Entropy and Information - the Deletion Problem

Tor Norretranders' excellent book The User Illusion is about, more or less, information technology, consciousness, and reality. In other words, it's broad.

I've been thinking about this book's information-theory driven claims that the problem with entropy and information is not gathering data, but throwing it away... this relates strongly to blogging in general, and our enterprise blogging idea.

The "publish" model of both ERP and traditional news does a good job of circumscribing what's true/real from the vast majority of maybe, early drafts, variants and versions, unproven, disproven, etc. While we chafe at its limitations, we get a lot of benefit from the assurance that the limited dataset within it is highly trustworthy.

Once something enters the blogsphere, it never leaves... even if disproven or expired due to time marching on... Is the best way to measure the truth/value of something the ongoing, continued active linking by people back to "old items" of real value, vs. the gradual decay of the "link-liveliness" of items which are either wrong, or increasingly irrelevant over time?

Studies of academic footnotes have (on a much longer timescale) done a good job of identifying "canon" texts vs. the vast majority of useless research (most published items are never cited by anyone else) plus the whole continuum in between. It should be possible to build a similar hierarchy of trust/truth over time, both across multiple data sources (mx. blogs) and within a single blog (e.g. I refer/link regularly back to the "good stuff" I have written, because it's relevant to my current writing, but I tend not to refer back to the less good stuff).