# How to recognize a Weblog tool by its permalinks

This list is inspired by Julian Missig, who thought I was using Movable Type on my own Weblog. (I’m using Manila.)

Cruft in each URI is emphasized. Cruft is any part which is an artifact of the Weblogging software, meaningless to most humans, while decreasing usability by making the URI unnecessarily long.

  • http://foo/index.php?p=501&c=1#comments
  • http://foo/arc20030718.htm#BlogID1234
  • http://foo/2003_07_18_archive.xyz#167509668584
  • http://foo/category_name/entry_title.html
  • http://foo/archives/000000002.php
  • http://foo/users/henry/1234.html
  • http://henry.foo/1234.html
Manila or Radio UserLand
  • http://foo/2003/07/26.html#a1234
  • http://foo/discuss/msgReader$1234
  • http://foo/storyReader$1234
Movable Type
  • http://foo/bar/000012.html
  • http://foo/2003/07/title_of_entry.html
  • http://foo/2003/07/26/title_of_entry.html
  • http://foo/cgi-bin/mt/mt-comments.cgi?entry_id=2
  • http://foo/comments.php?id=P1411_0_1_0
The Ultimate Weblogging System™
  • http://foo/2003/07/26/keyword

Additions and corrections are welcome. (Whining “But it’s customizable!” is not.)

I should point out that removing all the cruft is not the only way to improve the usability of a URI. There are a few other common problems with Weblog URIs.

  • Don’t use in-page anchors (URIs ending in #whatever) as permalinks, like Blogger, Manila, and Radio UserLand do. In other words, don’t put multiple entries on the same archive page. If you do, server logs can’t show you the popularity of individual entries, you can’t set up good redirects for each entry if you migrate to another system, and your site will annoy search engine users when their search terms are spread across unrelated entries on the same page. Instead, put each item on its own page.

  • Don’t use the ? character, like b2 and pMachine do. ? should be avoided for any persistent resource, one whose content will remain much the same from month to month — like a Weblog entry, for example. Unless a resource with ? in its URI is heavily linked-to, search engines will avoid it, because it’s likely to be just one node in a very large (or infinite) nonsense site. This is why, for example, Brian Tiemann’s Weblog is hardly indexed by Google.

  • Don’t use the _ (underscore) character, like Blogger and Movable Type do. Many computer users have trouble finding _ on their keyboards. And even for those who can, it’s harder to type than - (dash) is. The same applies to BiCapitalization: capital letters are harder to type than dashes are.

… 2003-07-31: Thanks to Alwin Hawkins, for supplying the pMachine example; to Bacchus at ErosBlog, for telling me about Blog; and to Már Örlyggson, for telling me about b2, for letting me know that the a in Manila’s #a1234 isn’t cruft (it’s actually required by XHTML), and for pointing to his article on how to decruftify Movable Type.

Posted on 7/26/03; 3:06:13 PM