RE: [ga] NC members, censorship and other absurd things

["Roeland M.J. Meyer" <rmeyer@mhsc.com>]

> Behalf Of 'Kent Crispin'
> Sent: Thursday, November 11, 1999 8:17 AM
>
> On Wed, Nov 10, 1999 at 11:44:00AM -0800, Roeland M.J. Meyer wrote:
> > > This is a meaningless distinction.  The list operator can
> > > *always* see
> > > email messages before they are distributed, and censor them.
> >
> > At what cost? This is a very time-intensive approach.
>
> Actually, it's not bad at all.  It just means that the moderator has
> to skim all the messages.  I moderate a list with about 475 members,
> and it is a trivial load.  Basically you just skim the messages very
> quickly.  99% of the time there is nothing to decide -- you just
> approve the message.

I think that the number, for these list, would be much lower than 99%. You
are optimistic. Then there is always the complaints of censorship, which the
moderator would be vulnerable to, regardless of justfication.

WRT the moderator load, increase your numbers by two orders of magnitude and
you begin to see the problem. The problem is that your approach DOES scale
linearly. If the GA is to be successful, the membership needs to increase by
at least three orders of magnitude. At those numbers, your brand of
moderation does not scale. There is also the issue that human attentiveness
drops off with inverse proportion to the number of items being reviewed.
IOW, the larger the sample size, the less attentive the reviewer,
consequently, the larger percentage of items that slip through the review.
This places an upper-bounds constraint on the number of mmessages that an
individual can effectively review. Ergo, you need more individuals as the
membership gets larger. This goes directly to cost.

> > As the population of
> > the GA rises, this approach is not scaleable.
>
> It goes up linearly with the number of posts.  Total time to moderate
> the ga list at the current rate would be less than 10 minutes/day.
> Bear in mind that a moderator at this level doesn't have to think
> very much about the posts -- most of the time the decision to approve
> can made in a few seconds.
>
> The real problem with moderated lists is not so much the time it
> takes, but the delay it can impose.

Which reenforces my claim that it doesn't scale well. Why would it impose a
delay if it were scalable?