Hello Jimmy Wales and other Wikimedia Foundation members,
I'm writing you to propose a Wikimedia project (tentatively called
"WikiSolve") no less important than Wikipedia.
We know Wikipedia can teach people knowledge (in terms of concepts),
but it can't directly help people find solutions to their problems,
because it's "concept-oriented" rather than "problem-oriented".
I envision a wiki that collects virtually every known problem in the
world and their corresponding solutions, so that people with a problem
in mind can find a solution on it.
The key problem in designing such a wiki is how such a wiki can guide
the user to the problem page he wants. I believe a hypertext-based
mechanism called "troubleshooting wizard" is the answer. A good
example of a troubleshooting wizard is
http://support.hubris.net/dialup/wizard/ . As you can see, this is a
way for the user to locate his problem in a wiki without knowing
keywords used to name or describe that problem, just like Wikipedia
allows a user to locate a concept without knowing its name or any
keywords used to describe it.
There is actually more background to this idea. I strongly recommend
you read the following article that compares how AI and a wiki tackle
two old problems differently: knowledge representation and problem
solving:
----------
(A formatted version of the following article is at
https://plus.google.com/u/0/102291835965130378165/posts/finzeoipR7A)
From formal to semi-formal: knowledge representation
and problem
solving in the AI way and the wiki way
1. Failure of the formal way to represent encyclopedic knowledge
Big thinkers like Leibniz, Dijkstra and John McCarthy all dreamed
about an encyclopedia written in formal language and an automated
reasoner that could solve a problem by reasoning on this formal
knowledge base. Unfortunately attempts at this like the Cyc project
still have a long way to go.
2. Success of the semi-formal way to represent encyclopedic knowledge
In contrast, Wikipedia is a big success. Most stuff on Wikipedia is
written in natural language, but Wikipedia does have some formal
elements. Most fundamentally, each concept on Wikipedia has a unique
formal name, and there are hyperlinks between related concepts,
enabling the user to navigate to a target concept without initially
knowing its name (which makes Wikipedia an important "global
positioning system" (GPS) for concepts).
3. What would be the wiki (semi-formal) way to problem solving, then?
When it comes to "problem solving", there are actually two kinds of
problem solving:
3.1. Wiki-based solution sharing
The first kind is when you have a problem already solved by experts,
and these experts want to create a wiki as a "solved problem base"
where you can easily find your problem and consequently see the
corresponding solution written by these experts. Now the question is:
how can such a "problem base" wiki be organized so you (the user) can
find your problem easily?
What I want to say is "troubleshooting wizard". Do a Google search for
[ troubleshooting wizard ], and the first result is a good example of
what it is like:
http://support.hubris.net/dialup/wizard/
As you see now, a troubleshooting wizard uses a series of questions to
let you specify your problem's characteristics (or "symptoms"), and
eventually leads you to a solution to your specific problem. You will
find this immediately familiar because you probably already saw this
kind of thing in Windows XP's Help System.
Now you can realize that a wiki as a hypertext system can surely
implement a troubleshooting wizard that walks the user to his problem
in a "problem base" wiki.
3.2. Wiki-based problem solving
The second kind is when you have an open problem that doesn't have a
known solution (otherwise you're supposed to find its solution in a
"problem base" wiki as discussed in Section 3.1). Now if you want to
attack this open problem on your own, creating a wiki may help, for
the following reason.
During your problem-solving process you may need to divide the
original problem into subproblems, or apply certain strategies such as
"generalization", "specialization" and "analogy" to the
original
problem to obtain some "derived problems", whose solving may help you
solve the original problem (this is what George Polya's famous book
"How to Solve It" talks about). To keep track of these "subproblems",
"derived problems" and other kinds of middle results, a wiki would be
a great organizer.