Groupprops:Subject wiki vision

The subject wiki vision is a derivative of the multiple wiki vision.

The problem
This is the problem we're trying to address: an increasing, and largely unfulfilled, need for specific, specialized information in particular subjects, where the information is available but not easily reachable. For instance, I might have a complex question like: What do you call a group which has no Abelian normal subgroup? Or I might have a question like: What are the various ways of proving that a given group is nilpotent?

More importantly than questions to which we feel the need for pressing answers, are questions which flit across our minds, and are then lost. Every question is the beginning of an act of creative exploration. The problem is the lack of tools to harness those beginnings and begin a progressive journey of creative exploration.

The problem is not the absence of a solution: it is the non-availability of a solution. I've had situations where some of my open problems in mathematics had solutions hidden within textbooks that I had at home all the time. But I didn't know where to look. In other words, the problem is an inherent inefficiency in the knowledge market: knowledge seekers and knowledge providers do not use each other as often as they could.

Does the problem really exist?
Ten years ago, we weren't aware of the problem, because we couldn't dream of a world where it would even be possible in principle for anybody to get any kind of information any time they wanted. Today, with the growth of search engines as a powerful tool to surf the Web (notably, Google), the growth of encyclopedias (notably, Wikipedia), and the growth of online courseware and textbooks (notably, MIT Open CourseWare and Connexions), we see that knowledge can, after all, be disseminated, not just with high quality and zero cost, but in a way that helps seekers and providers meet each other very efficiently.

In other words, the knowledge market today is revealing needs that simply weren't expressed in earlier knowledge markets, because there were practically no opportunities. Fifteen years ago, it was unthinkable that a student in India could be part of a classroom at MIT. Today, it's possible and widespread.

As a new collection of needs is met, we also expose ourselves to a new collection of unmet needs. This is the need for specific subject-related information, in a context-independent fashion. I'll explain what I mean by each of these in a little bit.

The problem of scale and priority
Compare:

http://harrypotter.wikia.com

against:

http://math.wikia.com or http://physics.wikia.com or http://chemistry.wikia.com

Some of the best and biggest wikis today revolve around cultural phenomena. This includes Harry Potter, World of Warcraft, and Star Trek. These are what I call phenomenon wikis. They're about a phenomenon that has captured the interest of a large segment of the population, some of which is willing to go to some lengths to document and organize the ideas and information of the culture. I'llcall these phenomenon wikis.

Phenomenon wikis cater to very specific phenomena, and the successful among those revolve around phenomena that appeal to a large number of people. Marketers often talk of the bell curve of marketing. The key characteristic of the Bell curve is that the top few items are extremely popular, and there's a long tail: a large number of items with low popularity and specific niches. Harry Potter, World of Warcraft, and Star Trek cater to the top segment.

On the other hand, phenomenon wikis are, after all, about phenomena. A lot of people may be interested in and inspired by Harry Potter today, and it's likely to remain a popular book for a long time. But Harry Potter isn't a subject. Fantasy fiction is a subject. Harry Potter is a specific phenomenon in the subject. So what do we mean by a subject.

What's a subject?
A subject should be something that:


 * Has long-term value and interest, that arises out of general applicability
 * Is important enough for at least some people to earn degrees, or do jobs, in. This should be over a sustained period of time. So Harry Potter may help some people earn money today, but it isn't a viable stream for revenue and research over the long term.

The next question is: what is the granularity at which we're looking at subjects? Do we consider science as a subject? Do we consider mathematics as a subject? Do we consider genetics as a subject? The granularity matters for a lot of purposes, as we shall see below.

What's a subject wiki?
The idea behind a subject wiki is simple: a wiki devoted to a specific subject. The principles behind organizing content on the wiki, the way pages are interlinked, and all the other policies, are specifically designed to meet the needs of that subject. The focal lens are those of that subject. There may be articles on related topics, but the structuring of those articles is determined by the subject.

Different subject wikis could have articles on the same subject from fairly different perspectives. For instance, compare group (on Groupprops) against group (on Diffgeom). The former is a page on group in a wiki devoted to group theory. The latter is a page on group in a wiki devoted to differential geometry. The perspectives differ.

The potential
Some people, when they hear about the group theory wiki, say: "A wiki on group theory? What could there be in group theory to fill an entire wiki? Isn't there enough stuff in the books already?"

The need for subject wikis is a latent need. When Wikipedia first started, people may have said the same thing: "isn't there Britannica?" When computers first started offering word-processing capabilities, people may have thought: "what's better than paper and pencil?" But once these things took off and started running, people found more and more of their latent, hidden needs, being met. So instead of asking "What's the need for subject wikis?", let's concentrate on asking, "What's the potential?"

Here's the potential:


 * Quick, one-point access to definitions from the subject perspective
 * Quick, one-point access to facts from the subject perspective
 * Quick, one-point access to survey articles, or expository articles, on a specific need, from the subject perspective
 * An easy way to navigate between definitions, see related terms, spot the interconnections
 * A tool to understand the important problems, ideas and methods in the subject

The difference between wikis and other online resources
MIT Open Course Ware, Connexions, and other open courseware is fantastic. Subject wikis are not meant to supplant this (and in fact it cannot). If open course ware is like somebody driving you through the city, subject wikis are like a map for you to explore. Subject wikis are a more passive exploration tool, but they can exploit the latent creativity of eager learners.

A paradigm pool
When I said, a little while above, that every subject wiki chooses its own organizational principles, this doesn't mean that the organizational principles have to be reinvented. Rather, I envisage a shared paradigm pool. There are two paradigms already in use in the mathematics wiki I've worked on: the property-theoretic paradigm and relational paradigm. The relational paradigm is fairly general, and is likely to be applicable to all subject wikis, though probably with more modification and refinement. The property-theoretic paradigm, on the other hand, may not be that useful in other subject areas (or it may need to be significantly modified).

As more and more subject wikis grow, each adds to the paradigm pool, and each modifies existing praadigms in their own subject wikis. An independent development of these paradigms can lead to a thriving ecosystem of paradigms.

Why wikis?
Some of the reasons why wikis are being suggested:


 * Easy, transparent markup and organization, and a number of features facilitating strong collaboration
 * Strong internal linking, leading to excellent internal cohesion
 * Reasonably good features for content management and organization, leading to easy navigability
 * A reasonable degree of modularity and programmability (through templates)

The subject wiki version isn't wedded to MediaWiki. We might move to a semantic MediaWiki concept soon, though perhaps not the same one as the semantic MediaWiki currently being experimented with. We might move to some system for conceptual organization that layers on top of the wiki. But at this stage, the wiki seems the best way to get started. What further tools we need will become clear as we create more and more subject wikis.

The challenge of licensing
The current plan is to have all subject wikis under a CC-by-SA license. This specifically means that others can build upon the work, and can use it for commercial purposes, including remixing it into lectures, writing books etc., provided they provide attribution. Being attributed is the coin of the realm in academics, so this could be a motivating factor.

The challenge of credibility
Can subject wikis be accurate and reliable? We'll look into ways of assigning credibility, and having approval and peer review processes. Also, the use of real names for all contributors may be a good idea. At the least, those working on managing the subject wiki, should use their real name.

However, peer review and approval features do not need to be built into a subject wiki from the start, because we as yet have very little idea of the kind of approval mechanisms that we'd need and the kind of audience interest that we'd generate. my guess is that in the nascent stages of a wiki (where nascent could mean anything ranging from three months to one year), the wiki should be free to experimentation, and temporarily incorrect content shouldn't be viewed as a deep problem.

For instance, the group theory wiki here has errors, some of them in the definitions (though this isn't very frequent) and some of them at more obscure parts of the article. It doesn't have the same degree of accuracy as a reviewed, edited and published book. But it has roughly the level of accuracy as notes for a first draft of a book, or class notes, or lecture notes prepared by a lecturer (without extensive editing for publication quality). As more and more people read the contnet, review it,correct it, improve it, or provide feedback, the level of accuracy is going to progressively increase. The increase will not be dramatic, but it'll make the difference from most probably correct to certainly correct.

The challenge of motivation
Why should people contribute to a subject wiki? This question cannot be answered straight-off, because what motivates people is a complex issue. Certainly subject wikis cannot rely on the same large-scale contribution effects that Harry Potter, Star Trek and World of Warcrafts do. On the other hand, because there are people making a living off the subject, there are more people who have a serious and vested interest in inrceasing the ready availability of information on the subject.

The challenge is to demonstrate that working on the wiki is a powerful way to meet both existing and latent needs for the subject. In this sense, the group theory wiki is a pilot project to demonstrate that powerful way. The two factors discussed above: the licensing model, and credibility, are extremely important. Additionally, it is important to have a mechanism for giving people credit for the tremendous effort that goes into organizing and improving a subject wiki. Mechanisms for this may require contributors to use real names, and may require a culture different from that prevalent on Wikipedia (though perhaps similar, in some senses, to that currently on the Citizendium).

The most important factor, in my opinion, is granularity.

The challenge of determining granularity
At what level do we pitch the subject wiki? Do we have a subject wiki on mathematics? My guess is no. I just want to point to:

http://math.wikia.com

It has less than 200 articles! Check out the Wikia physics and chemistry wikis, as well.

The problem with having a subject wiki on a very broad subject is that if a few people start it out and take it in a direction that other people do not like, then it just dies down. So, a reasonably specific subject wiki can allow people to do individual experiments. Too broad a scope could constrain creativity.

While keeping the scope specific, it's still important to make it a subject wiki rather than a phenomenon wiki. Wikis that are devoted to a specific course taught at a specific university are important and useful, but they do not qualify as subject wikis. Subject wikis need to be out of a context or phenomenon, and be sufficiently widespread in space and time, even if they're at the far end of the long tail.

This doesn't mean that subject wikis have to. always be on advanced topics. In fact, we could have a subject wiki on high school geometry or on math for toddlers. But a subject wiki on mathematics is too broad.

Competition among wikis
Should we promote the creation of multiple wikis revolving around the same topic? My guess about this is yes, as long as each new wiki proposes to offer some new type of value, or perspective, that is missing from existing wikis. The idea is not to tell people to set aside differences and work on the same wiki, but the idea is not to create meaningless competition among wikis that do exactly the same thing (though the license doesn't forbid this). So if there's already one wiki in group theory, another wiki in group theory should offer some distinctive new flavor or value. It can, of course, build on content in the previous wiki (that's the point of the Creative Commons license). This could lead to healthy competition between the wikis.