Monday, December 29, 2008

The joy of namespaces

I promised (threatened?) in this blog's introductory post to talk a little bit about why I needed Blogger to be able to publish under my domain, zigg.com.

In 1998, Tim Berners-Lee, the guy who invented the Web, wrote a little bit for the W3C called Cool URIs don't change, in which he advocates keeping resources available under the same URI forever in the face of the beginnings of a worrying trend: URLs that we used just a few years ago are increasingly being invalidated by the march of progress.

You've seen it yourself, if you've used the Web for any length of time, I'm sure. A link used to go to some page you're interested in, and it drops you off at a dead page—or worse, asks you to try searching their site for it—an exercise that is usually fruitless. It doesn't have to be this way; a conscientious and competent webmaster can either keep the same URL scheme forever or provide some server-side trickery so that URLs from many years ago will still work, with one key catch: they have to have full control over the domain that the original URLs lived on.

This is why I need my journal to live under zigg.com; if, down the road, I do quit using Blogger, I can continue to serve up all the content that I've created since I started transparently—even if I switch software and end up using a new URL scheme, I can have the old mapped to the new. If I hosted my blog elsewhere, though, I wouldn't have that option. I'd be at the mercy of whatever that host provided me. If I fully control the domain I'm using, I can do what I need to do, and if the people I'm paying to host zigg.com can't give me what I need, I can go elsewhere.

I'm glad I can do this for my blog, but it's troubling that so many other sites (including many that may simply not be here in a few years) are lining up to have me host resources without giving me any sort of control over the namespace itself. If it's a Facebook-like situation, where there's a walled garden of sorts and no expectation of external linking, the stakes are obviously much lower; but if I am going to be creating public resources, I need to be able to promise the public that I am committing to continuing to serve that resource, as long as it exists and I am able, at the address I originally give for it. Publishing things under anyone else's domain just doesn't give me that ability.

And to not be able to make that promise, well, that's just not cool.

Footnote: If you're confused by "URIs" and "URLs", you're not alone. The types of addresses most people have been exposed to on the Web, those that start with http:// and ftp:// and what not, those are URLs because they locate a resource. There are also things called URNs, which name a resource, but don't actually locate it on the network; something like urn:isbn:068486293X would be a URN. Both URLs and URNs are URIs. The more you know!


posted by zigg 9:00 PM


Further comments are disabled indefinitely. Here's why.