Marko Schulz <in6x059 / public.uni-hamburg.de> wrote:
> On Thu, Apr 25, 2002 at 04:37:23AM +0900, Stefan Schmiedl wrote:
>> 
>> Well, then we have two possibilities:
>> a) there is no String#prev
>> b) there is a STring#prev with a preference.

> I strongly favor a): 

But note likewise that none of this applies to Integer...

> - String#succ is a convenience method. It is nice to have, but one
>  should not conclude the necessity to have the reverse.

Whereas Integer#succ and Integer#prev are far more fundamental, and it would
be nice to have fast implementations for when they are needed.

> - String#pred is not really necessary. Even a lot less than
>  String#succ.

Integer#prev has a more or less equal footing with Integer#succ 

> - String#pred would require a lot of questionable distinction of
>  cases. A lot more than String#succ. (The predecessor of "foo"
>  might be "bar", "baz" or "zoo". What do we go for? Hacker 1: "bar"
>  of course. Hacker 2: No "baz" is more reasonable. Hacker 3: "zoo" is
>  what most people will expect.)

Integer#prev is well defined :)

Basically, the line should be drawn at "the analogy between String and
Integer breaks down here" rather than "neither of them need a prev method".

(And while we're at it, IMO .pred(ecessor) is a more natural
complement to .succ than .prev is)

-- 
Martin DeMello