Monthly Archives: April 2006

Weighted Car

This is a pic of Nick Hill’s car after being loaded with servers, rack cabinets and stuff. It took hours of work, and Nick drilled out some rivets to take the cabinet apart. Currently we’re waiting on the official green light on hosting, at which point all this stuff will get installed.

Podcast: Nick Hill

Nick and I spent a day tearing apart donated computers and racks last week and afterward ended up in a pub. We talked about openstreetmap and how he’s getting data in there. Here is an 11 meg mp3 which lasts 25 minutes. Again, if you subscribe to the RSS feed in iTunes etc then you’ll magically get the audio in your iPod. Enjoy!

Old Maps, Aerial Photographs and Experimental Derives

Richard picks out a great Guy Debord quote in his review of Simon Sadler’s The Situationist City. I’ve found a longer version of the quote which I’ll reproduce for posterity here:

“With the aid of old maps, aerial photographs and experimental derives, one can draw up hitherto lacking maps of influence, maps whose inevitable imprecision at this early stage is not worse than that of the first navigational charts; the only difference is that it is a matter no longer of precisely delineating stable continents, but of changing architecture and urbanism.”

Fantastically resonant, that.


Google Maps vs Mapstraction, pt II

This was going to be part of an earlier post, but should now stand alone.

When we’re not working on free geodata creation for OpenStreetMap, we like to make the best of the current free beer commercial mapping APIs to make neat things like GPX viewers and pub maps. In the light of some of the more amazing Google Maps mashups out there (here’s a recent decent), these sites might seem underwhelming, but if you consider how quickly they could (or couldn’t) have been put together a year ago then it’s clear that thanks to Google Maps we’ve come a very long way.

Tim O’Reilly recently asked the Mapping Hacks people, and others, why Google Maps was so much more popular than all the other APIs available. Schuyler pretty much nailed it, though after my Mapstraction research I think he’s too generous towards Yahoo and Microsoft.

I note with disappointment that neither Yahoo nor MS offer polyline support, which leaves my desired cross-vendor demo of a GPX viewer dead in the water. It also means that anyone requiring that functionality for drawing routes or boundaries is stuck with Google for the time being. Ironically for Google however, their recently developed open source Explorer Canvas might offer a cross-platform way for Y! and MS to compensate for this. Further disappointment with alternatives to Google Maps comes with the realisation that Yahoo’s maps only cover North America so far (although I was pleasantly surprised to find Microsoft is the only provider to give any road or placename data outside of North America, UK and Japan).

Microsoft’s Virtual Earth SDK feels clunky to code with, shirking the fashionable javascript idioms which help make Yahoo and Google‘s offerings feel elegant, and their API is completely malnourished with regards to easily placing nice looking map pins. Sure, it might be flexible, but Google and Yahoo’s offerings look good out of the box which is really important. (The same can be said of the developer documentation, too).

Whilst looking at all of this I took a look at OpenStreetMap’s slippy maps interface to see if I could allow plotting of pins. Unfortunately, although the Civicmaps code base we started with supports plotting GeoRSS, it looks like in the process of adding in support for Mercator projection that the marker features were broken. This means that for an OpenStreetMap-powered javascript maps implementation you can put on any website, there is still lots of work to do. However, following this exercise I’ve discovered that if anyone wants to seriously compete with Google in this space, then we all have a lot of catching up to do too!


Free Web Mapping APIs and Mapstraction

When I first got involved with OpenStreetMap, my aspiration was to make free maps to enable something like Wayfaring or Platial to be built. Even though those sites and many others are storming ahead without us and creating great services on top of Google’s Maps API, I still think it’s important for them to be able to access and build upon free geodata and avoid lock-in with commercial vendors. This is especially important for sites like Placeopedia that use Google Maps to geo-code things, because as Mikel Maron and Richard Fairhurst have pointed out (see the comments) it’s entirely possible that the map owners could claim ownership of the geodata they are creating.

With all that in mind, I took a look at what could be done to make it as easy as possible to switch mapping providers. In the short term, this would mean making it easy to flip between Google Maps, Yahoo Maps and Microsoft Virtual Earth in response to changes to T&Cs, introduction of ads, addition of better data etc. In the long term, a common javascript mapping API would mean that free offerings such as OpenStreetMap or Worldkit could quickly be adapted as drop-in replacements for commercial providers. I’ve been collecting my thoughts on this at based on discussions with Steve and Mikel, and there’s an illustration of how it might work here.

I’m not the only one thinking along these lines, of course. The Open Source Geospatial Foundation has a cross-project mailing list for people interested in similar issues, and numerous hacks and comparisons abound as people work out how similar but different the various mapping platforms are. Hackers everywhere have been murmuring about their dependence on Google for some time, and maps aren’t the first area where Google’s initial offering invites obvious commoditisation.

Looking at the common features of the mapping APIs, it’s clear that the bar is currently set quite low: it’s scrolling tile-based maps with map markers, folks. With tongue in cheek, Schuyler Erle refers to this as “red dot fever” (more context here, though I don’t know if that’s Schuyler’s writing or a like-minded collaborator). Numerous open source equivalents have emerged, but none has taken hold to my knowledge. This cements the general feeling among the responses to Tim O’Reilly’s open question about Google’s Mapping API dominance: Google Maps is actually very good, it was first to market, its maps look nicest, its terms and conditions are reasonable and as a developer it’s pretty easy to work with. It’s clear to us that OpenStreetMap has a lot of catching up to do, but it’s also become clear to me that Yahoo and Microsoft do too. I’m watching this space.

Tom Carden

Our first spam

Openstreetmap had its first spam yesterday from someone uploading a html file (a web page) as a GPX file (a GPS trace file). Being XML (similar format), it got parsed and uploaded with spam-like tags linking to the individuals business. The point being that google etc would crawl the pages and increase his page rank. It exposes the soft underbelly of openstreetmap’s policy of crossing bridges when we get to them – we’ve only had good bridges so far!