No index.php
On a mailing list I’m on someone just pointed to no-www.org. It’s been awhile since I’ve run across that site (or, before it existed, Slashdot commenters condemning use of TCWWW — The Cursed WWW), but I strongly agree — www. in a domain name is pointless.
Even worse is index.php in the path. You’ve taken the time to publish a website, now take a few minutes to make its URLs less ugly. I’m not going to bother setting up no-index-php.org, but someone should. However, in the spirit of no-www.org, here are a couple resources for removing index.php from popular software installations:
- WordPress Codex on using permalinks — in many cases WordPress will produce the right
.htaccessfile automatically, but if you have to fiddle, the default behavior of identifying a post in the querystring with an integer (e.g.,?p=1913) has a strong appeal, while the “almost pretty” option is extremely ugly. - MediaWiki Manual:Short URL — when this page lived on meta.wikimedia.org, it was appropriated called Eliminating index.php from the url.
Please remove index.php from your URLs, or signal that you have no taste, no technical abilities, or both.
Thanks!
May 20th, 2008 at 14:44
I was long in the no-www camp.
But then via the Yahoo ‘Best Practices for Speeding Up Your Web Site’, I was reminded that if your site’s cookies are set on domain.com, they will be sent to every subdomain.domain.com, too. That presents problems for caching, even if you segment all static resources off to a ’static.domain.com’ subdomain.
Solution: either set up a different owned domain like domainstatic.com for cacheable resources, or embrace the ‘www.’ and set your application cookies only on ‘www.domain.com’.
So the www is not all bad.
See: http://developer.yahoo.com/performance/rules.html#cookie_free
(Point taken about index.php, of course. Clean URLs are one of the dozens of little things that send a signal of craftsmanship to a savvy audience. Others that spring to mind: ALT/TITLEs for images; making radio/checkbox labels active; intuitive tab orders; layouts that work at different font sizes; form submits that redirect to non-POST reloadable pages.)
May 20th, 2008 at 14:49
Also: for a site that has (or may have) different language versions, the wikipedia per-language subdomain approach is nice. It offers a subdomain on which to hang the login cookie — such as en.wikipedia.org — that’s more meaningful than ‘www.’.
May 20th, 2008 at 16:08
Good point re caching. It seems that Yahoo takes the domainstatic.com approach — yimg.com — though yahoo.com redirects to http://www.yahoo.com.
I do like meaningful subdomains where appropriate.
May 21st, 2008 at 2:30
‘www’ could have been selected to cue the speaker to greater precision in their speech and the listener to greater concentration in correctly interpreting the following identifier. A sort of line conditioner like the ten pulses of the zero that starts all telephone numbers (well, it worked like that once).
It also serves as a good audio marker for any speech analysis software.
Ok, so TBL just thought it looked cute - and then made up some lame excuse about it being memorable and standing for ‘world wide web’.