Mark Baker points in todays link collection to an article by Nick Gall on the question if URLs need necessarily global scope. (What a sentence!)
Let's try again: are URLs the same for everyone? Stu Weibel asks the question, Nick Gall elaborates and Mark Baker agrees: they don't have to.
All those are really smart people. Which means I am probably wrong with my gut feeling. So, let's toss the idea around a bit and see what comes out of it.
Location dependant representations sound absurd at first glance since we are used to talk about urls like http://www.w3.org/TR/2006/REC-xml-20060816 and definitely, positively like the fact that it is the same for everyone. What happens when we give up that assumption in regard to URLs? I don't mean that a single web app somewhere in Absurdistan does it, but what if a good chunk of the web uses them like that? If it becomes an architectural style?
Somehow I am reminded of the tower of babel story. It starts with: Now the entire earth was of one language and uniform words.
And further: And The LORD said, "Lo! [they are] one people, and they all have one language, and this is what they have commenced to do. Now, will it not be withheld from them, all that they have planned to do? 7 Come, let us descend and confuse their language, so that one will not understand the language of his companion".
Myself being an agnostic, I do not draw truth from bible citations. But the author of that babel story did clearly see the potential of a universal language. So, if URLs lose their global scope, the language of the web becomes a soft-of context-sensitive one. And would that be bad?
Well, for one, it would be tricky for crawlers for sure. Depending on where the crawler is located in relation to the user doing the query later on, the result may make sense or not. That does not sound very attractive.
On the other hand, context-sensitive URLs are there. The interesting flickr photos from the last seven days is an obvious example. And it is a damn useful URL indeed.
Or take google. The query for url
has its own url, http://www.google.com/search?q=url, and we surely would be disappointed if it always gave the same result. Instead we expect it to list the best resources about url
at the time we call it. That is what google does after all.
So, the discussion boils down to the question if URLs should be sensitive about anything else besides time. Or, to narrow it down: if it's a good idea to make them sensitive about location. I don't know. I am certain however, that the opposite, namely only context-free URLs, would make a very boring web.
Technorati Tags: context-free, context-sensitive, url