🚀 go-pugleaf

In article <a4rd38.775.ln@anacon.demon.co.uk>,
  Mark Evans <mark@anacon.demon.co.uk> wrote:
> Nigel Orr <news@river-view_dot_freeserve.co.uk> wrote:
> > On Wed, 08 Dec 1999 11:58:17 GMT, R.E.Ballard (Rex Ballard)
> > <r.e.ballard@usa.net> wrote:
>
> >>Actually, there is a rather simple sampling method.  First, you take
> >>very popular sites, and have them assign cookies to each user.
>
> > Far from perfect results though- unless you know how many people are
> > using JunkBuster and similar cookie-confusers.  For example, I
browse
> > as an old Mac version of Netscape, which is the default with my
> > version of Junkbuster...
>
> Also people might do this sort of thing if they encounter
> braindead sites which make silly assumptions using the
> user agent field.

Remember,

The goal here is to come up with a set of relative proportions of
similar users that is likely to represent "1 user 1 vote".

Like any "election" or "poll", you are going to have some odd-balls.

What you are trying to avoid is coming in as 1 IP address 1 vote,
since your chances of being the one vote counted out of thousands
of dial-up users using "ms-chap" POPs vs the few pops that are
"Linux friendly" could skew the votes.

Search dejanews for the StatMarket debate of a few weeks ago.

--
Rex Ballard - Open Source Advocate, Internet
I/T Architect, MIS Director
http://www.open4success.com
Linux - 60 million and growing at 2%/week!


Sent via Deja.com http://www.deja.com/
Before you buy.