Re: Generic Package

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Mon, 30 Apr 2007 20:29:22 +0100, Simon Wright wrote:

> "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de> writes:
> 
>> On Mon, 30 Apr 2007 16:48:53 +0200, Georg Bauhaus wrote:
> 
>>> "The foreach procedure operates as if the following body
>>> were written
>>> 
>>>     procedure foreach(S: Set) is
>>>        C: Cursor;
>>>     begin
>>>        C := first(S);
>>
>> You forgot that first is not defined on Set, which is not
>> well-ordered and thus does not have visible implementations of
>> first, last, next, prev etc in its interface. It has Length, =, /=,
>>:=, and, or, xor, /, Empty, in, not in. [*]
> 
> First is not *publicly* defined on Set but this is *implementation*,
> for Pete's sake! It's likely to be hard to implement iteration/foreach
> efficiently without private knowledge

You missed the point. It was that if foreach were a primitive operation
defined on an unordered set, then its contract could not be stated. The
example I posed before Georg was to prove that counting members using
foreach with an atomic integer would indeed have returned the number of
members.

(that is - without looking into the implementation, because there it is
obviously ordered, iteratable, whatever)

> .. but you could, for example,
> copy the Set and then pick a random member, remove it from the copy,
> and process it, until the copy was empty.

I answered to this in a subthread. To be able to pick a random member <=>
to have an order.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de