One of the limitations of Twitter is that it’s hard to tweet a URL without exceeding the 140 character limit. A result of that has been the growth in URL shorteners, such as bit.ly, is.gd, ow.ly and so on. Last year, Twitter got in on the act itself and registered its own short domain, t.co, to use as a shortener.
Recently, you may have noticed that Twitter now automatically shortens anything longer than 19 characters using t.co, so you don’t need to use a third party shortener any more. This is applied both to tweets made via the web and via any third party application which uses the Twitter API (such as phone and desktop apps). You can read Twitter’s announcement of the change on their developer blog.
Sounds like a good thing, no? Well, no. For a start, it now means that any URL which is already being shortened by another shortener gets re-shortened. That includes bit.ly, as well as various “own brand” shorteners such as bbc.in (used by the BBC) and gu.com (The Guardian). As well as being pointless, that adds another, entirely unnecessary, layer of redirection into the system. It also means that it’s impossible to tweet a longer URL unmunged even if you wanted to – and, while that may be uncommon, if you do want to then you certainly don’t want Twitter rewriting it.
What’s more of a concern is that t.co, like all Twitter domains, is blocked in places like China. While Twitter itself can be accessed via intermediary applications which connect through a VPN to the Twitter API, if a URL is transformed into a t.co link then the ultimate destination is much harder to reach as it can’t simply be clicked on no matter what software is being used to read the tweets.
More worryingly, Twitter plans, eventually, to use t.co for all URLs, even those that are shorter than the 19 or 20 characters that a t.co URL takes up. The reason for that is that Twitter wants to eliminate all non-Twitter URLs fro the stream. As Twitter developer Taylor Singletary puts it:
It’s limiting to think of this service as simply a URL shortener. When you view it this way, some of its functionality may seem absurd to you. t.co is a URL wrapper. It wraps URLs to primarily protect users from malicious content and to better understand how users interact with shared content — understandings that will yield product features and additional APIs in the future.
The anti-spam argument makes a bit of sense. Being able to switch off a short URL that is being used abusively is a valuable defence against spam and phishing. But other shorteners already offer the same option, so forcing all URLs through t.co isn’t going to help a lot with that. But it’s the second part of Taylor’s comment which most concerns me. It seems to me that this is being driven more by Twitter’s own need to find ways to monetise its operations. Can we be sure that, in future. t.co URLs won’t redirect via interstitial pages containing advertising, for example? I’m not opposed to advertising – I’ve got it on this website – but I am unhappy with it being used intrusively. Forcing all URLs through t.co also means that Twitter will build up a huge database of URL clicks, allowing it to see who follows which links. There are some very good privacy reasons why that’s not necessarily a good thing.
If you think it’s a bad idea too, can I encourage you to go to the Twitter developer blog and add your own comment. You don’t need to be a user of the developer API, all you need is a Twitter account to sign in and comment. If enough people make their voices heard, then it’s possible that minds may change.