3

I'm looking for an open source knowledge base system that uses tags, rather than free-text search to identify articles (a lot like serverfault does).

I've looked at twiki, which many people suggested, but haven't found what I'm looking for. Basically I want to be able to create and tag articles, and provide an easy way for anonymous users to search based on tags.

Edit:

OK, here's some more detail regarding what I want. Basically, all the knowledge base systems I have seen so far are a collection of articles, each article with a title. Most of them allow you to categorise articles into groups and sub-groups. Users of the system can search for information using a title search, for example "How do I print from AwesomeProduct?" - which then shows a list of any articles that match that search text.

This is fine and dandy when your KB is for one version of the software product (the mythical AwesomeProduct ver 1.0). However, the development team then go ahead and create a new version (ver 2.0) that adds many new features and changes some existing features.

Now, how do we support both products in the same KB? The Naive method is to copy all articles from 1.0, and update them for 2.0, adding and removing articles in 2.0 as required. We can then add text at the top of every 1.0 article that says:

"this articles applies to 1.0 only, to see the 2.0 version, click here" (or something similar)

The problem with articles being indexed in the system by title is that it's very hard to filter based on meta-data like version. What happens when we create version 3.0 or 4.0?

The end-situation here is that you have a mess of articles. They're hard to search, hard to filter, and even harder to manage.

The solution (it seems to me) is to use tags, rather than text as the article index mechanism. So articles can be tagged with a tag representing the software version, topic area etc. etc. Users can then filter based on tag - an example search might be "version_1 printing" - which straight away gives a list of articles with all these tags.

So that's what I'm looking for - a KB system that uses tags, rather than text to index many articles. I'm sure I could build something with drupal, but I was hoping for something that worked out-of-the-box.

Thomi
  • 251
  • 1
  • 3
  • 9

6 Answers6

2

A tag based system is also known as a folksonomy (as opposed to a taxonomy). Clay Shirky has a great article entitled Ontology is Overrated. Currently there is some research being done with microformats. And there seems to be a little talk of it in the Drupal groups.

If all you're looking for is a loose ontology, why not just use drupal with the nodewords module? You can also just use wordpress with tags as well.

Tagging's easy, creating content is hard.

Joseph Kern
  • 9,809
  • 3
  • 31
  • 55
  • Tagging may be easy, but it seems finding a CMS specifically designed for tag-based knowledge bases is very very hard. While useful, none of the links you have provided are really build for a knowledge base system (which is a LOT more involved than a simple collection of pages with associated tags). I'm starting to think I may have to start looking for commercial software instead. – Thomi Aug 26 '09 at 13:34
  • I think it really depends on what you consider a "tag-based knowledge base". By that definition any CMS that has tags could be used in this manner. Can you be more specific? Perhaps in your main post ... – Joseph Kern Aug 26 '09 at 13:53
1

I think this may help... it is an open source CMS for documentation based on tags.

I myself have been looking for something like this.

sourceforge project page: http://sourceforge.net/projects/nexi/

project page below: http://nexi.com/33

Tell me what you think... hope this helps.

Jariell A. Perlman
  • 11
  • 1
1

For small knowledgebases, Tiddlywiki works well. It has tags and full-text search.

Bob
  • 940
  • 5
  • 7
  • ...Not really suitable for a knowledge base though is it? – Thomi Aug 26 '09 at 13:36
  • Well, we use it that way. I type in "scsi" and "linux" and I get all our internal troubleshooting articles for SCSI on Linux. I can access it from a webserver or a thumb drive, which is handy when I'm in the machine-room fixing a server which is firewalled off. Somebody was working on a TWiki->Tiddlywiki export, which would have been ideal for us. – Bob Aug 28 '09 at 12:17
0

I am right now also searching on something similar. seems like everyone is just sugesting wikis and blogs. But i see there is also a php for this: http://faq.phpmyfaq.de. as far as i see it has tags, search, related results...

0

http://www.kbpublisher.com

it is a paid version (but not very expensive - around $300 i think).

It uses tags, categories, article ids etc

I just did a experiment for you by putting a unique tag to one specific article on my KB and instantly with that tag I was able to pull up just that Article - so that works !

There is a 30 days or 21 days full functional trial available for download too, you would haev to use MySQL, ioncube loader, php etc to get it working. I use it on Windows 2008 IIS7 and works great.

Hope that Helps

Mutahir
  • 2,347
  • 2
  • 32
  • 42
0

From your description of how it is to be used I can't really see the need for tags. Instead you need a decent categorisation scheme and a decent search system. With that in mind I would say mediawiki would fit your needs very well.

John Gardeniers
  • 27,262
  • 12
  • 53
  • 108
  • possibly. The whole point of using tags is that multiple tags can be applied to a single article. I.e.- "version1" & "version2" can be applied to an article that describes a feature in both products. Using tags also stops the user having to guess the text phrase to use when searching for a specific article. – Thomi Aug 28 '09 at 07:01
  • What about blog software? You can have tags and full text searching as well? – John Gardeniers Aug 28 '09 at 21:48