Five ¢ents

When Short URLs Break

Update: 301Works.org announced to solve this problem

Update 2: Google and Facebook announce their own url-shorteners. Things are getting heated…

One of the wonderful aspects of the modern demo-to-market web environment (aka Web 2.0) is the rise of small, cheap utility services. Sure, they’re very often copycats that offer marginal if any improvements on the services that inspired them, but theoretically it creates a breeding ground for competition.*

In no arena is this more true than for URL-shortening services.

URL-shortening services

A shortening service takes a long URL (such as the one for the page you’re now reading: http://fivecents.ianwessman.com/post/67353947/when-short-urls-break and provides an alternate, such as http://bit.ly/break.

The URLs one receives when using a shortening service are opaque, necessarily so. Some of these services now provide ‘vanity’ URLs, because http://bit.ly/NYTimes is much preferable to http://bit.ly/TdATv.

Pioneered by tinyurl.com in 2002, the playing field has grown to include dozens of ‘competitors’, including ping.fm and bit.ly. The raging success (as measured by users, not revenues) of these shortening services is in no small part thanks to Twitter, the also-not-profitable microblogging/asynchronous chat service that has become de rigueur for those who needed something to post to in between longer blog entries.

The Danger

But what if your shortening service goes belly-up, as most eventually must? You’ve already thrown away most of the information in your link by using this service, and there’s no automatic way to recover the original when bit.ly bites the dust.

Today, for a time, ping.fm was unreachable. While this did not translate into any great economic harm for anyone (I was merely trying to follow this link), it got me thinking:

URL shorteners should make their databases public if they go away

This is not a complicated goal, but it is a necessary one. Dead links are just as dangerous to the health of the web as missing pages, and there’s no Internet Archive or Google Cache for short URLs.

These databases, simple hash-to-string lookup tables, have no direct monetary use to creditors. That value lies in the ‘extras’ of these services, such as usage and traffic data. Giving these simple databases away would cause no further harm to the developers, but would be a boon to the web as a whole.

What I’d like to see is simple: a public agreement, signed by as many shortening providers as possible, to release their databases as either unencrypted MySQL stores or as large CSV files if they cease operation. This data should be either CC-licensed or, preferably, made public domain.

In that scenario, an interested but mostly-neutral party (Google, I’m looking at you) can provide a stopgap service that keeps all those old, shortened URLs relevant.

Who’s up for the challenge?

* We’ll side-step why this is good for the market as a whole, even though few will profit from it.