Emerging structure from tags on the DAR wiki.
I decided to spend a little time systematically tagging some more of the pages in the DAR wiki.
The mediawiki software has a facility to extract a list of all uncategorised pages automatically, and this is accessed from Special:Specialpages and then Special:Uncategorizedpages.
So I went through the list and added one or more tags (called categories in mediawiki) to each. For the sake of consistency, I kept another browser tab open with the list of all categories Special:Categories to refer to.
At the end of this process I had about 40 categories for the 70+ pages of content, and that’s far too many categories to list on the front page in the manner which has evolved up until now. I wondered if I could group clusters of them together into much broader categories, perhaps ending up with just a handful of top level tags. At this point I remembered noticing another special page endearingly named – Uncategorized categories. This suggested that category pages themselves can have category tags added to them and that this is also cross referenced somewhere in the database so worth a try. I added the category “Projects” to the category pages for both “drupal” and “Barn Raising” lo and behold a protopage for category Projects can be accessed which states
Category:Projects
Subcategories
There are 2 subcategories to this category.
B
* Barn raising
D
* Drupal
So I had instantly created a hierarchical structure with three levels. Main category, sub-category, and individual pages. And all without having to declare any categories or structure in advance. The meaning can emerge out of the development of the resource itself rather than having to be pre-defined with all teh constraints that imposes. This is a very powerful tool indeed, highly flexible and emergent, and particularly rewarding of the emergent design methodology which I ventured to adopt in the first place, although it did take nearly 12 months to reach this stage, nearing completion of one of the long cycles begun in my final year at Ultraversity.
To explain in more detail what can happen here:
New pages of content can be created in isolation, without any reference to how they might fit in with the rest of the dynamic resource hosted on a mediawiki.
At the point of creation or any time later, pages can be tagged with category descriptions just by inserting the code similar to the following at the bottom of the page:
[[Category: COPs]] [[Category: Links]]
The software then automatically picks up the creation of a new category code or the addition to an existing one. Category pages are special pages which list all of the pages to which the code has currently been applied. Category pages can themselves be tagged by the same method and this is also recognised by the software which can then list all of the sub-categories hanging off that tag.
Thus one of the recommended methods for emergent coding analysis of data is perfectly facilitated by this type of system, with the proviso that the data being analysed is loosely structured around a concept of pages. It probably wouldn’t work with lists or conversations. The amazing thing to me is that there doesn’t seem to be any conflict or tension between the loosely organised, unstructured, emergent method of collecting and developing the data and ending up with a highly organised strictly hierachical taxonomical index.
I then added the tag “Main category” to the page for “Projects” just to round things off, and all I have left is to work my way through the uncategorised categories from time to time and perhaps merge any duplicate concepts with similar names.
Thanks for subscribing to Andy Roberts blog
Emergent coding analysis facilitated by mediawiki