Categorization Schemes

We've been thinking of these as schemes in which each pattern fits in one, and only one category. Also see tags, which are more open.

Draft Primary Categorization Scheme February 2011:

The core group (Dave, Sue, Tree), after lengthy discussion, has converged on a draft Primary Categorization Scheme which consists of nine categories, and assigned each of the 82 patterns to one and only one 'primary' category, with one identified 'keystone' pattern (shown in bold) for each category. The categories each have between 5 and 12 patterns assigned to them. This scheme can be downloaded in two formats:

• Download Primary Categorization Scheme
  in Powerpoint (.PPT) format
• Download Primary Categorization Scheme PDF
  in .PDF format

Note that the colors and symbols on those charts are still being worked out.

To update these:

Process (by Tree):  When a few people first set out to categorize the patterns early on, they came back and said it was not yet manageable.  Some months later a group of us at a work session did the first "quick-&-dirty" categorization.  During 2010 several people tried doing their own, some solo and some joint, some at work sessions and some independently.  Eventually a document was created listing the schemes thus far side by side on a spreadsheet, so they could easily be compared.  Dave synthesized several of the schemes together into a proposal, which the core team then further refined, until finally reaching agreement.  This included considering and laying aside the possibility of listing patterns in more than one category.  The hardest part in my opinion was working through a disagreement between Dave and me about what it takes to create a thorough categorization scheme, particularly what is the role of group insight in that vs. one person sitting down alone to think it through.  In the end i'd like to think we got the best of both individual and group effort; at the same time i recognize that our result is subjective and necessarily imperfect. 

Additional Schemes:

John Abbe+Categorization schemes (close to a convergence now)

List of possible criteria that a categorization scheme should meet:

(note that these criteria apply to categorization schemes only, not to tagging schemes, which may be much more flexible)

1. The categories are mutually exclusive and collectively exhaustive i.e. each pattern reasonably fits in one and only one 'primary' category. (We are allowing some overlap at the moment.) An 'other' or catch-all category should not be used.
2. There is an intuitive coherence to the categories. If someone is reading a training manual or book about the pattern language that uses this categorization scheme, they should intutively appreciate why this categorization scheme was chosen, and it should help them learn about and organize in their own thoughts, the complete set of patterns.
3. There is an internal thematic consistency to the categories. It should be possible to say "this categorization scheme classifies the patterns by (quality x)".
4. For practical reasons (we don't want the chapters of the book and of training materials to be of wildly different lengths), the number of patterns in each category should be reasonably consistent. Ideally there should not be categories with, say, more than twice as many patterns in them as other categories. Unless there is some compelling reason, there should not be ca|tegories with only one or two patterns in them.