[CoLoCo] [OT] warning - my kid is posting here

Neal McBurnett neal at bcn.boulder.co.us
Sat Feb 9 21:22:48 GMT 2008


I'm very glad you're turning this over to your daughter :-)

I'm still dubious about what a 6-year-old can get out of running
billions of simulations of pseudo-random number generators.  The
standard ones in Linux are really VERY GOOD these days, as you've seen
- more a demonstration of statistics than anything, since she is
unlikely to try her hand at writing a new PRNG.

I suggest turning the attention back to a real experiment with dice
and coins - way more interesting for most kids, I think, than nearly
perfect PRNGs, and of very current practical interest for our
democracy.  Why?

I've recently been working hard in Colorado for a good audit of
elections - what we do now is a sad farce:

 http://mcburnett.org/neal/elections/auditrule.html
 
Elections officials need to audit the real data, and need real,
verifiable random numbers to do the selection of the precincts to
audit.  And it has to make sense to non technical observers.

So I dug out this great article on public random numbers for the
clerk:

 Arel Cordero, David Wagner, and David Dill, The Role of Dice in
 Election Audits--- Extended Abstract, IAVOSS Workshop On Trustworthy
 Elections (WOTE 2006), (June 29, 2006)
 http://www.cs.berkeley.edu/~daw/papers/dice- wote06.pdf

and helped the clerk do a real audit of the Longmont election:
 http://mcburnett.org/neal/elections/longmont-audit/

Then I found the article that I posted here before.  Hopefully this
motivation will clarify why I was excited about it.

....

This article has some great insights on randomness and dice and coins
- spinning coins vs flipping them, biasing dice by rounding corners or
making the side lengths different, etc.

Great insights to design real experiments that everyone will "get".

 You can load a die but you can't bias a coin
  Andrew Gelman and Deborah Nolan April 26, 2002

 http://citeseer.ist.psu.edu/605305.html

http://citeseer.ist.psu.edu/rd/83657428%2C605305%2C1%2C0.25%2CDownload/http%3AqSqqSqwww.stat.columbia.eduqSq%7EgelmanqSqteachingqSqpublishedqSqdiceRev2.ps

AKA: http://tinyurl.com/2cmkyz

-Neal

On Sat, Feb 09, 2008 at 01:51:48PM -0700, Jim Hutchinson wrote:
> My daughter (Leina) is working on her science fair project (all the
> RNG traffic) and so far I've been asking all the questions for her.
> That is not a good approach in my opinion so I'm starting to turn this
> over to her. I think she needs to ask her questions herself. When I
> told her that she asked, "so, am I supposed to ask the Ubuntu people
> myself?" I told her yes. She will soon send out her questions to the
> list. Please keep in mind that she is 11 and is not a computer expert
> by any means but she does all her work on an Ubuntu computer and will
> be doing her project using only tools available in Ubuntu (I will
> encourage her to add an Ubuntu/FOSS credit to her project display).
> 
> I know it's redundant, but if those of you who posted scripts would be
> willing to post them again to her, that would help add to the
> authenticity of her project and allow her to share her communications
> as part of the project. Part of doing good research is connecting with
> a community.
> 
> On behalf of my daughter and myself, thanks for your help.



More information about the Ubuntu-us-co mailing list