🚀 go-pugleaf

wunderwa@informatik.tu-muenchen.de (Jens Wunderwa) writes:

>I have edited both max-versions for better comparability:
>
>max_heiner( [Max], Max ) :- !.
>max_heiner( [MaxSoFar, First |Rest], Max ) :-
>	MaxSoFar > First, !, 
>	max_heiner( [MaxSoFar |Rest], Max ).
>max_heiner( [_|Rest], Max ) :-
>	max_heiner( Rest, Max ).
>
>
>max_hamer( [First |Rest], Max ) :-
>	max_hamer( First, Rest, Max ).
>
>max_hamer( Max, [], Max ) :- !.
>max_hamer( MaxSoFar, [First |Rest], Max ) :-
>	MaxSoFar > First, !, 
>	max_hamer( MaxSoFar, Rest, Max ).
>max_hamer( _, [First |Rest], Max ) :-
>	max_hamer( First, Rest, Max ).

>uses a temporary variable for the greatest element seen so far 
>whereas Heiner uses the first element of the input list for that 
>purpose. 
>
>I think it's bad style to misuse the input for constructing
>the result instead of applying the standard accumulator pair technique.
>
>Heiner's solution needs heap space for storing the intermediate 
>result making it 30 % slower.

Hamer's first post has expired here, but didn't it use one '->'
instead of two cuts? Using arrow is much better style than using cuts,
which are never justified for deterministic code like this. It is much
more readable, and can also be more efficient. I think that this is a more
important stylistic issue than the use of an accumulator.

It's also not good style to use the same functor name for max_hamer/2 and
max_hamer/3. I would much rather name the latter max_hamer_2/3, which gives
the compiler a better chance of catching errors if I accidentally use the
wrong number of arguments somewhere.

-- 
Fergus Henderson             fjh@munta.cs.mu.OZ.AU      
This .signature virus is a self-referential statement that is true - but 
you will only be able to consistently believe it if you copy it to your own
.signature file!