> From: Scott Brim <swb@xxxxxxxxxxxxx> > > On 03/11/13 14:41, Mary Barnes allegedly wrote: > > This year's set of nominees was far more diverse than in the past and > > yet the IESG will still be entirely male and entirely North > > American/European. Of course, only people that bothered to use the > > tool to input comments would see that. So, indeed the nomcom process > > is part of the problem. > > Mary: I believe you would agree with this but your language doesn't seem > to say so: just because the nomcom chose a less diverse set of nominees > from a more diverse set of candidates doesn't mean there is something > wrong with the nomcom or the nomcom process. That is true, but this message of Mary's is one of the few items of *data* we have on the subject: The diversity of the selected people (along the dimensions we are considering) is noticably smaller than the diversity of the pool from which they were selected. So we can conclude that there is some factor operating within the Nomcom process that we should examine and analyze and may wish to change. (As opposed to the possibility that the output of Nomcom has the same diversity characteristics as the input of Nomcom, in which case, trying to fix the Nomcom process would probably be ineffective.) In regard to data, I see that the statistics on nation of authorship for "recent RFCs" (http://www.arkko.com/tools/recrfcstats/d-countrydistr.html) is even more US-heavy (~70%) than the statistics for "all RFCs" (http://www.arkko.com/tools/allstats/countrydistr.html) (~50%). It's possible that IETF participation even at the purely technical level has become less diverse over the past decade. In regard to data, I expect that there are several levels of participation which may have very different statistics due to the difference in time/money costs involved: 1. participants in WG mailing lists 2. authors of drafts 3. regular attendees of the meetings 4. WG chairs 5. AD/IESG/IAB members My guess is that groups 3 and 4 are fairly similar, but that group 3/4 may differ strongly from group 5 and groups 1 and 2. Dale