Short URLs and the Future of the Web

While Maverick Revolution does develop and maintain the urlShort open-source URL shortener (and also runs our own service using the code.), it is important to take the time to think about the problems with URL shortening, and what we can do in the future to combat these.

The chief problem is reliability. When the Cligs URL shortener fell victim to a hacking, over 2.2 million URLs were hijacked and redirected to another site. Those 2.2 million cli.gs URLs were the only point of entry to websites posted on Twitter, and other social networks, but suddenly they weren't going to the right place. 2.2 million may seem like a lot, but it really is miniscule compared to TinyURL, which more or less invented short URLs. TinyURL claims they have over 275 million URLs shortened, plus 2 billion redirects per month. Now, imagine if suddenly 275 million URLs disappeared from the internet. It would cripple both old and new content alike.

I'll be straight to the point on this: There is no solution. Aside from constant backup solutions, server redundancy, and an endless amount of cash and time to maintain the most popular of URL shorteners, there is no way to guarantee any of them will exist in 10 years, or less. There needs to become a standard for short URL services to follow, that includes redundancy and regular backup in order to be used by services like Twitter, and applications like TweetDeck and Seesmic Desktop. Otherwise, all we are doing by using these services and apps is enlarging the potential problems in the future.

Not only does it create a gaping hole in the infrastructure of the internet, it also destroys the transparency in URLs. Usually, if a URL hasn't been shortened you can tell what it is, or at least get the general idea of where you are going. For example http://mavrev.com/site/story/urlshort_120_whats_in_store looks to be on the site mavrev.com, looks to be a story, and also about something called urlshort 120. It's not perfect, but you can tell that you're going to a site that you (possibly) know, and that it's a story/article of some kind. But with http://u.mavrev.com/e5yp or http://tinyurl.com/makt6v, you have no idea where you're going, aside from the domain u.mavrev.com, or TinyURL.com.

Now, there is a partial-solution to the issue of transparency; Custom short URLs, or aliases. Rather than http://u.mavrev.com/e5yp or http://tinyurl.com/makt6v you can instead have http://u.mavrev.com/urlshort120 or http://tinyurl.com/urlshort120, which helps to give some information to the browser about where they are going. A better solution is having services like LongURL or Long URL Please offer Firefox extensions, GreaseMonkey scripts, Ubiquity commands, and APIs for replacing the short URL with the full URL. Of course, these require action on the part of the user - Going to the website, downloading and installing the extension or script.

The best solution would be for browser developers (Google, Mozilla, Apple, Microsoft, Opera) to include this functionality on their own as part of their phishing and malware protection (on by default). Of course, this still relies on the servers for these URL shorteners being available to offer the expanded URL, but it is, however, a good way to make up for the transparency lost by URL shorteners.

Similarly, services like Twitter, where short URL services are frequently used should also automatically expand the URLs for users, to show them where they are going when they click a link. That, or they should operate their own short URL services if they insist on counting URLs in the length of messages. Ping.fm shortens all URLs sent through their service with their own shortener, eliminating the middle-man that exists when you post through Twitter without Ping.fm. With coding teams that these services have, a simple short URL service wouldn't be hard to do. And to make it easier, they could also use an open-source project like urlShort to start with, and just limit it to their servers backend from access.

Many URL shortening services, such as Digg, Ow.ly, Adjix, and Su.pr (StumbleUpon), frame the content they "shorten" with their own toolbar. This is wrong. Not only does it ruin the point of URL shortening (and instead makes it into URL sharing), it also (essentially) steals the content from the creators. To the average person, a frame at the top of the page saying "Digg" makes it seem like all the content on the page is from Digg, or whatever name appears at the top. In addition, both Digg and Ow.ly display ads in their frames. They are making money off of other people's content, and possibly stealing revenue from the content creators, as the ads appear at the top of the pages and don't move. With the case of Adjix, many publishers use the short URL service for the frames, with ads, because they receive a portion of the revenues. However, there is no way to determine if the person owns the content they are shortening, which can lead to the same problem seen by the Diggbar and Ow.ly.

In addition to portraying a false sense of ownership, and also potentially stealing revenue, framing websites destroys link architecture. When you go to a link that is framed, and then click through to other links and decide to close the frame, you're taken back to the first page where you were, which was initially framed. Not only that, but it also masks, and therefor further destroys the transparency of URLs, the URL of the page you're on for your entire visit, or until you close or disable the frame.

Luckily, services like Digg and Ow.ly allow you to permanently disable frames with a cookie. Digg, actually, goes a step further and changed the behavior of frames on short Digg URLs to have no frames, as opposed to automatically displaying a frame like the default behavior was when the Diggbar first launched. Other services, like Adjix, don't allow you to disable frames at all. So not only are they displaying ads on other people's content, they're also locking you into viewing them, by not offering a way out. It isn't only morally wrong to profit from other people's work, but it is also wrong to force people into viewing your ads on other people's content.

In short it's quite simple: There needs to be a standard for services to follow if they want to be used by default (or at all in general preferably), transparency is important and URL shorteners are breaking it, browsers and services should automatically expand short URLs to show their true destination, and framing content is bad.

Have some thoughts on the subject? Leave a comment, and let's get some discussion going. After all, this is a fairly touchy subject, not to mention how very relevant it is.

—