Microcontent and Purple
August 08, 2005
I originally wrote this as a comment at a Microcontent Musings post on PurpleNumbers but thought it should go here too (with some slight edits). I felt perhaps the point of purple numbers hadn't quite hit home. (PU7)
PurpleNumbers provide granular addressability to pieces of content smaller than a document. Originally the idea was to be able to create references into a page so the pointing one does when making reference to something is more complete. For example I can say "Yeah, take a look at the third item on that list after the third header on this page" or I can say "Look here" (unfortunately there a bit of an error in my css that scrolls that chunk up under the header instead of putting it in view, I should fix that). That's perhaps a micro optimization, but it is one that adds up immensely over the history of one's browsing by reducing ambiguity. (PU8)
More interestingly, though, is that once we got PurpleNumbers working, we saw that they could be used to give the chunks identifiers that raised them to the level of first class elements on the web: they could have their own URI and the purple numbered content could therefore be dynamically transcluded into other documents or used to compose new documents. For a pretty old example see TransClusion. (PU9)
Arnaud says: (PUA)
Each paragraph, each item in a list, might thus have it's own permalink. The only purpose that I can think of, is when you want to quote some text from someone else. But the question is whether the quoter and quotee have the same idea of quotable pieces. I doubt that. (PUB)
Indeed there is a big difference between what the quoter and quotee might think of as a quotable (or referenceable or transcludable) piece. PurpleNumbers currently take the political stance that the quoter should be able to quote whatever they want: every chunk gets a purple number. Some similar tools let the author decide what chunks get numbers. Purple numbers choose the more anti-authority approach. They also try to avoid being ridiculous. (PUC)
Where the chunks gain meaning is in the way they can be reused in other documents. Knowledge creation is a dialectic of sorts. Reusing existing content to create new content springboards new understandings without reinventing the wheel. (PUD)
Comments
In the past systems I've developed where chunks are addressable, I've gone so far as to try to make every word and punctuation mark individually addressable such that any combination of phrases could be quoted. (PUG)
Turns out, block level elements are a much better building block for quotes and other uses of "microcontent". Better in terms of being managable, both by the author/editor, and by the quoter/reader. (PUH)
So, the solution I've worked out is providing join and extract functions for block level elements, e.g., join these blocks as a set and then quote everything within it after the first ":". (PUI)
For a URI-based approach, the join and extract functions can be web services that take the chunks URIs as part of their arguments. (PUJ)