LOGO-L> Re: Eliminating items randomly from a list.

[George Mills <mills@softronix.com>]

Nice recap.

In essence if you think of your list as a "set"
(by definition the set can even have duplicates) then
Franks original solution or my later repaired solution
would work fine.

If you think of your list as a "list" (by definition
the list implies an order which also implies duplicates
are allowed) then my later repaired solution would do
the right thing.

Some folks incorrectly use the term set and list synonymously.

Frank Caggiano wrote:
> 
> I thought it would be a good idea to recap this.
> 
> The original post asked for a way to remove a random item from a list.
> 
> In my original response I made the double mistake of; one, assuming that
> the list would be unique or that the order of removal didn't matter
> (more on this latter) and two, not mentioning this in my reply.
> 
> George was good enough to point this out and he then supplied a routine
> that used the position in the list to pick the item to be removed rather
> than the item itself. His routine is (with the correction of a typo in
> the item selection statement) data independent.
> 
> Now what did I mean when I said that my routine would work if the exact
> order of removal doesn't matter. Let's say your simulating selecting a
> marble from a bag of different colored marbles. If the bag contains 3
> red, 3 blue and 3 green marbles (simulated by the list [r r r g g g b b
> b] or [ r g b r g b r g b]) the odds of selecting a green marble is
> 1/3.  If I then select a green marble it makes no difference which green
> marble I remove from the list.  In fact to talk about position in the
> bag of marbles is meaningless.
> 
> So where does this leave the original questioner. The data structure and
> routines used to manipulate that data depend a great deal on the
> intended usage.  However as George's routine will work in either
> situation it's probably best to use his routines unless there is some
> overriding reason not to.
> 
> Hope this helps and clears up any confusion this discussion may have
> caused. As George mentioned it's important when posting code, especially
> in response to a question, to be as explicit as possible as to the
> limitations and intended uses of the code and to test the code as much
> as possible, especially at the extremes.
> 
> regards,
> 
> --
> Frank Caggiano
> caggiano@atlantic.net
> http://www.atlantic.net/~caggiano

-- 
===============================================================
George Mills
email: mills@softronix.com
http://www.softronix.com
The www page contains some very powerful educational software.
Our single most important investment is our kids.
---------------------------------------------------------------
Please post messages to the Logo forum to logo-l@gsn.org.  Mail
questions about the list administration to logofdn@gsn.org.  To
unsubscribe send    unsubscribe logo-l    to majordomo@gsn.org.