Mixins and Tags
http://search.cpan.org/dist/mixin/
This is a rave: bear with me.
Mixins are cool. They seem to make sense to me in the way labels/tags/categories make sense, and aspect-oriented programming makes sense. There seem to be many cases wherein a certain piece of functionality is needed for many parts of an application, and subclassing this functionality doesn’t necessarily make sense. For instance, logging. Everything should be able to log. It probably doesn’t belong in the core of a programming language, but every part of your application might need access to logging functionality. Mix it in.
By “tags” I’m referring to something like what’s used on del.icio.us. Any link can have any number of tags. This idea makes sense to me. A link about scuba diving in London should have tags like scuba, diving, and London. Actually, del.icio.us has tag intersections to help one find exactly what one seeks. Flickr uses a similar tagging system as part of their image classification scheme. Apparently a lot of time and energy is devoted to studying tagging.
These ideas (mixins and tags) seem fundamentally related. You give something meaning and/or functionality without changing it’s ancestry. Take humans, for instance. We have ancestry that doesn’t change, but our life choices make up our “tag cloud”, each thing we associate with becomes a more or less temporary tag.
One place where I’d really like to see tagging is in financial software. It makes zero sense to me that the $10 spent at Joe’s Diner must fit into one category. I want to give it a bunch of labels, like: entertainment, food, dinner, and local. In this way you could “mix in” the parts of life where a transaction fits in.
Something to think about, anyway.