An addressable knowledge-base

There has been a lot of discussion about Wikipedia of late. Much of this has been about 'authority'. There is another major issue at play here as well, which is really quite interesting for libraries.

Let me step back a moment to talk about the URL. The URL is the currency of the web. For something to be referenced, to be talked about, to be shared in the web environment, it needs to have a URL. Those things that are referencable are more likely to enter the web conversation; those that are not referencable in this way are off-web and much less visible.

Wikipedia is an addressable knowledge-base. It allows me to incorporate additional 'knowledge' in my communications by simply including a URL. The economy and convenience of doing this is enormous, and it is only possible because the resource is on-web.

Comments: 0

Mar 30, 2006
Justin Pitts

How are libraries ever going to bridge that gap in the face of copyright?

Mar 31, 2006
Jerome McDonough

An important point. One of my concerns in the Wikipedia context is that the URL into the knowledge base doesn't have a stable referent. The content the URL addresses today might be different than the content it addresses tomorrow. This obviously isn't a wikipedia specific issue, it's one shared by the whole web, but to the extent we increasingly rely on those knowledge bases, I think the stability of what's being addressed is going to be of more concern.

I've actually been thinking about this more in terms of how to do citation in papers which are discussing web resources, and I'm leaning more and more towards always including two URLs for a something I'm referencing: the original URL where I actually obtained the content, and a URL into the Internet Archive's copy which was archived on a date as close as possible to the date on which I obtained the original. URLs into the Archive not only are likely to stay 'live' longer, but I can rely on the content's fixity in a way that I can't for any resource on the general web.

Apr 01, 2006
Justin Pitts

Many wikis posses the ability to version their page history over time. For example, The wikipedia page for the OCLC http://en.wikipedia.org/wiki/OCLC has a long history of edits at http://en.wikipedia.org/w/index.php?title=OCLC&action=history
This makes it possible, though not easy with current implementations, to provide a stable referent (why? because wikis wisely choose to prefer pretty, easy to understand URL formats over opaque numeric IDs). When we wish to provide a stable referent....but, do we always wish to? Sometimes. Often, maybe. But not always. I suspect sometimes we want to refer to the most current version of a resource. I wonder if HTTP has a mechanism for this. I'll go read the docs, but I bet the sematics are there.

Apr 02, 2006
Jerome McDonough

You've homed in on the key aspect of the problem:

> When we wish to provide a stable

> referent....but, do we always wish to?

> Sometimes. Often, maybe. But not always.

> I suspect sometimes we want to refer to

> the most current version of a resource.

The stability of a referent of a URL is something that's going to be decided in the particular context of those running a website. Some will want stability, others will not, and it will be *very* difficult for anyone not running the website to have a firm grasp on what the 'stability policy' governing a particular URL is. You'll almost certainly see various policies in place for one site (and I'm interpreting policy broadly enough here to include 'barely formulated hazy notion of what I ought be doing'). Policies for a site may also change over time. Which is why I'm leaning towards using the URLs for the Archive, because they are explicitly dedicated to a policy of stability and appear to have the organizational wherewithal to support that policy. I'm not 100% confident their URLs have stable referents either, but I have more confidence in them than I have in 99.44% of the web.