Bethlehem Olivewood Burls - FreeStuff Winners (June 10, 2004)

[Aaron Koehl]

FreeStuff Winners from Bethlehem OliveWood : John Cioffi and Jim Ketron

Let us thank our own Diane Darwesh and Bethlehem OliveWood for their generous contribution!Per this thread (http://www.sawmillcreek.org/showthread.php?t=9529)--

Congratulations to Jim Ketron of Weber City, VA, and John Cioffi of Peabody, MA! They have won 15
beautiful olivewood pen burls to be shipped from Bethlehem, Israel, and have until
midnight, Thursday June 17th to post in this thread to claim their prizes.

Thanks again to Bethlehem Olivewood (http://www.bethlehemolivewood.com/Us.htm) for such a great contribution to our FREEStuff program!

PS. We expect pictures of the finished product, right guys (and gals)? :D

Regards,

[Chris Padilla]

Doh...I didn't win...AGAIN... :(

[Aaron Koehl]

There's always this coming Monday! (And another FREEStuff I'll announce shortly)

[Terry Stapleton]

Winning is always fun. The trouble is that I am not a turner, at least that's one addiction I haven't let myself start yet. So I've told Aaron to run the drawing again. That means all of you "losers" get another shot at it. Good luck!

[Chris Padilla]

Whoo-hoo!! You all right, Terry, you all right! :D :D

[Lloyd Frisbee]

Wow! Terry just went from "You bumb!" to "You Da Man!" Congrats....now draw my name!

Chris, you aren't a spinny guy either are you? I want some of that olive. I can turn it in my tiny workshop where the mice are hunchback.

Lloyd

[Chris Padilla]

Lloyd, no, I'm not. I'd just be so blindingly happy to win something for once in my life! :D I'd probably send it all to Ken...could give you some, too. :)

[Aaron Koehl]

Looks like I had to draw again--

Well, here goes nothing ... And the winner is...

<a href="http://www.sawmillcreek.org/member.php?u=1471">Tom Hintz</a> of Concord, NC.

[Aaron Koehl]

I had to draw again-- Congrats to Jim Ketron of Weber City, VA!

[David Wilson]

When is my turn?

[Chris Padilla]

Aaron, Aaron, Aaron...we need to have a chat about this algorithm of yours...it must be broken!

[Aaron Koehl]

Aaron, Aaron, Aaron...we need to have a chat about this algorithm of yours...it must be broken!Then, I submit my algorithm for approval and peer review: :D :D



select username from user where lastactivity > unix_timestamp(adddate(now(),interval -60 day)) order by rand() limit 1;
(Finds a random user who's visited within the past 60 days)

[Ed Falis]

How do you now rand() returns a different value each time if they're different invocations of the program?

[Jim Ketron]

Thanks Diane Darwesh, Bethlehem OliveWood, and Sawmill Creek I will put them to good use. I have several preachers and friends that would love a pen especially when they find out where the wood comes from. I don't own the items needed to make pens but I have wanted to get them and I have a good reason to get them now. any suggestions on what to get to start? I will get some blanks to practice on (not to mess up the good ones).
Hope to get some feedback so I can order tonight!
Thanks Aaron
Jim Ketron

[John M. Cioffi]

Aaron,
Thanks again for pulling my name.
Thanks to Dianne & Co. for the generous gifts given to & your involvement in this forum. I already responded to the other post, but wanted to say thanks, again.:D

[Aaron Koehl]

How do you now rand() returns a different value each time if they're different invocations of the program?Ed,

I know you're looking for a pseudo-technical explanation, otherwise you wouldn't have asked: :)

Empirical testing is applicable here--I haven't yet drawn the same name twice. While not a perfect uniform distribution of values over very large iterations, mySQL's random number generator works perfectly for our purposes--ad hoc random user selection with minimal constraints (e.g. the user must have been active within 60 days).

A typical programming failure is not seeding the implementation of the random number generator (usually achieved by seeding with the current timestamp, to decrease the probability of getting a repeated sequence). If I seeded the random number generator with '3' every time, I would always get the same number as a result.

In the case of a database server which stays resident (never terminates), like mySQL, the random number generator need not be seeded at all, because the queries are all going to the same invocation of the program, despite using different connections to the database. (The server, since it doesn't terminate, maintains its 'sequence' state between database connections).

If I were to stop and start the database server every time I ran the query, I would get the same names drawn each time, and in the same order.

Pitfalls:
Seeding the random number generator each and every time I ran the query [e.g. ORDER BY rand(NOW())] would, upon simple inspection, appear to be a feasible solution. Unfortunately, there is a problem if I draw names rapidly (within 1-2 seconds) in succession. The timestamp (now() or unix_timestamp()) doesn't change fast enough (only changes every second, right?), therefore if relying on the timestamp to seed the generator, we have to wait a second before drawing another name (which is fine for our purposes, since we only draw one name).

The solution, of course, is to seed your random number generator with the timestamp only once, when the database server first starts, (in our case a month ago), to ensure a random start point in the number sequence.