In the process of importing my delicious bookmarks into evernote, I discovered that around half my bookmarks - mostly the older ones - went to dead links.
The dead links were mostly (but not all) links to magazine and newspaper articles, and the links that persisted mostly to academic papers. No, I didn't keep notes and this is all impressionistic and anecdotal.
Now, all my professional life I've kept notes, notes of presentations, conference papers, and also newspaper clippings (initially), later on printouts of web pages, bookmarks, and now evernote notebooks.
Of all my notes, about 80% are never referred to again, and of the remaining 20%, around half of it is irrelevant after five years.
Guessing which 20% is going to prove some use and of that which is worth keeping is the difficult bit, especially as happenstance can be a wonderful thing. That presentation that you went to many years ago on a long dead product's peer to peer architecture and it's use of a low bandwidth gossiping algorithm suddenly assumes significance when talking about data replication and resilient cloud storage - not because of anything other than the idea picked up on wet Tuesday long ago.
This doesn't come as a surprise - when we moved to Australia from the UK we got rid off, in one way or another, about 75% of our books. We started out the culling process by trying to be rational, but ended up being pretty random in our approach. And while we've had to rebuy a few things, it hasn't worked out too badly. The key point being that while there might be music you want to listen to again or books you want to reread, or material to refer to, you actually don't really need to as far as most of the content is concerned. (One wonders how many people will end up deleting books out of their e-readers to make space, just as people delete pictures out of flickr to stay below the 200 image limit).
Now, personally, I'll continue to keep notes as it helps me remember things, just as writing blogs about stuff helps me clarify my thoughts. And I've always been a sucker for buying books.
But this does have implications for digital preservation. People rightly worry about the sustainability of content, particularly in the current economy. However we can approximate forever to be something between five and ten years. After that time most things are fairly irrelevant, and probably could be let die.
In fact, using my experience as a guide, only about 10% of anything is worth keeping for more than five years. The problem is which 10%.
And the real problem is that the 10% will vary according to audience. For example, J, as an artist, images, ideas, pictures of textures, information about exhibitions, is important.
For someone else it could be book reviews. For my father, as an engineer, it was specifications, engineering change notices, schematics and diagrams. We're all different. Equally, I have the suspicion if we were to genuinely, randomly, delete 80% of all online content after five or so years we'd probably not be too badly off ...