As I write this, I’m trying (and failing) to load twitter.com/home. Oh - it just timed out. According to Is Twitter Down? it’s not down, but I can’t get to it. There have been no updates on my timeline for two hours now, which is quite rare for a working day.
I just wonder whether Twitter has reached a natural end. You know “it was fun while it lasted” sort of thing. And it was fun, until relatively recently. Then the number of “spam” followers increased, thanks to the ease of scripting against the API, no doubt; on the other hand, the API also allowed fantastic tools like the Twitter Twerp Scanner to me made, so I don’t think Twitter shouldn’t have an API.
With or without an API, the problem with Twitter (as has been said numerous times) is that it doesn’t scale. It’s a centralised (if clustered) service unlike email, blogs or Plain ol’ Websites, which can exist anywhere and conform (roughly) to a standard. To Tweet, you must Be On Twitter. This is its fatal flaw while being central to the way it works. Mike Arrington wrote on Techcrunch a while ago on how Twitter might be decentralised.
Twitter: a highly-addictive social experiment that just goes to show, by counter-example, the merits of decentralisation.
June 2nd, 2008 at 5:18 pm
I’d advise ignoring all content coming out of TechCrunch, on all topics.
Spam followers might be bad, but why does it even bother you?
Twitter being down a lot is a bad problem, but it can scale. They just haven’t done it.
June 2nd, 2008 at 9:02 pm
Hi Phil,
I normally do ignore Techcrunch; the blog post there just came up when I searched for “Twitter decentralization” proving, if nothing else, that it’s not just me talking in such terms.
I block “spam” followers. Why do they bother me? I’m not sure, really. The whole “follow thousands of people in the hope that a small percentage follow back” thing pollutes the pond for the rest of us, in that I don’t suppose it can help the performance and uptime of the site.
But if the only way to scale is to do an IBM and “throw more iron at the problem”, isn’t there something fundamentally flawed in the system?
June 2nd, 2008 at 9:38 pm
I’m not sure decentralisation is the answer…
It appears Twitter wasn’t built to scale and there appear to be plenty of suggestions for how to fix that - whether they will work is another question of course.
Of course the real problem behind an alternative to Twitter is if I go will my friends go too?
June 3rd, 2008 at 3:25 pm
I don’t think anyone’s said there isn’t something fundamentally wrong. It seems like they got into a bad place where the resources required to keep twitter running were equal to those needed to re-engineer.
Would people have been happy if twitter had gone down for a fortnight whilst they re-wrote? Do you think there’d be an audience by the time they came back? It’s a set of tough calls and they’re muddling through.
I definitely don’t think decentralisation is a good idea for this service, although I don’t discount it for a separate service. Also, it’s worth making clear that, as you imply, just because other people think the same thing, doesn’t make it true (argumentum ad populum innit).
June 8th, 2008 at 5:49 pm
Of course it not just Twitter, services like tinyurl also carry the same risk. Having said that in my experiance del.icio.us seems pretty reliable they must have invested in sufficient infrastructure.
I came across SMOB
http://www.johnbreslin.com/blog/2008/05/09/prototype-for-distributed-decentralised-microblogging-using-semantics/
an interesting looking little project, I haven’t tried installing or running in anger but it may be one way to take a little more control over your micro-blogging.