[RFC] Rethink the voting / decision making process

[Jo-Philipp Wich]

> do you want default-allow or default-deny?

I'd favor a default-allow model.

> How many votes do you think you will have? If you are thinking in
> terms of <10 votes, then a couple bad voters in the mix can be a huge
> problem.

Personally my vision is to have several dozens at least, simply to let
voting results reflect somewhat common consensus among the wider community.

> But if you have 100 voters, a handful of bad voters isn't going to 
> matter. If you have 1000 votes, even less so.

That's my thoughts as well.

> how many people with commits in the last 6 months (don't forget to 
> include commits to the feeds)

I think the info can be easily extracted from Github, see e.g. here:
https://github.com/lede-project/source/graphs/contributors

> can someone look at the OpenWRT forum logs and extract some commit 
> counts (anonomised if needed, but ideally you should be able to
> filter out the people who also appear in the first group)

https://forum.openwrt.org/userlist.php?show_group=-1&sort_by=num_posts&sort_dir=DESC&search=Search

> I suspect that the voting pool is going to be much larger than you 
> expect. If you are thinking in terms of the half dozen former
> OpenWRT developers who founded LEDE, every vote is extremely
> significat.

We tried to avoid having only former OpenWrt people early on and having
the feed maintainer participate in the decisions would be great - so I
guess it will be something in the 40-80 people ballpark (just a hunch
though).

> But if you find that in 6 months you have a few hundredpeople with
> commits (even if they are drive-by commits), and then add a similar
> number of other contributers, a few upset voters are just noise.

Yes I think so too.

~ Jo