Cross Dressing URLs

When developing geospike.com (with my pretty small dev team) it always amuses me when I implement thing that to me are no-brainers, but that the big guys have not yet done.

Take this tweet (randomly selected) URL.
http://twitter.com/#!/Mr_Lexington/status/79427291343630337

Now have some fun with it:
http://twitter.com/#!/hello_kitty/status/79427291343630337
http://twitter.com/#!/you_suck/status/79427291343630337

The urls work! Of course you could make the URL quite obscene if you wanted to, I’ll leave this as an exercise for the reader. They don’t even redirect (meaning it’s entirely possible for such URLs to be indexed in Google!).

Or take this news.com.au story:

http://www.news.com.au/technology/alaska-publishes-palin-email/story-e6frfro0-1226073420363

and tweak the text a bit:
http://www.news.com.au/technology/sarah-palin-is-elected-president-of-the-united-states/story-e6frfro0-1226073420363

OK, at least news.com.au 301 redirects to the real story, so Google won’t cache it. But still – is it good these URLs work at all? What valid reason is there for them to work? [and no, I won't accept "my sysadmin/developer is lazy" as a valid reason]

Had some fun?

Now try a geospike one:
http://geospike.com/divzero/spike/27435

tweak it a bit:
http://geospike.com/eat-poo/spike/27435

oh, it doesn’t work. geospike++. different case versions of the username *do* work (well they 301 redirect), as to any past usernames.

Dev’s note: I fixed this today (had you tried this a few hours ago, the geospike.com link would have worked too). I can certainly see why the twitter and news.com.au (and no doubt countless others) exhibit this behaviour. The reason they work is that the URL already has a unique ID that is used to look up the content, and the rest is purely dressing. I think the dressing is good. It makes the URL pretty, and in News.com.au’s case may help with SEO. I just think the dressing should be valid.