very interesting thoughts! How shall we get a response about a potential problem from a dissatisfied user that is about to leave the page, or even the site. I don't know. John E
Lars Aronsson wrote:
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?