1. No graceful degradation. Par for the course with many web apps these days, but it’s an url-shortener! Two textfields and a submit -button- link require javascript. Welcome to the wonderful world of web 2.0.
2. Speaking of which, the input form is not even that. Come on, it’s not like you can’t make ajax calls with forms.
3. No url validation. Try shortening the url “learn2validate”, and you get a helpful alias to http://learn2validate.
4. Breaks Firefox autocomplete. Because they aren’t using a form and triggering the submit on keycode 13, if you type “ht” in the textfield, scroll down in the history to select your url of choice, and hit enter to choose it, what’s going to happen is that you’ll end up submitting the url “http://ht”.
5. Does not respect cache headers. Jason Levine has a detailed breakdown here
6. Their bookmarklet is not one-click (which disqualifies it as a canonical bookmarklet imo). The bookmarklet simply takes you to the their home page and helpfully fills in the url field. Compare that to TinyUrl’s which creates the url for you with that one click.
I really don’t understand all the hype over Bit.ly. A glowing review from ReadWriteWeb, and people on twitter generally seem to like it. I guess all the shiny stuff distracts from the gaping holes in core functionality. At times like this I really wish uncov was still around to smack them down.
Update
Looking through their page source reveals a ton of WTFs. They have both inline css and a separate stylesheet (I can understand arguments for optimization with one or the other, but both?). The same goes for javascript—directly included in the page as well as in a separate file. Alternating javascript and css inside head. More javascript in body. Unnecessary use of tables (check the source for the footer). It’s embarrassing.
