Talk:Quickselect algorithm: Difference between revisions
Content added Content deleted
(→How is pivot chosen?: new section) |
(→How is pivot chosen?: Not stipulated?) |
||
| Line 2: | Line 2: | ||
Should the task specify how the pivot is to be chosen? Or should it leave it up to the implementer (and perhaps specify explicitly that the implementer can choose it how they want)? For example, the Python implementation chooses a random pivot, so it is randomized quickselect. There are other ones, e.g. choose the first element, or median of three, etc. --[[User:Spoon!|Spoon!]] ([[User talk:Spoon!|talk]]) 01:23, 1 October 2013 (UTC) |
Should the task specify how the pivot is to be chosen? Or should it leave it up to the implementer (and perhaps specify explicitly that the implementer can choose it how they want)? For example, the Python implementation chooses a random pivot, so it is randomized quickselect. There are other ones, e.g. choose the first element, or median of three, etc. --[[User:Spoon!|Spoon!]] ([[User talk:Spoon!|talk]]) 01:23, 1 October 2013 (UTC) |
||
:Yep I saw that too, but thought I would just leave it open. It would allow some one to write an additional (or extend the current) Python solution using other pivot choices. I guess if it means a lot to someone then this is the time to create a separate task if they think it is necessary then restrict this task to be for random pivots, but I didn't think it was needed (which sounds better than I am lazy, however I can't truly remember if laziness applied in this case :-)<br>--[[User:Paddy3118|Paddy3118]] ([[User talk:Paddy3118|talk]]) 22:06, 2 October 2013 (UTC) |
|||
Revision as of 22:06, 2 October 2013
How is pivot chosen?
Should the task specify how the pivot is to be chosen? Or should it leave it up to the implementer (and perhaps specify explicitly that the implementer can choose it how they want)? For example, the Python implementation chooses a random pivot, so it is randomized quickselect. There are other ones, e.g. choose the first element, or median of three, etc. --Spoon! (talk) 01:23, 1 October 2013 (UTC)
- Yep I saw that too, but thought I would just leave it open. It would allow some one to write an additional (or extend the current) Python solution using other pivot choices. I guess if it means a lot to someone then this is the time to create a separate task if they think it is necessary then restrict this task to be for random pivots, but I didn't think it was needed (which sounds better than I am lazy, however I can't truly remember if laziness applied in this case :-)
--Paddy3118 (talk) 22:06, 2 October 2013 (UTC)