On Wed, 11 Jul 2007, Derek Holzer wrote:
Dear Libero, [...] where $0 will be replaced by a new random number which is unique to each abstraction).
Dear Derek, please see http://web.archive.org/web/20011027002011/http://dilbert.com/comics/dilbert/...
$0 numbers are just as predictable as the random numbers in that cartoon. they form the sequence: 1000, 1001, 1002, 1003, 1004, 1005, 1006, 1007, 1008, 1009, 1010, 1011, ...
In pd, it's not really important how those numbers are picked as long as there are no duplicates. In practice, it's easiest to count up and not bother reusing numbers when abstractions are destroyed. This is what pd does. This strategy is not without shortcomings, but few people ever noticed it.
_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - tél:+1.514.383.3801, Montréal QC Canada
Hi Mathieu,
yes, of course I meant a non-repeating number. I was concentrating on the end-results rather then the details of the process, but it's a useful distinction when trying to document it. Thanks.
d.
Mathieu Bouchard wrote:
On Wed, 11 Jul 2007, Derek Holzer wrote:
Dear Libero, [...] where $0 will be replaced by a new random number which is unique to each abstraction).
Dear Derek, please see http://web.archive.org/web/20011027002011/http://dilbert.com/comics/dilbert/...
$0 numbers are just as predictable as the random numbers in that cartoon. they form the sequence: 1000, 1001, 1002, 1003, 1004, 1005, 1006, 1007, 1008, 1009, 1010, 1011, ...
In pd, it's not really important how those numbers are picked as long as there are no duplicates. In practice, it's easiest to count up and not bother reusing numbers when abstractions are destroyed. This is what pd does. This strategy is not without shortcomings, but few people ever noticed it.
_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - t�l:+1.514.383.3801, Montr�al QC Canada
On Wed, 11 Jul 2007, Derek Holzer wrote:
yes, of course I meant a non-repeating number. I was concentrating on the end-results rather then the details of the process, but it's a useful distinction when trying to document it. Thanks.
The end-result that you should stress is that there are going to be no duplicates, because they would conflict. All of the rest is secondary, especially randomness.
Half the point of referencing the Dilbert strip is the repetition. Independently-random numbers can be duplicates.
example 1. Picking two numbers from 0 to 99 you have a 1% chance that they are the same.
example 2. If $0 uses uniformly-random numbers from 1000 to 9999 and you instantiate at least 110 abstractions you have more chances of getting duplicates than not. (This is usually known under the name "Birthday paradox")
example 3. the original Birthday paradox. the probability of N people of having shared birthdays: http://en.wikipedia.org/wiki/Image:050329-birthday2.png
_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - tél:+1.514.383.3801, Montréal QC Canada
-
Derek Holzer -
Mathieu Bouchard