Page:Aaron Swartz s A Programmable Web An Unfinished Work.pdf/23

'''2. BUILDING FOR USERS: DESIGNING URLS    11''' found interesting. (The original browser didn’t have a URL bar, in part to force you to keep track of pages this way.)

It was a brilliant idea, but unfortunately it was written for the obscure NeXT operating system (which later became Mac OS X) and as a result few have ever gotten to use it. Instead, they used the clone created by a team at the University of Illinois Urbana-Champaign (UIUC), which never supported editing because programmer Marc Andreesen was too dumb to ﬁgure out how to do page editing with inline pictures, something Tim Berners-Lee’s version had no problem with. Marc Andreesen made half a billion dollars as UIUC’s browser became Netscape while Berners-Lee continued doing technical support for a team of physicists in Switzerland. (He later became a Research Scientist at MIT.)

Image: http://www.w3.org/History/1994/WWW/Journals/CACM/screensnap2_24c.gif

The result is that we’re only reacquiring these marvelous features a couple decades later, through things like weblogs and Wikipedia. And even then, they’re far more limited than the wide-reaching interactivity that Berners-Lee imagined.

But let’s turn away from the past and back to the future. Sir Tim argued that to protect your URLs into the future, you needed to follow some basic principles. In his 1998 statement “Cool URIs don’t change”, described as “an attempt to redirect the energy behind the quest for coolness... toward usefulness [and] longevity,” he laid them out:

However, I go on to disagree with Tim’s proposed solution for generating Cool URIs. He recommends thoroughly date-based schemes, like “http://www.w3.org/1998/12/01/chairs”. As far as I’ve noticed, only the W3C has really thoroughly adopted this strategy and when I’ve tried it, it’s only led to ugliness and confusion.

(You may notice that Tim says URI, while I say URL. URL, the original term, stands for Uniform Resource Locator. It was developed, along with the Web, to provide a consistent way for referring to web pages and other Internet resources. Since then, however, it has been expanded to provide a way for referring to all sorts of things, many of which are not web pages, and some of which cannot even be “located” in any automated sense (e.g., abstract concepts like “Time magazine”).