Re: [ga] More thoughts on a Registrants Constituency

[George Kirikos <gkirikos@xxxxxxxxx>]

Hello,

--- Karl Auerbach <karl@xxxxxxxxxxxx> wrote:
> So the formula I suggest is this, where Y is the number of years that
> have 
> elapsed since registration.
> 
>    Votes = 2**(Y-1)
>    (i.e. the number of votes is 2 raised to the power Y less one)
> 
> Thus the registrant would get votes according to the following table:
> 
>   YEARS     VOTES
>       0     0
>       1     1
>       2     2
>       3     4

Not a very good formula, IMHO (and I own many old domains, so would
benefit from this). I'd dub this the "AARP voting scheme" (see
www.aarp.org for those unfamiliar with that organization).

Continuing that table above:

4 years ---> 8 votes
5 years ---> 16 votes
6 years ----> 32 votes
7 years ----> 64 votes
8 years ----> 128 votes
9 years ----> 256 votes
10 years ---> 512 votes
11 years ---> 1,024 votes
and so on.

In the real world, 30 year olds don't get 1024 times more votes than 20
year olds, 40 year olds don't have 1024 times more votes than 30 year
olds (and rougly 1 million times more votes than 20 year olds) and so
on. If they did, those 65 years of age and older (the AARP members)
would rule. :)

Since registrants contribute linearly (per domain) to the current
funding costs of ICANN, a better formula would reflect this (although
registrants of different TLDs pay differing amounts).

Assuming all the domains were from .com, a formula that reflected the
domains per registrant, and percentage of voting domains relative to
the number of domains might be something like:

R = total registrants
Q = voting registrants ( Q<=R)
N(1) = domains owned by registrant #1
N(2) = domains owned by registrant #2
N(j) = domains owned by registrant j
T = total number of domains = N(1)+N(2)+....+N(R)
V = total number of VOTING domains = N(1)+N(2)+.....+N(Q)  
V <= T

Votes for Registrant J = [N(j)/T]**[V/T]
Total Votes = sum(j=1 to R, V(j)) 

If V = T, we have linear weighted voting by number of domains

With V<T, the above should ensure that a single large domain registrant
can't capture the constituency, because the exponent becomes less than
1 and thus their vote becomes less significant the bigger they are.

The above is just off the cuff, though, and other weighted schemes
could make sense too (although not the AARP voting scheme!).
Sincerely,

George Kirikos
http://www.kirikos.com/