Probably you've heard of Greasemonkey, the Firefox browser extension that makes it easy for people to write little bits of code that transform web sites. What? Greasemonkey makes it possible for users to fix badly designed web pages. Without Greasemonkey, if a site's broken, the author needs to fix it. But usually, the author doesn't think the site is broken (or isn't willing to fix it).
So far, Greasemonkey scripts are being used to filter out junk from various web sites. Ads are usually the first thing to go, followed by all the junk formatting that gets between the user and the information he wanted. These scripts seem to be straightforward to write, and another 10 or so are popping up every day. What they all have in common is that they try to make the browsing experience better for users, and more often than not, they succeed.
Some people think this is bad. Hell, for every idea, there's someone who disagrees, even for unmitigated goods like donuts, puppies, and Radiohead. The outcome that these people seem to worry about is a bunch of scripts interacting badly, or one script going crazy after a site redesign, and rendering a page unreadable. In other words, they're worried about bugs. Bugs. The latest argument is based on HTML being a poor foundation for these kinds of hacks. As the current crop of scripts demonstrate, HTML works well enough. If the hacks persist, maybe that'll encourage publishers to move to something better to reduce confusion. Meanwhile, Microsoft's Scoble, whom I ordinarily think of as reasonable, was anti-script, then decided it was ok provided the script conformed to his own extensive guidlines. In other words, he hates Greasemonkey, but he can't come up with a convincing argument against it.
I think they're really worried that the publishers are going to pack up and go home. If I can reformat CNN's pages to be 80% content and 20% ads (rather than the other way around, as they currently are), maybe CNN is going to stop giving away its content. It's probably way too late for CNN to give up on the web. What it might do is make publishers realize that users aren't going to stand for the ad-laden, "user-hostile" designs that are so popular today. Publishers will lose control of their content. The question is only whether the users taking over will do so out of love or anger.
As always, this isn't going to transform the world overnight, and the process is going to be painful. We'll know Greasemonkey is going mainstream when publishers start attempting counter-measures. Possibly they'll try to obfuscate their HTML in an attempt to confuse the reformatting scripts. It'll be an interesting fight to watch.
In the meantime, here's a collection of Greasemonkey scripts. Enjoy!
Update, March 21, 2005: Glassdog has covered the greasemonkey/autolink issue optimally.
Also, Scoble left a comment (see below) saying that he doesn't hate Greasemonkey. Maybe "hate" is the wrong word, but I think there's at least a whole lot of ambivalence there. Scoble's guidelines for proper use of linking technologies are here and then here. By my reading, a browser that ships with the default behavior of blocking popups violates Scoble's first, second, third, and sixth guidelines (there are 8 in total). But blocking popups is exactly what I'd want the default behavior to be, for me or grandma. Admittedly, I'm not sure I'm applying the guidelines properly. But the impression I'm left with is that they're an obstacle to widespread adoption of greasemonkey. You want grandma to download a plugin pack and click through a bunch of screens to enable/disable various features? Forget it.
This whole situation is starting to remind me of the debate about whether to accept or reject non-well-formed XML. Eww.
I buy 
