On Sun, 5 Nov 2000, William X. Walsh wrote:
> > I understand. In fact the estimate is correct. Unfortuantely I don't
> > have the time to teach you statistical analysis. But we can be confident
> > that 5% of internet traffic is non USG. My results are representative and
> > can be extrapolated accordingly with some confidence.
>
> You have absolutely no basis for making that assumption. The number
> of nameservers queried is in no way directly representative of the
> number of nameservers actually used for end user name resolution, nor
> of the number of queries made on a per server basis.
No it in fact does. There are some 300,000 dns in the dot.com file, to
date 60,513 have been surveyed (15%). It's easy and completely acceptable
to extrapolate from there. And furthermore the existing population
enumerated is more then enough to extrapolate from.
>
> To make the claim you made, you would need a lot more data than you
> can get from doing the simple queries you are doing, and as a matter
> of fact this would require the co-operation of the nameserver
> operators.
>
> Your method is not scientific, and it no statistician would ever make
> the leap in logic you tried to make here.
Get yourself a statistician and I'll be happy to provide him with the
numbers. Like I always say - talk is cheap and evidence dont lie '=)
regards
joe
--
Joe Baptista
http://www.dot.god/
dot.GOD Hostmaster
+1 (805) 753-8697
