The social web is largely based on the principle of “If you can’t say something nice, don’t say anything at all.” You can ‘friend’ and ‘follow’ people, but you can’t take some action to express the idea that person X is a useless waste of protoplasm who is consuming oxygen that could be better used by bacteria. (You can ‘unfriend’ and ‘unfollow’ them, but that just restores the status quo, instead of registering a negative). You can ‘bookmark’ or ‘favorite’ web sites, but there is no corresponding way to mark them publicly as worthless crap. And while you can sometimes vote things up or down, a down vote is typically an after-the-fact reaction to someone who thought it was worth voting up in the first place.

I’m aware of the reasons for this asymmetry. An “anti-social web” composed only of negative judgments would probably be a nonstarter, or at best a playground for extortionists (“Nice little website you’ve got there. It’d be a shame if four thousand people decided to vote it down.”) Still, there are times when you come across something so repellent that you long for some way to punish it.

Today’s candidates for swift and righteous execution are the companies that offer those in-text advertising links. I’d been vaguely aware of the issue before, but I’d never realized quite how loathsome it actually is.

Picture this; you visit a ‘tech’ blog that has an article summarizing a report from a security vendor. The article is short on detail, but the name of the vendor and the word ‘report’ are underlined and shown in a different color. Conditioned by your experience of the conventions of the web, you make the mistake of assuming that these are hyperlinks that will take you to a source of more detailed information.

But wait - the words have a double underline, signifying - in someone’s mind - that these are not hyperlinks, but something else. In fact, they are ads, pure and simple. If you mouse over the name of the security vendor, you get a popup that displays … information about a paper shredding company in New Jersey. Same thing if you mouse over the word ‘report’. If you click on the links, you get taken to the paper shredding website.

The word ‘India’ is also underlined. That gets you an ad for an Indian restaurant. And if you go back and mouse over the other links again, you get a completely different set of ads, most bearing only the most limited relation to the highlighted word.

The whole model is built on breaking one of the most fundamental assumptions of the web: that a marked hyperlink will link to more information relevant to the highlighted words. It is an absolute anti-pattern. It clutters the page and tries to deceive the user.

“Ah, but”, the inventors of this fetid little scheme would probably say, “We are establishing a new convention. Users will quickly learn that a single underline denotes a hyperlink, and a double underline denotes a worthless piece of shit. There is no danger of confusion.” And they may be right. Users will learn, or at least the smart ones will. The dumb ones will keep clicking the links and, who knows, may even do business with a company that has already demonstrated its contempt for them by deceiving them. But all of us will have to deal with the extra load of having to figure out which links are useful and which are time-wasters. And that whole ‘double underline’ convention? I’d be willing to take a small bet that when users start to wise up and the click-thru rate falls off, the ‘convention’ gets quietly dropped.

I don’t run ad blockers. I accept that websites have to cover their costs, and if that means that I have to endure a certain amount of advertising cruft, so be it. The ads that zoom out and cover the text you were trying to read or that suddenly zoom open and cause the rest of the content on the page to move out the way leave me feeling a bit abused, but I’ve no objection to a discrete column of well-chosen text ads. Everyone has to eat.

But an ad scheme that deliberately breaks the implicit social contract between publisher and reader seems to fall into a different category. ‘Intent to deceive’ - and I find it hard to call it anything else - strikes me as a significant breach of trust. I think users would be within their rights to open up the source of the page, find where the Javascript library that injects the links is loaded from, and block that sucker.

Or, more simply, to make a note of the so-called websites that use this scheme and resolve never to go back.