A proposal about how to replace categories

[David Hoffman]

As you can imagine, it's hard for a newcomer to step into such a long set of threads. Here's a stab at it...

It seems to me that CZ folks are too concerned with arranging the database with some kind of top-down logic. The workgroups, disciplines, etc look like rather static "folders" or "directories," which is fine for paper. However, esp for Internet search and organization, folders are not as useful/dynamic as would be tags/labels. Sure, it's taken me time to adjust, esp since I'm way into logical trees and order -- yet it's clear Google, gmail, technorati, delicious, spotlight, etc etc are moving us to be more productive through tags. I don't think we need to try to control or manage the use of tags -- we can make systematic suggestions*, of course -- but when people end up coding with tags, the most useful tags rise to the top. For wikis, the closest thing to a tag is a category. So, it seems important for CZ to develop categories, let them flourish. They are a great ease for readers, i.e. the audience we hope will eventually use CZ, both for searching and for getting a sense of how knowledge is organized and inter-related.

* To give guidance on tags, perhaps we can be given data on the tag distribution maps for major sites. And then somebody can write an app to generate tag map for CZ.

From an academic standpoint, tags/categories also make sense because (at least in humanities and social sciences) we are moving further away from static disciplines. How can we list an article only under workgroups, when new interdisciplinary innovations (i.e., "tags") are constantly emerging?  My 2p. Thanks. 

P.S. Perhaps we would be wise to move away from framing the problem in terms  of "taxonomy," which comes out of an arguably outdated, idealistic epistemology (see Foucault, The Order of Things).

[Chris Day]

Perhaps we would be wise to move away from framing the problem in terms  of "taxonomy," which comes out of an arguably outdated, idealistic epistemology (see Foucault, The Order of Things).

Aren't workgroups a bit like tags?  Articles can be in more than one.  I think the workgroups are more about organising contributors than articles. In that sense it helps like minded editors find each other.

As far as i can tell the CZ taxonomy will be defined by links in articles so it will be much more like a network than anything.

[Derek Harkness]

From an academic standpoint, tags/categories also make sense because (at least in humanities and social sciences) we are moving further away from static disciplines. How can we list an article only under workgroups, when new interdisciplinary innovations (i.e., "tags") are constantly emerging?  My 2p. Thanks.

Technology has advanced beyond categorisation and tagging. We no longer need to summerise the content of a article into a tag list or category. Instead we can make use of full text and link based search indexing to organise documents into virtual, custome made, groups on demand.

You can see the progression if you look at the history of sites like Yahoo! Ten years ago, yahoo was a massive directory. It was full of categories and each page was listed under one or two categories. The listing within each category was simply alphabetical and it required manual matenance by humans. This worked well for a while, but then the web got bigger. More advanced technology was needed to keep up with the number of pages. Human input an review was replaced by automated spiders that crawled the web. Meta tags were replaced with full body searching. Other technology like link tracking and page ranking made the automated systems more accurate and the result more relevent.

The web has moved on beyond categoriesing into directory structures or manually tagging. It's about time the wiki moved on also.

[Robert Winmill]

Unfortunately with the use of spiders, corrupt (pages gaming the search technology) web pages, and the search engines themselves being gamed for revenue production most search results end up with a lot of junk.  And one can get lost and frustrated in the information super bog.  Categories and other forms of structure are necessary.  However, they are hard to do and maintain.   But isn't that a role of the editors at CZ?

[Derek Harkness]

Unfortunately with the use of spiders, corrupt (pages gaming the search technology) web pages, and the search engines themselves being gamed for revenue production most search results end up with a lot of junk.  And one can get lost and frustrated in the information super bog.  Categories and other forms of structure are necessary.  However, they are hard to do and maintain.   But isn't that a role of the editors at CZ?

Actually the converse is true. The directories an tag based searches were easy to game - or rather the word is spam - since they rely on human input and those humans lied. It is still possible to spam the search engines, but it requires significantly more skill and technical knowledge than was required to tag a web page to a different keyword.

But why would you spam the wiki's search engine. Search engine spam is done to promote a web site, usually for commercial purpose. Nobody makes anything by pushing one encyclopedia article to the top of the search results. The only reason I can think of is malice and that sort of vandalism would be rare in the CZ community where you have your real name and reputation to stake.

[Pieter Kuiper]

Discussion seems to have stopped...?

I am new here, and I find this ban on categories incomprehensible. While the workgroup category allows me to see what articles there are about physics, it is difficult to see this site's holdings on say Sweden.

The great advantage with categories is that they self-organize. However, lists require maintenance, and the "related articles"-scheme requires maintenance of many different lists.

[Stephen Ewen]

The great advantage with categories is that they self-organize. However, lists require maintenance, and the "related articles"-scheme requires maintenance of many different lists.

I agree.  Categories, though faulty, are a greatly labor-saving device.  With the current system, we may have 30 articles on Sweden but they are too hidden until someone creates a catalog page about them, rather than simply adding upon creation, or afterward flying through them and placing [CTL+V] [[Category:Sweden]], [[Category:Fauna of Sweden]], etc.  But the categories could greatly facilitate the catalogs. We'd just need to create a canon of categories so as to avoid category chaos. 

[Larry Sanger]

Pieter, Wikipedia's conceptually confused category feature took considerable time to get started as well; it was useless for a long time, too.  (I think it still is, but that's another matter.)  Obviously, the Wikipedia approach requires maintaining many lists as well, and it is considerably harder to do so, because one must go to a separate article to edit every single entry in a given category.

In the long run--when we have filled in the broad strokes and more people understand how to create Related Articles pages--it will be obvious that the Citizendium approach is easier both on the user and on the contributor.

[Aleta Curry]

The great advantage with categories is that they self-organize. However, lists require maintenance, and the "related articles"-scheme requires maintenance of many different lists.

I agree.  Categories, though faulty, are a greatly labor-saving device.  With the current system, we may have 30 articles on Sweden but they are too hidden until someone creates a catalog page about them, rather than simply adding upon creation, or afterward flying through them and placing [CTL+V] [[Category:Sweden]], [[Category:Fauna of Sweden]], etc.  But the categories could greatly facilitate the catalogs. We'd just need to create a canon of categories so as to avoid category chaos. 

Don't be fooled.  Categories require their own set of maintenance, and having done this sort of thing, I can tell you it's no joke.

Also, you're not kidding about the "canon of categories" to avoid chaos, Stephen.

One big problem is that you have everyone creating their own categories, and not necessarily knowing what exists already.  So no one category is comprehensive, and indeed, one can believe that one is looking at a category with all the articles about X, when actually, those are split.

I'll try to give an example (and I'm just making this up, got no time to go looking for examples at WP)

*Category of American Actresses.  Okay, is that North and South American actresses, or US only?

*Category of United States Actresses.  That one's clear.

*Category of United States Actors - For some, this includes actresses.  For others, that is sexist, and so female actors go in

*Female Actors in the United States.

*US film actors

*Actors in US theater.

*Broadway actors

On and on and on and on....  It's a mess.  Some people are in several, some in one only, some in none....

In the long run--when we have filled in the broad strokes and more people understand how to create Related Articles pages--it will be obvious that the Citizendium approach is easier both on the user and on the contributor.

The cluster approach is definitely going to be a great help here.

Time will tell as to whether or not this works.  If it doesn't, the worst that will happen is that we'll all have to start putting categories at the bottom of our article pages.  I don't see that as major disaster.

However, what we still have to figure out is how to deal with interdisciplinary topics.  And those that fall into no current workgroup.

[Pieter Kuiper]

Categories require their own set of maintenance, and having done this sort of thing, I can tell you it's no joke.

Of course. And I have also done it, quite a bit at Commons. Many new pages are not categorized, others overpopulate high-level categories, etcetera. I feel that such work is worthwhile, much more so than compiling lists.

[Joe Quick]

I expect (and maybe it's just me, but I don't think so) that there will be categories for sub-workgroups once we are big enough to justify sub-workgroups and that there will be categories for area studies once those workgroups get started.  That should cover things pretty well, I think.

[Stephen Ewen]

http://ontoworld.org/wiki/Semantic_MediaWiki

[Joe Quick]

http://ontoworld.org/wiki/Semantic_MediaWiki

That looks handy but it's beyond my ken.  It would be prone to the same sort of proliferation as categories, I assume.

[Daniel Mietchen]

Background: I confess to a perhaps overextreme animus against Mediawiki's Category feature.  Yes, the feature is useful when it comes to categorizing articles for, say, workgroup assignments, or for categorizing images.  But when it comes to categorizing articles, well, the information contained on category pages is the same sort of information that one looks for in an encyclopedia article.  That is, a list of subtopics is perfectly valid as a sort of end matter as are lists of Further Reading and External Links.  This suggests, then, the following proposal.

It's worth noting that several MediaWiki developers are working on turning the present MW categories system into something closer to "tagging" - with the ability to combine tags in a search.

What's the current state on these tagging initiatives?