Re: [Eligibility-discuss] Exhausted pool [draft-knodel-nomcom-gender-representation-00]

[Michael StJohns <[e-mail]>]

On 11/25/2023 4:53 PM, Brian E Carpenter wrote:
> On 25-Nov-23 23:11, Michael Richardson wrote:
> ...
>
>> I would like the document to deal with what happens if one gets to 
>> the end of
>> the pool and there are still empty slots.
>
> I think you have to withdraw the new rule, and accept that all ten 
> will have
> the same gender that year.
>
> (Geek question: What would happen with today's rule, if *all* volunteers
> were from four large companies, such that the maximum-2 constraint could
> not be satisfied? That won't happen, but it illustrates why 
> over-constraining
> the criteria could get us into a dead end.)

Last I looked, that's unlikely to happen.  I've been seeing between 
60-75% volunteers from large companies.  With the current document, that 
would result in a group of 8 voting volunteers.

>
>    Brian
>
Nah -

It's all in how you sequence things.  Pick the required people first, 
then fill in with the rest.

A is the set of all volunteers, W is the number of women to select, V is 
a selected volunteer

1) Create a subset U from the set of all nomcom volunteers A where 
gender != male

2) For i = 0; i <= W; i++; randomly select a member from subset U, add 
them to their organizational class count.

3) For i = W; i <= 10; i++ {

     do {

      3a)  randomly select a member V  from A

      3b) if (V is a member of a capped class (e.g. count of 
organization class >= 2) continue;

      3c) add V to its organizational class count; break;

       } while (true);

}