[Beowulf] OT: public random numbers?

Lux, Jim (337C) james.p.lux at jpl.nasa.gov
Thu Aug 11 17:55:30 PDT 2011

Low order digits from weather stations are not likely to be random.
They're almost certainly converted from some quantized converter, and may actually have a double conversion (Celsius Fahrenheit)

NWS and NOAA are actually part of the same organization, aren't they.  (since the NWS web page at weather.gov is titled "NOAA's National Weather Service")

Jim Lux
From: beowulf-bounces at beowulf.org [mailto:beowulf-bounces at beowulf.org] On Behalf Of Peter St. John
Sent: Thursday, August 11, 2011 5:44 PM
To: David Mathog
Cc: beowulf at beowulf.org
Subject: Re: [Beowulf] OT: public random numbers?

I was thinking the National Weather Service, instead of NOAA; it's a vital public service that such information is recorded and diseminated for airfields and the like, e.g.:
So I would write a script to scrape least significant digits from that, for agreed times, dates, and locations. Whoever writes the script and wherever it is run, anyone can check its results manually.
However, that item has a disclaimer that the data is subject to review :) So it may matter how far back in time you need to be able to go, and how long into the future you need the data to be available at the same place. But nobody promises their website will stay unchanged indefinitely, they can't. But at any given time, a group can agree on (say) the lowest significant digits of the temperatures at time T in cities X, Y, and Z as reported at time T2 by the NWS.

On Thu, Aug 11, 2011 at 7:56 PM, David Mathog <mathog at caltech.edu<mailto:mathog at caltech.edu>> wrote:
Since this is very OT, I'll try to keep it short.

Here is the problem - imagine a group of people who neither know nor
trust each other, yet must agree on the fairness of a single random
number.  Basically they are going to have a lottery.  They aren't
organized enough to generate such a number themselves - it must be found
from some process already active on the web, and be so obviously "fair"
that they won't argue about that.  Everybody must be able to obtain it
freely from a web connection.

Can any of you think of a source on the web for a set of small files
with these properties:

1.  from a trusted source (here this mostly means the data is generated
   for some other innocuous purpose)
2.  represents a largely random process (temperature readings,
   stock market values, etc.) with a set generated at known intervals,
   preferably daily (at least M-F)
3.  are never, ever, revised
4.  are distributed reliably (for instance, signed files)
5.  are publicly and freely available
6.  can be obtained reliably (is available from many sites)

So far I have looked at stock market values and weather data - without
much luck.

You would think the S&P 500 is the S&P 500 and one could look it up on
any site and get the same data.  Not so! Check the Yahoo and Google
financial sites for the first few weeks of Jan. 2011 and you will find
digits that differ between the two sites in every single column.  Not
every day mind you, but often enough that it isn't reliable.  Heck, the
volume numbers differ by large factors between the two sites.  So just
choose one site and go with that?  Not so fast - if the single source
goes down the data is unavailable, and there is no guarantee that the
site (which is not party to this particular use of their data) might not
revise the page or choose to block it entirely.

Or weather data, right?  Lots of random bits there and we trust NOAA.
But good luck with criteria 3-6.  In particular, they don't give data
out for free.  In theory no US Government site should, since they are
supposed to charge to recover distribution costs.

Criteria 4-6 are typical of software distributed on mirror sites, but so
far I have not found any physical measurements which are distributed in
a similar manner.


David Mathog
mathog at caltech.edu<mailto:mathog at caltech.edu>
Manager, Sequence Analysis Facility, Biology Division, Caltech
Beowulf mailing list, Beowulf at beowulf.org<mailto:Beowulf at beowulf.org> sponsored by Penguin Computing
To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.beowulf.org/pipermail/beowulf/attachments/20110811/530523ff/attachment.html>

More information about the Beowulf mailing list