Twitter said recently it plans to roll out its own link-shortening service, known as t.co, and “wrap” any links that are posted by users in a t.co link, even those that have already been shortened with other services such as Bit.ly. The company said it needs to do this in part for security reasons, so that phishing attacks and malware links don’t get through, but that it also plans a number of analytical services based on the link-sharing data of users. In the meantime, however, Twitter is coming under fire from developers — including a prominent member of the Google Buzz team — for the way it’s planning to implement the new link-shortening feature.
In a post on his Buzz account today, DeWitt Clinton took issue with a number of the elements of the new Twitter service. Although the engineer made a point of saying that his criticisms applied to other link-shortening services as well (and he stressed in a footnote that his post represented his own views and not those of his employer), it was obvious that most of his concerns are aimed at Twitter. Among other things, the Google engineer said link shorteners are bad for the web because they deliberately obscure the ultimate destination of a link from both the user and from search engines such as Google (he didn’t discuss Google’s own link shortener).
Clinton said that by convention, link shortening should only be used “at the boundaries of the web,” and that services that want to be open to users and the web in general should use shortened links as little as possible, and “unroll” them whenever possible to show the actual link being posted. He added that:
Twitter’s recent decision to make every link a short URL positions Twitter outside the web, rather than a part of it, which likely does a disservice to the long-term health of the young platform.
The Google engineer also said that because of the way the company had configured its new link-shortening service, it was no longer going to be a neutral content-delivery service, but was now going to be modifying user’s messages in ways that they might not understand or appreciate, since their links would be checked and effectively redirected.
This is a subtle but profound shift in the Twitter platform dynamics, and one that I feel undermines Twitter’s considerable current strengths as a neutral content distribution network.
And Clinton criticized Twitter’s attempt to get developers to display a different URL or link instead of the t.co one when using the feature. The Google engineer said that it was an example of “link masking” or “link cloaking,” and that it was the kind of behavior that was discouraged by search engines as a hallmark of “affiliate link spam.” Clinton recommended that Twitter not require the link masking, that it make the use of its shortener opt-in instead of the default, and that it “consider relaxing the 140-character limit altogether” since that is the main reason that users need a link shortener in the first place.
The Google engineer’s post was echoed by developer Kevin Burton of Spinn3r, and a number of others wrote posts in support and gave his Buzz post the virtual thumbs up, including Google open standards advocate Chris Messina. But are his criticisms fair? More than anything else, Clinton and some of the company’s other critics seem to be complaining that Twitter is wrapping links so that it can track what’s being shared and clicked on — but of course it wants to do that. Why wouldn’t it? And just because that behavior irritates Google doesn’t mean it isn’t appropriate.
Whether the engineer’s criticisms get any support or not from the broader web community, it seems obvious that Twitter’s move into link shortening has stirred up some strong feelings about its behavior, and some concerns about how it might behave in the future, as it tries to become more than just a neutral content-forwarding service. I’ve asked Twitter for comment on Clinton’s post, and will update if and when I get a response.
Related content from GigaOM Pro (sub req’d): Lessons From Twitter: How to Play Nice With Ecosystem Partners