The problem with labels

A couple of forays worth into the application of labels (as Blogger refers to tags) and I decided I better move the cloud display further down the sidebar because it was clearly going to rapidly get out of hand.

The problem as always with labels is that there are so very many things, and the map that could supply every detail of the terrain would have to be as big as the territory.  So you end up with a handful of drastically recurring objects because they happen to be so broadly defined or general followed by a massive scrum of overly specific one-offs.  I guess time will tell.  It's less of an issue of course in a case like this where the labeling is pretty much just a parallel work of art.  When it is supposed to be functional these problems represent real obstacles.  In the real world of say biological taxonomy actual disputes over where lines are drawn, or whether a particular label represents over-specificity is a regular occurrence.  Classification is useful but on some level always artificial.

When you start to weight stuff new issues arise.  The meta tag I'm using here is a good example.  It's simple function is to label the non-song posts - the posts about the blog itself (not to be mistaken with the solipsism tag, which is used for when I refer to the song of the day project within the song itself, an occurrence I fear I will find to be all too common).  "Meta" posts are by a very great margin the minority on this blog, but because the tag is unambiguous and consistently applied it already carries a rather high weight in the tag cloud.  Somewhere out there there is a whole series based on Dante's Inferno - and it's horrible! Unreadable!  But for that sole exercise I expect the hell tag is always going to loom large in the cloud once I reread (ugh) and label that series.  Oh well, I guess it is what it is.  I can just be glad I don't have contextual advertisements in it anymore, the heavy weight of the hell and religion tags would probably attract an unsavory element.

I'm hoping getting to a more representative sample of the songs labeled will see legitimate common themes arising in the songs will play this down by and by - in many of the seemingly obscure labels I've come up with so far I'm anticipating motifs, themes and personal jargon I know will arise repeatedly as I continue through the lyrics.  The problems of redundancy and over-specificity are harder to know what to do with (at this point I'd invite certain of my compatriots to enjoy the opportunity to suggest the redundancy of having separate tags for religion, mythology, and superstition...)

So the evolving tag cloud is going to be an ongoing element of interest to me, at least...

what