When the web was infantile, there was no real way to know what was new. During its toddler stage, information about new websites was passed around via email or on USENET. Eventually, there became too many sites being created for this to be a reasonable way to keep up. (A similar issue plagued the single /etc/hosts file, but that’s a different story.)
Finally, there came Yahoo. Now, there was a directory for the web. You wanted to know about the NRAO? You went to Science, then selected Astronomy then Radio Astronomy, finally Observatories, and there you found the link.
But any hierarchy has certain troubles. What if you didn’t know what the RA in NRAO stood for. You get to Astronomy but there’s no Observatories sub-category directly under Astronomy! How well do certain new things fit in the existing structure? What if things reasonably belong to more than one category? If I found something once, how do I find it again? So, Yahoo added site searching capability. Now, you could search (within yahoo) to find your site. Since yahoo was relatively definitive, this still worked pretty well. But finally even link pages within yahoo became huge. Not to mention the administrative nightmare of keeping a huge directory up to date with sites coming and going at an ever increasing rate.
But google had an answer to that… Don’t keep a hierarchy. Just search the whole thing. Search the whole web. To be fair, I don’t know that Google initiated this approach. Only that they were the first ones I knew about to do it effectively.
Then came the googling of email: GMail (If anyone needs gmail invites, leave a comment. I’ll set you up.) You no longer need to sort your mail into a hierarchy of folders. Just leave it in a big pile and search for what you need.
One thing you may notice about both of these, is the presence of meta-data. Data about the data. Web pages have indicators labeling the title of the page, many have explicit meta-data in the form of keywords and descriptions, and of course there’s the (in)famous page rank indicating a page’s popularity. Email has similar meta-data: To:, From:, Date:, and Subject: all right there for easy computer analysis.
And now, with del.icio.us comes the end of the bookmark hierarchy. But with bookmarks, there’s much less natural meta-data. Bookmarks have (at beast), the link itself, the title of the link (usually pre-defined as the title of the linked page), and the date that the bookmark was added. So, del.icio.us lets you add your own meta data in the form of “tags.” You tag a bookmark with keywords to aid in your finding them at some later date. And it treats bookmarks more like a stream, or a queue. Last in are first available.
Bookmarks have interested me for a while. Mine keep getting lost. Or they’re available on one machine, but not another. They are far more transient than I’d like. As a result, I had been occasionally working on an online bookmark tool. But, like my thinking, it had been hierarchical. You could make folders and bookmarks. Folders contained bookmarks and other folders. Now, I’m going to give delicious a try. It may be the evolution of the bookmark before I even managed to finish creating the interim step for myself. (Other versions did exist, but none impressive enough to make me want to use them.)