[Foundation-l] It's not article count, it's editors

[Gregory Maxwell]

On Wed, Sep 23, 2009 at 12:23 AM, Erik Zachte <erikzachte at infodisiac.com> wrote:
> Gregory Maxwell:
>> Why are people without computers or reasonable access to computers
>> considered potential audience for editing a website?
>
>> Why are people whom are effectively illiterate considered potential
>> audience for editing an encyclopedia?
>
> Your second question is redundant, few illiterate people will possess a
> computer.

Not really, they are both factors and one or the other may dominate in
some places. Possessing a computer isn't the criteria I'd impose, it's
having reasonable access to one and there are many people who are
illiterate who could be reasonably counted as having access even
though they can't make much use of it.


> Agreed, participation level does not tell the whole story,
[snip]
> So the metric is far from ideal. But still rather practical. It paints
> vividly how we are not waiting for a few minor languages to catch up, but
> how we have not reached large parts of the world yet.
[snip]

It does but that vivid painting is not a relevant one, and as such it
is potentially deceptive.

The reason "how we have not reached large parts of the world yet" is
because access to Wikipedia is significantly influenced by things
outside of Wikimedia's control and scope.

I think a reasonable argument can be made that Wikimedia's actions
could not produce a statistically significant improvement in the
penetration vs population metric; simply because the causative factors
outside of our ability to influence are so large.

You sound like you've heard me making this point and you acknowledge
that this way of measuring has weaknesses, but perhaps I did not state
this bluntly enough:  I think you're measuring something which is
largely irrelevant to the activities of the WMF.

If you instead took each language as it was and simply said that we'd
be doing well to increase the users of each language,  I'd agree. But
saying that the WMF base its decision making on a measurement of "not
reached large parts of the world yet" is effectively equal to
measuring our performance by observing that Wikipedia is only used on
1 of ~382 planets known to exist in the universe.

Failure to consider this leads to bizarre conclusions like "the lower
birth rates in the developed world compared to the developing world
decrease Wikimedia's success over time". Which is simply silly— we
shouldn't ignore our progress just because it may be dwarfed by
shifting populations.

Surely someone must have a respectable count of internet users by
language that we could use for comparison? That would be a much better
metric for our success today; while raw literate speaker numbers would
be a useful comparison for what we could start reaching with
non-internet mechanisms.

[snip]
> And the same argument of being too blunt and non discriminating can be used
> for most metrics. In fact I made it in so many words against article count,
> which treats all articles as equally important.

I fully agree with you that article count stinks as a metric in many
ways— but at least it is measuring something which is pretty much
fully under our control.

And this is really the point I was trying to make: While article count
is a bad metric because not all articles are equally helpful.
Penetration vs Population isn't even bad, because it's quite possibly
not measuring something which is generally useful to our decision
making. (Except perhaps in the decision to chase non-internet
distribution; but without literacy calibration and a comparison to a
measure of online reach it's even of questionable value there)

If we had a goal to double the number of articles in some reasonable
period of time we could do it. If we had a goal to double the
penetration, ... well, unless the WMF changes its mission nothing it
could do would get us there.