Monday, 17 August 2009

Backing up and the trouble with short URLs

40% of everything on Twitter is 'pointless babble', according to a new study by Pear Analytics,* with only 8.7% of tweets having 'pass-along value'. (The company neglected to comment on the possible existence of babble that has a point).

Despite these statistics, however, the existence of services such as BackUpMyTweets (among other options) suggests that at least some of those tweeting their babble (pointless or otherwise) may want to preserve it for future reference. Especially as there seems to be some doubt as to how long Twitter will store your tweets, or at the very least regarding how long tweets remain easily accessible and searchable. Indeed, you might want to back up your other online accounts too, just in case. But are these back up services capturing everything? Well, no, not quite.

The problem - as highlighted by the recent ructions over - which first died, then decided it felt a bit better, then finally dispersed into the universal consciousness (or something like that) - is with URL shorteners. What, if anything, is being done to back up their databases? Or to put it another way, is the connection between, say, and (or between signifier and signified, if you want to get all structuralist about it) being recorded anywhere, besides (in this case) on's servers?

Actually, I do want to get a bit structuralist about it, as it turns out. Because perhaps one way of looking at it is that URL shorteners, at present, represent a threat to the future relationship between what we are saying now and to what we are referring. A potential glitch in the language of history.

With the internet and social networks playing an increasingly significant role in current events themselves, as well as in recording them as they happen, and indeed with short URLs being employed by offline news media to reference the online world, it seems increasingly important that something is done to preserve the semantic relationships these URL services create. After all, on one level, the net is essentially a gigantic, complex, system of languages. A modern day Babel that somehow works.

Of course the net has always had to contend with broken hyperlinks. But a short URL deprived of its reference is something else again - a broken link at least hints at where it was supposed to lead, but a short URL is arbitrary. It isn't even a code that can eventually be cracked, just a random alphanumeric sequence assigned to whatever standard URL a user has fed into it.

Happily, though, a group called now seems to be attempting to deal with the problem. As yet, all we have to view is a homepage containing slightly opaque bits of blurb, but what the group appears to be aiming at is something a bit like the Internet Archive, but for short URLs.

Already, though, the project isn't without controversy - the people behind dispute the altrusim of 301works, hence their decision to release their code and database to the open source community, rather than signing up. But at the very least it seems that the issue is now on the collective radar and might soon be addressed - and lets hope it is, or much more than 40% of Twitter might one day be pointless babble; never mind various other parts of the net.

Still, if you really want to be certain about backing up your every online utterance, maybe you could do your own shortening? (Or, then again, maybe not. Life's plenty short enough already).

*Based on a highly representative sample... of, er, 2,000 tweets. Probably mostly from people bored at work, given the times and days they were selected.

No comments: