The Spatial Miscellany


A weblog. A website. A geospatial miscellany…

Raster to Vector Magic

A tantalising blog post from Steve Coast in his new role at Microsoft. What if we could build vector datasets from a raster image? Well it’s been tried before, not least by my friend and colleague Bong Khin Fah, but Microsoft have stood up an experimental service specifically designed to capture roads from their BING imagery.

You can see a video of the new service from Microsoft in action below, or click here to try it out. Fingers crossed this is the tip of an iceberg of some pioneering R&D in Seattle.

Faster web mapping with Google’s new browser?

Last month Google released a web browser…Google Chrome. It appears to have debuted to mixed reviews, personally I really like it, but others have had less favourable experience. On the surface, it looks much like any other web browser, but underneath, it’s a bit of an animal.

Google have acquired a new JavaScript engine (V8) written from the ground up to work more efficiently with websites that have large amounts of JavaScript rather than the little snippets of JavaScript which was typical of websites developed when JavaScript was first integrated into Netscape Navigator in the mid 1990s.

Typically JavaScript engines use a dictionary-like data structure as storage for object properties – each property access requires a dynamic lookup to resolve the property’s location in memory. V8 works differently…the first time V8 encounters an object it interprets how the object would be represented as a class, creating a hidden class, which means the next time the object is encountered, its properties can be accessed from memory without the time consuming dynamic look up.

Google give a far more comprehensive introduction to this approach in their documentation of V8, conveniently they use the example of an object common to all GIS developers…a Point.

A class based approach to JavaScript

So What?

The novel approach taken by the V8 JavaScript engine, presents a new opportunity for GIS web developers to work with Points, Lines and Polygons on the web client instead of the web server as is typical of a web based GIS. Using JavaScript in the web browser removes the need for lengthy round trips to the Server, which will make for faster web mapping applications.

Here and now, this will allow web mapping API developers to work with more markers, the typical 100 marker limit of Google Maps applications (and similar web mapping API’s) – is no more; Mike Williams and his team have reported working with as many as 2000 markers when using Google Maps within Chrome.

Chrome isn’t the only web browser to recognize the importance of working with JavaScript heavy web sites, for example, much work is being done on a new JavaScript engine for Firefox. Perhaps in the future, as this approach gains support, spatial analysis functionality can move from the server to the client, which would make for a more engaging web mapping experience?

If anyone is looking for a dissertation project, or has time of their hands, it would be interesting to see the outcome of taking some topological operators (e.g. Java Topology Suite), Google Web Toolkit (GWT) and GWT for Google Maps, all served up with Google Chrome?

GIS Software Above the Level of a Single Device

A couple of weeks back I surfed over to the Nokia website to check out the latest developments for their Series60 device platform (S60); unsuspecting I stumbled upon some software that really got me thinking.

It turns out, a couple of years back Nokia embarked on a project to port the Apache web server to the symbian operating system that underpins their Series60 device platform. The project was subsequently handed over to the open source community and you can get involved here, project raccoon. Interestingly, Nokia have recently wrapped the web server as user friendly software with a supporting website:

I installed the application on my N95 and as you might expect it provides complete access to the contents of my mobile phone via the web. I can fire up the web browser on my desktop PC and browse to a web page, and then click a button on the web page to take a photo with the camera on my phone, wherever my phone might be. I can then use my desktop web browser to browse the photos I have taken, or any other information I have on my phone e.g. contact details or calendar events.

My mobile phone via a desktop PC web browser

This struck me as a unique piece of software, I tried to think of other software that functioned in a similar manner. After roaming the web for some ideas I found an article from Tim O’Reilly that suggested similar behaviour could be observed with Apple iTunes, he’s coined a term to describe such software as…‘software above the level of a single device’.

Installing the mobile web server software on your phone, allows you to use the software on any number of devices…your phone; a desktop client with web browser; a games console; or any other internet enabled device, even someone else’s mobile phone? Software above the level of a single device – just as Tim O’Reilly describes when he observers that you can control your iPod from an iMac.

From a geospatial perspective, couple the GPS enabled N95 mobile phone, with the mobile web server, and we have a tracking service that can be consumed by any internet enabled device, powerful stuff. Perhaps worthy of more consideration, I look across the GI industry, GIS software vendors, the open source community, and other corners, but I don’t see ‘GIS software above the level of a single device’ as an overriding design architecture?

There is a bigger question here…what happens when 3 billion mobile phones run as personal web servers?

Summer of Code & OpenLayers

Google Summer of Code (GSoC) is a program that offers student developers cash to write code for various open source projects. Google will be working with several open source, free software, and technology-related groups to identify and fund several projects over a three month period.

OSGeo and Summer of Code

I think it’s a great opportunity for students, surely better than spending many summer months obsessing on an esoteric thesis that will never see the light of day – perhaps that was just my experience? Christopher Schmidt and others have offered their support for students wishing to further develop OpenLayers, it would be great to see some of the stuff on the list tackled.

Fight spam with reCAPTCHA

When you submit details to a website via a web form, increasingly you’re asked to interpret a picture of a word and type the answer in a text box, this type of puzzle is known as a ‘CAPTCHA’ – if you’re not sure what I’m talking about, check out the comments section of this post for an example.

The CAPTCHA was created by Luis von Ahn, in an attempt to fight Spam…if the CAPTCHA is completed successfully you are assumed to be human and the web form is submitted, if the CAPTCHA is failed you are assumed to be a computer and web form submission is prevented. The CAPTCHA is a classic example of a Turing Test, as proposed by the eponymous researcher in his 1950 paper ‘Computing Machinery and Intelligence’ – indeed the name CAPTCHA coined by Luis von Ahn is an acronym for a “Completely Automated Public Turing test to tell Computers and Humans Apart”.

An example captcha

Ahn was initially proud of his highly effective solution for preventing Spam, but was subsequently frustrated with the cumulative amount of time being consumed by millions of people across the globe filling in CAPTCHA’s and producing very little in return. Personally I found the CAPTCHA mightily offensive and avoided them like the plague, until I discovered them to be the only effective way to stop computers spamming this blog. Fortunately, Ahn has recently worked with Ben Maurer to address his frustration and recently released reCAPTCHA. Here’s the deal…

The Internet Archive is attempting to automatically digitize old books using optical text recognition software, they are largely successful but the text recognition software struggles with recognizing ye olde English, meaning that roughly 8% of words are digitized incorrectly or not at all. reCAPTCHA addresses this problem by asking a user to interpret a picture of two words (instead of just one), the first word is a known word and the second word an unknown word. If you correctly interpret the first known word, it is assumed you have also correctly interpreted the second word that wasn’t recognized by the text recognition software. So when you fill in a reCAPTCHA, not only are you proving that you are human, but you are also helping to digitize old books! For me, application of technology in this way is poetry.

Obviously Luis von Ahn has applied this technique to digitizing old books, but it had me thinking of the potential to apply such an approach to digitizing maps…is there scope for a geoCAPTCHA?

The above cartoon is the copyrighted work of David Farley

Continue Next page

Free GIS Software...

Download ArcGIS Explorer, a free geobrowser from ESRI!


Use ArcGIS Explorer to visualise geographic datasets. The latest build provides full access to Virtual Earth imagery and comprehensive support for several data formats including GeoRSS, KML and ArcGIS Layerfiles.

Before you go

Going so soon? Test your geography with the...


Do you support the campaign? Should government-funded and approved agencies such as the Ordnance Survey collect data with significant indirect contributions from the UK tax-payer, but then charge users and companies for access to it?


Download Flash plugin