geobloggers

Last week we (as in flickr) launched the whole new Places pages. Something of which I’m inordinately proud of. Although I’m still a bit too tired and busy with other stuff to be able to fully enjoy it, I think I’ll feel a lot happier and content when things have settled down and I have a chance to tune and tweak it at in a slightly more relaxed reflective state.

[Now, I was going to do a post where I got to talk about all the cool and interesting things about the project, how it came about, development strategies and prototyping we used, how the photos are picked, why we did it and so on. But in starting that I realized how overwhelming all that would be, not only for you to read but me to write. So I'm splitting it up. Also to ease myself back into blogging after having my nose to the grindstone for a fair while. First off, URLs]

Anyway, just incase you haven’t seen it yet, you can dive in at the Places landing page, (or try Reykjavik my new favorite place on the whole planet) or take in a pretty screen shot. In which you can see how we moved some of the buttons and layout around.

The “Places” project is one that’s been brewing for quite a while. My glib reason for having them, is actually to have a page from which we can hang location based geoRSS (and KML) feeds off. I mean we have geoRSS feeds for San Francisco and so on, but up to now no-where on the site that you can get to them from. As it happens we still don’t have the feeds on the page, but, for a good reason!

George insisted, I agreed, and between us we forced (although it didn’t take much) Kellan to make the URLs for Places as human readable as possible. On the grounds that’s it’s very important (link selected to potentially confuse Tom as to why a post from 2004 is getting a couple of extra visits). Although human readable in this case doesn’t always translate to hackable URLs, until you start to get used to how they work, then you can have a good stab at it. Example, which as it happens I did just hack together …

http://www.flickr.com/places/United+States/Texas/Austin?q=sxsw

… seems fairly obvious what you’re about to get yourself into, which is good. Sometimes the URL is not quite what you’d expect …

http://www.flickr.com/places/Iceland/Reykjavik/Reykjavik

But you (should) still know where you’re going to end up.

There’s a terribly good reason why sometimes the placename is repeated in the URL like that. And one day I’ll sit down with Kellan and pay more attention when he tells me for the 20th time why it does this. I do know however know that turning essentially a record ID from the backend into something that a human can understand, and yet the backend can once more turn back into the same record ID wasn’t as easy as I thought it aught to be when we asked Kellan to do it :) and lots of effort went into making it so.

Basically you have to ask the backend, where something is the ‘child-of’, to build the hierarchy. However as we don’t have full control of the backend you often get a city, let’s say London, belonging to a place called Greater London, which, as far as I can make out from the bounding box, is about 2 inches taller and wider than London. Or York (in England) belonging to something like Duke Elligton’s Marginal Lower Land Barrows of the Kidsworth Council Academic Elective Region. Because that particular piece of geographic data was extracted from the government’s parish records from 1895. It wouldn’t surprise me to find that some towns in England were a child of a mobile library’s service region.

Anyway, hidden away from me in a black box constructed by Kellan and Aaron is all the complex magic that works out what’s important and what to miss out. All of which means that Places URLs are slightly different to the geoRSS URLs we already have. So, once we have the older rss URLs pulled inline with the Places URLs we’ll slap the RSS (and KML) feed back onto the place. Meaning you’ll not only be able to view a page for a place, but subscribe to new photos from that place too.

The main thing I did learn from all these URL shenanigans is that working with location is still surprisingly hard (who knew?). I do however have more to say on that matter in it’s own post all of it’s own :)

So, yes, Places, Yay! Soon, Part II, photo buckets and how Place’s photos !=== Interestingness.

As a side note, new version of ecto is a bit odd, so I hope this works.