std.array suggestion

[Stewart Gordon]

Oskar Linde wrote:
> Stewart Gordon wrote:
>> Oskar Linde wrote:
>> <snip>
>>> T[] doSort(T[], delegate|function wrongOrder(T,T));
>> <snip>
>>
>> You've left off the return type of wrongOrder.  Moreover, why the name 
>> wrongOrder?  To simply return whether or not these two elements are 
>> out of order?  Why not use what people are used to, i.e. a function 
>> that returns -ve, 0 or +ve for <, = or > respectively?
> 
> The return value would be anything usable as a condition in a 
> conditional function if(wrongOrder(a,b)) {...}
> 
> The reason for not having -ve,0,+ve is that you only need a binary 
> predicate for sorting.

But depending on the sorting algorithm, I can imagine that either may be 
more efficient.

Quite a lot of things at the moment rely on a three-valued ordering.

C: strcmp, memcmp, qsort
Java: compareTo (sic), Comparator
D: opCmp, std.string.cmp

Even if you do have it documented that doSort is unusual, somebody might 
either forget this or be already thinking on the three-valued terms when 
they see code that uses it.

> It is also in many cases simpler to define a 
> ordering predicate than a 3-valued ordering. The name wrongOrder just 
> helps as a reminder of what the predicate should return.
<snip>

If that's so, I wonder why so many things have stuck with the 
three-valued ordering.

Stewart.

-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GCS/M d- s:- C++@ a->--- UB@ P+ L E@ W++@ N+++ o K-@ w++@ O? M V? PS- 
PE- Y? PGP- t- 5? X? R b DI? D G e++>++++ h-- r-- !y
------END GEEK CODE BLOCK------

My e-mail is valid but not my primary mailbox.  Please keep replies on 
the 'group where everyone may benefit.