Difference between revisions of "Talk:Concurrency with oracles"
Jump to navigation
Jump to search
(Original remarks transferred) |
m (Spelling correction >_<) |
||
| (2 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
(unknown): Someone can add a reference here; I'm too lazy right now to do so. |
(unknown): Someone can add a reference here; I'm too lazy right now to do so. |
||
| + | |||
| + | [[User:Atravers|Atravers]] 05:02, 18 March 2021: (''references finally added'') |
||
(unknown): I think we could equivalently assume a given oracle not only stores the prediction about the computations and events it's going to be applied to, but also a couple of additional oracles. Thus the same kind of thing can function for the origial idea of an oracle and as the infinite binary tree of oracles that seems to come in handy. |
(unknown): I think we could equivalently assume a given oracle not only stores the prediction about the computations and events it's going to be applied to, but also a couple of additional oracles. Thus the same kind of thing can function for the origial idea of an oracle and as the infinite binary tree of oracles that seems to come in handy. |
||
| + | |||
| + | [[User:Atravers|Atravers]] 22:35, 18 March 2021: So instead of e.g. streams of them, combine the oracle concept with a tree type, to have the features of both: |
||
| + | * see Hancock's <tt>name_supply</tt> in in Peyton-Jones's ''The implementation of functional programming languages'' (section 9.6); |
||
| + | * Augustsson, Rittri, and Synek show how to make it practical in ''On generating unique names''; |
||
| + | * Launchbury and Peyton-Jones provide an implementation of it in ''State in Haskell'' (section 10.2) |
||
Latest revision as of 22:38, 18 March 2021
(unknown): Someone can add a reference here; I'm too lazy right now to do so.
Atravers 05:02, 18 March 2021: (references finally added)
(unknown): I think we could equivalently assume a given oracle not only stores the prediction about the computations and events it's going to be applied to, but also a couple of additional oracles. Thus the same kind of thing can function for the origial idea of an oracle and as the infinite binary tree of oracles that seems to come in handy.
Atravers 22:35, 18 March 2021: So instead of e.g. streams of them, combine the oracle concept with a tree type, to have the features of both:
- see Hancock's name_supply in in Peyton-Jones's The implementation of functional programming languages (section 9.6);
- Augustsson, Rittri, and Synek show how to make it practical in On generating unique names;
- Launchbury and Peyton-Jones provide an implementation of it in State in Haskell (section 10.2)