cool

[nkuipers]

Yes you are right, this case is just as effectively solved non-recursively and 
the reminder not to get carried away is well received.  If I were to chide 
your solution however, I might say that it is very difficult to read to anyone 
not familiar with a breadth of perl idioms, everything from shift and qw() 
which I also used, to preceding parens with + for making something not look 
like a function call (if I remember correctly), => comma equivalence, and some 
rather heavy operator chaining.

;)

nathanael


>===== Original Message From Peter Scott <Peter at PSDT.com> =====
>At 12:58 PM 2/3/03 -0800, nkuipers wrote:
>>I played around in recursion a bit, wanting to get the practice but
>>also to do
>>something useful.  So I framed it to myself as an exercise.  This was
>>my first
>>idea:
>>[snip]
>>sub makeDNA {
>>         my $n = shift;
>>         my @alphabet = qw( A C G T);
>>         if ($lengthRemaining == 0)
>>         {
>>                 return;
>>         }
>>         print $alphabet[int(rand(4))];
>>         makeDNA($n - 1);
>>}
>>
>>makeDNA(10);
>>print "\n";
>>
>>This could develop into a small host of other recursive, random
>>functionalities useful for bioinformatics, such as taking a sequence and
>>randomly mutating n residues.  I will be incorporating above subroutine and
>>possibly others in my BIO::Basic module.
>
>Well, hold on a second... fun as recursion may be, it is not a
>panacea.  It is only optimal in the cases where you really do find
>yourself wanting to call the same function from within itself.  I
>suggest familiarizing yourself with a number of classic uses of
>recursion before deciding to use it.  Look for: The Tower of Hanoi,
>Ackerman's function, and anything to do with parsing a stream which may
>be lexically recursive in some fashion.
>
>I'd write yon function above as
>
>sub makeDNA
>{
>   join '' => map +(qw(A C G T))[rand 4] => 1 .. shift;
>}
>
>In general, don't print from inside utility functions.  Return strings
>instead so you can either print them or decide to do something else
>with them (e.g., run them through tests).
>
>--
>Peter Scott
>Pacific Systems Design Technologies
>http://www.perldebugged.com/