. Many thanks. [[User:Hout|Hout]] ([[User talk:Hout|talk]]) 11:53, 11 October 2018 (UTC) |
::: Good edits and explanations. Many thanks. [[User:Hout|Hout]] ([[User talk:Hout|talk]]) 11:53, 11 October 2018 (UTC) |
||
::::Done and dusted... |
|||
::::While I was making sure everything compiled, I did some performance tests, with some interesting results... |
|||
::::I didn't change the first and second parts much other than to make the code producing lazy lists functions so the lists can be consumed as evaluated as discussed above, and forced evaluations so they wouldn't be deferred until the end of the calculation, which need showed up in performance testing. |
|||
::::However, performance and capability as far as range of usefulness stunk, so I added back the working simple function recursive versions, which are at least ten times faster, don't take so much memory, don't crash for large ranges, and which compile easily without showing up compiler bugs. I'm a little harsh in evaluating the usefulness of the Y-combinator unless someone can show the error in my analysis... |
|||
::::As always, I monitor this Talk page and the main page if anyone would like to discuss it...--[[User:GordonBGood|GordonBGood]] ([[User talk:GordonBGood|talk]]) 12:05, 11 October 2018 (UTC) |
|||