Date: Wed, 10 Feb 2021 18:27:57 +0100
I'm not much for bikeshedding, which this in deed has turned out to
be, but I do recognise that there might be some people who might take
offence with the name colony.
If we really need another name, the skipfield implementation made me
think of hopscotch (https://en.wikipedia.org/wiki/Hopscotch).
You start by drawing the field and see how fast you can skip through
it, then you at some point add a stone (delete an item) and when you
need more challenge you draw another one besides the first one.
If this is not completely correct, I apologise but the description
made me think of that.
My beef with this is that it is very implementation specific and
exposes quite a bit of detail about the underlying structure.
/Paul
Den ons. 10. feb. 2021 kl. 17.07 skrev Ville Voutilainen via SG14
<sg14_at_[hidden]>:
>
> On Wed, 10 Feb 2021 at 17:54, Jeff Garland
> <jeff_at_[hidden]> wrote:
> > True, and that name wouldn’t help the user decide if they want to use it like ‘unordered’ does - it actually obscures the function. But, I’m actually starting to think that in this case ‘bucket_list’ might really be better — I was half joking when proposing, but ‘buckets’ have their own analogy here from the real world and from implementations of hashed containers. They are often partially empty. You put stuff in and out of them. In hash buckets you skip over empty slots in the traversal. Buckets get allocated/deallocated as a unit. And it’s pithy ;0
>
> Well, it's at least trying to describe the high-level structure of the
> container. There's some snags, like items moving between
> hash buckets if you rehash, but not between bucket_list buckets. So I
> don't think this name describes the element stability
> much better than the current one, or hive.
>
> Toss a coin, Matthew.
> _______________________________________________
> SG14 mailing list
> SG14_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/sg14
be, but I do recognise that there might be some people who might take
offence with the name colony.
If we really need another name, the skipfield implementation made me
think of hopscotch (https://en.wikipedia.org/wiki/Hopscotch).
You start by drawing the field and see how fast you can skip through
it, then you at some point add a stone (delete an item) and when you
need more challenge you draw another one besides the first one.
If this is not completely correct, I apologise but the description
made me think of that.
My beef with this is that it is very implementation specific and
exposes quite a bit of detail about the underlying structure.
/Paul
Den ons. 10. feb. 2021 kl. 17.07 skrev Ville Voutilainen via SG14
<sg14_at_[hidden]>:
>
> On Wed, 10 Feb 2021 at 17:54, Jeff Garland
> <jeff_at_[hidden]> wrote:
> > True, and that name wouldn’t help the user decide if they want to use it like ‘unordered’ does - it actually obscures the function. But, I’m actually starting to think that in this case ‘bucket_list’ might really be better — I was half joking when proposing, but ‘buckets’ have their own analogy here from the real world and from implementations of hashed containers. They are often partially empty. You put stuff in and out of them. In hash buckets you skip over empty slots in the traversal. Buckets get allocated/deallocated as a unit. And it’s pithy ;0
>
> Well, it's at least trying to describe the high-level structure of the
> container. There's some snags, like items moving between
> hash buckets if you rehash, but not between bucket_list buckets. So I
> don't think this name describes the element stability
> much better than the current one, or hive.
>
> Toss a coin, Matthew.
> _______________________________________________
> SG14 mailing list
> SG14_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/sg14
-- • − − •/• −/• • −/• − • •/− • • •/•/− •/− • •/• •/− • • −/•/− •/• − − •− •/− − •/− −/• −/• •/• − • •/• − • − • −/− • − •/− − −/− −//
Received on 2021-02-10 11:28:12