notes.husk.org. scribblings by Paul Mison.

Above London: a retirement notice

Over three years ago, at the Yahoo/BBC Hackday in Alexandra Palace, candace and I knocked up Above London and Above SF, two Twitter bots that would alert followers in one of those two cities that the International Space Station (or an Iridium flare) would be visible.

I’m happy with the reception it got, and I still prefer its output to that of some of the subsequent services that provide the same service (such as @overlondon or @twisst). However, it was always a bit of a pain to look after (true to the word “hack” in the event title, I cut corners when it came to handling DST, and occasionally the cron jobs running it would fall over).

Two things have finally done for it: Twitter’s move to OAuth, and more importantly, the fact I managed to leave the only copy of the code on a server that’s now sitting, unplugged, in the UK. Even the service on which I posted the write-up of the hack has now closed. Given that, it’s probably best that I post a message to the Twitter accounts, and formally shut up shop (for now, at least).

Thanks to everyone who followed either the San Francisco or London account, and good luck with one of the aforementioned alternatives. I hope you got to see the ISS at least once. It’s always warmed my heart to look up and see the few humans that circle the world, shining brightly in the evening twilight.

Archiving Twitter Via IRC

A couple of people have started posting the sort of thing I’d post here, but split into multiple Twitter posts. Unfortunately, given the absence of meaningful Twitter archiving and the way that the conversations (which can admittedly be useful, but are usually just distracting) drown them in the what you can see on Twitter, they’re quickly lost.

Even using URLs is tricky. For a series of posts, you end up having to open multiple links, and there’s no cohesive narrative. At least that gives you as a user a method of archiving them, but I though there’d be an easier way of doing so.

#2lmc have used a bot to fetch page titles for years. Pairing namer.pl, which only does that, with a from-Subversion copy of URI::Title, which fetches the full content of posts to Twitter, lets you quickly drop in post URLs and extract a conversation. Here’s John Siracusa, last week:

And here’s Fraser Spiers, this morning:

Obviously, it’s far from perfect, but at least it’s a single block of text that can be easily edited down to some meaningful sentences. (While the above is captured as text in my IRC logs. I thought it was a bit mean to turn what people might have considered ephemeral into archived HTML, so I’ve posted screnshots of my Terminal window containing the IRC client.)

I may try and do this more in future, or even build a more web-native method of doing so, but then I wonder if, despite the fact that you can’t copyright a post, people might get offended. It also fails (as so many Twitter apps do) when one participant is private. Ah well.