URL/URI Redux

The question of how to construct the best URL made the rounds recently with a concentration on time sensitive publishing, eg. blogs and newspapers, with suggestions on how to make them “Future-proof.” I posted some resources that suggested best practices previously. Here are some additional ones:

• Common HTTP Implementation Problems
  This mentions one of my pet peeves, “[w]hen you change URIs, use standard redirects” ie. don’t do it with javascript or a meta refresh tag in a file at the old location. That just disables the back button.
• Article URLs with an emphasis on news sites
• Friendly, Lasting URLs is a good collection of links.
• Clean url’s uses an Apache rewrite rule to get rid of file extensions
• How to Future-proof URLs in MT advocates posting date and time
• Cruft-free URLs in MT uses the keyword field to enter a short title

There's tension here between making them unique and unlikely to change vs. making them semantically meaningful. It all goes back to the fact that URL's weren't really meant to be part of the interface, but they are. So, it seems most advocate using a short title (or slug¹) and/or date and time to attempt both.

Keep in mind that just setting a one to one mapping of the title to the file or directory name isn't a good idea. Such mappings can result in some rather long url's. Even setting whatever you use to publish to truncate after a certain number of characters isn't satisfactory since that results in weird, non-intuitive, chopped words. It works better if the concept of a short title or slug is used. Kind of like what we use at A List Apart. For example, Cross Browser Variable Opacity with PNG has the slug "png opacity" in its URL. We, however, do that by hand. I use movable type here. I've thought about using the keywords field for a short title entry (similar to what Mark did) but that would prevent me from using it for, well, keywords. Oh what to do?

1.Fun fact. From the OED, slug "Journalism, an identifying title, usu. occupying one slug." No, not the slimy animal you're picturing, it's printing press terminology. OED again, "1871 RINGWALT Amer. Encycl. Printing 416 Slugs. Pieces of metal of various lengths and thicknesses, but always thicker than leads, which they resemble in other respects...1888 JACOBI Printers' Vocab. 127 Slugs, numbered divisions of metal between different takes of copy.

The traditional definition of the term "slug" seems to indicate that a slug wasn't ever intended to be seen by readers, but instead it served as a catchphrase or place holder in the production process.

1925 G. M. HYDE Newspaper Editing (ed. 2) ii. 89 Use expressions that will not offend readers..if the slugs slip into print... 'Kill widow'..may be misunderstood.

But of course slugs used in URL's are intended for public consumption and not just part of the production process. Plus you might want to use "kill widow" for the shock value.