[Foundation-l] Article Feedback Tool 5 testing deployment

[Jussi-Ville Heiskanen]

On Sat, Dec 24, 2011 at 10:14 AM, Liam Wyatt <liamwyatt at gmail.com> wrote:

> IMO, a good attitude to have so long as you're not trying the same thing
> and expecting a different outcome. In the case of the AFT for example, you
> can clearly see in the documentation that each iteration (versions 1-4 and
> now 5) have built on the best and discarded the worst of each previous
> model. As I've stated, I hold much hope for the usefulness of AFTv5, which
> is being build learning from past experience. Which is why I would like to
> see AFTv4 removed from the remaining 99.7% of en.wp now that it is being
> superseded.

This is both true in the absolute, and absolutely misleading in the real world,
as a Jesuit might put it.

You paint the current effort as one of iterations which will "improve" but in
actual fact they are iterations that approach asymptotically the one thing
that was demonstrably a total timewaster and failure, namely Nupedia.
Just because you are fiddling the bits doesn't mean you are improving
things. Usually you are just making them fail less badly. Not a way to
design winning mechanisms.


-- 
--
Jussi-Ville Heiskanen, ~ [[User:Cimon Avaro]]