Looking at the current
http://quality.wikimedia.org/
it seems that "Wikimedia Quality" is just a new name for
Wikipedia 1.0 or "stable versions". It might be a good name.
From another perspective, the word "quality"
can be understood in
many different ways. One of them is the international
standards
in the series ISO 9000, deployed in most industries in the western
world in the 1980s and 1990s.
Do the founders of Wikimedia Quality have any background in
industrial quality management, quality assurance or "six sigma"?
I'm not saying that all methods used in manufacturing industries
are directly applicable in Wikipedia. But I think that being
completely clueless can be harmful. So is there a need to read up?
In the modern industrial sense of "quality", it is always a
measurable entity, compared to a stated goal. It is essential
that the producer and consumer share an understanding of the
purpose of the delivered product or service, before you can start
to measure how well that purpose is met. A car with an expected
life of 5 years can be of good quality if this is what the
customer wants and expects and the car does last for 5 years.
This is in sharp contrast to the common view of the "man in the
street", who believes a car is of better quality if it lasts for
70 years than if it lasts for 40 years, the longer the better.
If a car lasts 40 years, increasing its life expectancy to 70
years might include gold-plating all electrical connectors. This
might make the car a lot more expensive, and that would be a waste
if this customer only needs this car to last for 5 years. Other
"improvements" might make the car more bulky and less economic in
other ways. Avoiding such suboptimal "improvements" is much what
quality control is about.
In designing for quality, it is essential that customer/user
expectations are investigated and fed back. There needs to be a
feedback loop of expectations and learning from experience into
the producing organization. This is usually illustrated as a
cycle of four steps: Plan, Do, Check, Act (PDCA). There's even a
Wikipedia article about that,
http://en.wikipedia.org/wiki/PDCA
In my opinion, the sharpest difference between proprietary,
commercial software and free software of the same kind, such as
Microsoft Office and Open Office, is that success or failure in
sales and marketing causes a strong feedback to the developers of
Microsoft Office. If the product fails to sell because it lacks
some feature, there is a very strong incentive to add that feature
in the next release. Even if free software is often stable and
reliable, its evolution is often slow and unpredictable and seldom
guided by the needs of potential users.
I'm not advocating properietary and commercial software. I'm a
Linux user since 1992 and an Emacs user since long before that.
What I'm saying is that "their system" (Microsoft's and Oracle's)
has a feedback loop that we could wish for.
Since Wikipedia has borrowed so much from the free software
movement, it has also inherited the lack of this strong feedback
loop. Both free software and Wikipedia do have another feedback
loop, where each user is encouraged to become a programmer and/or
text editor, but this mechanism is a lot weaker. Most
dissatisfied users will not become programmers/editors, but will
just silently drop out of the loop. In terms of control theory,
this is equivalent to attenuating the feedback loop, causing a
much slower signal response,
http://en.wikipedia.org/wiki/Control_theory
So, one way to improve the quality of Wikipedia could be, I think,
to somehow capture that lost feedback.
Is this on the agenda for Wikimedia Quality? Should it be?
--
Lars Aronsson (lars(a)aronsson.se)
Aronsson Datateknik -
http://aronsson.se