october 2004 archives
Sunday, October 24, 2004
Behind the scenes: web maps
Last year's article The Making of a Web Map charted the technologies used on the web sites submitted to the Web Mapping Contest 2003. No conclusions were drawn, but the graphs showed clearly what client-side and server-side technologies were used, what databases and what online GIS software. Only a few days ago, this year's contest closed. What technologies were used on the web sites submitted this year?
It's difficult to present the data on this in bar charts. As noted in the article not every entry listed all of the information that was requested. In addition, more than one technology may have been used for each platform
. However, I feel there's a lot of technical confusion among developers of online mapping web sites. Some entrants list a particular server-side technology, e.g. php, as a client-side technology. Others mention DHTML, JavaScript, and HTML, but omit CSS.
Both general purpose and proprietory plug-in technology for vector mapping is on its way back. Fewer web sites use the Flash player, a plug-in that is in almost every browser. Java applets, ActiveX controls, and the SVG player are being used less. Despite the ongoing efforts to promote web standards, CSS is mentioned only a few times.
It's difficult to summarise the statistics on server-side technology. Entrants could add any technology; some focussed on the web server, others the operating system, or scripting language. Especially the use of Microsoft products spawned long lists of components. Well, of course you run your .NET application on IIS: duh! But it would have been interesting to learn on what platforms open source GIS software installations run.
ArcSDE on top of Oracle and SQL Server is the most popular database configuration. MySQL is the most popular open source database, whereas PostGIS is generally known for its more mature spatial extension.
Love it, or hate it. Arc IMS is the online GIS software of choice for most web sites. MapServer is second best followed by MapGuide and MapXtreme. There are many more smaller solutions being used. Another interesting trend is the slow adoption of web services, be it an OGC compliant WMS, SOAP or XML-RPC. Little by little, online mapping web sites are using mainstream web technologies.
![]()
Wednesday, October 20, 2004
More of the same
With the US presidential elections coming up, there's a lot of interest for maps in election campaigns. Just recently, I highlighted the practice of gerrymandering and how maps can be used to reveal where presidential candidates get their money from. Wired featured an article last week, discussing ways in which maps change the way politics and elections are done
. For example, voter-outreach specialists can easily identify specific areas to get the people to the polling booths who wouldn't be targetted otherwise.
They're only just starting to explore the political benefits of GIS: desktop technology has incited a revolution! Desktop technology? That's just so 1990s! Why not use Open Source online mapping tools, for example IndyMapper. It's a web-based map authoring interface where a non-technical person can dump in demographic data developed by Jo Walsh and Schuyler Erle.
Clagnut pointed me to Plazes. A Plaze is a location with a local network. The system uses a fingerprint of the network you are logged onto as a unique identifier for a specific Plaze. The key is being connected with the geographic location of that network. Users have access to pictures, comments and mapping information left by others and can connect to the people currently online at that Plaze. One of the Plazes is the Telematica Institute in Enschede. You guessed it: they use the Multimap.com Free Linking Service.
![]()
Monday, October 18, 2004
Location and social networks
Last year I presented a paper at the Cartography 2003 conference on collaborative mapping. In collaborative mapping, the value of annotations (geo-tags) depend on both physical and social proximity, expressed in distance and “degrees of separation”. Dodgeball is yet another take on FOAF networks. Users do not only obtain insight into who they know through their friends, but they also learn the location of their friends and acquaintance.
Tell us where you are, and we'll broadcast your location to all your friends and let you know if any friends-of-friends are within 10 blocks.
Once signed up, you go about your life moving freely through space and receive a text message (SMS) on your mobile when someone in your network is nearby. Scary as in “Big Brother”? Well, as long as you don't send out any text messages, nobody knows where you are. Acoording to a recent article in Directions Magazine, Dodgeball is not as compelling as it seems at first glance:
- Users have to keep sending new text messages to update their location.
- Dodgeball reports last location, even if it's seriously outdated.
Other drawbacks mentioned seem to be tied to the fact that text messaging is just not as popular in the US as it is in Europe. Secondly, the mobile networks operate differently. For example, users may send text messages only to others on the same network. If Dodgeball were to be launched in Europe, these drawbacks may not be applicable, thus ensuring a much more positive reception.
By the way, this reminds me of an observation I overheard a few weeks ago. The likelihood you are willing to surrender your location to others can be modelled to a bagel. The distance from the centre indicates the social proximity. The user broadcasting the location is positioned in the centre. The thinkness of the bagel indicates the likelihood of you passing your location to people you are related to. Thus, the likelihood you inform your relatives of your location is very small, or even zero. Indeed, the bagel is typified by the hole in the centre! Further away in your social network, the likelihood you inform your friends about your location is much bigger, but up to a certain degree. Maybe the accuracy of the location should be modelled into this as well? Just a thought...
![]()
Sunday, October 10, 2004
Maps and politics
Maps are a political statement, sometimes on purpose, sometimes unconsciously. When Eurostat published its Statistical Compendium 2004, they'd never imagined it would stir the emotions that much! On their website, it reads: It provides a good understanding of all topics on which Eurostat has statistical data available.
However, at first glance it's obvious Eurostat does not live up to this statement: the country of Wales is missing on the cover. Fair enough, the map of Europe is highly generalised, but there is a fine line between proper cartographic generalisation and leaving out a country altogether! One Welsh MP remarked:
Mistakes like this make it more difficult for pro-Europeans like me to convince people of the benefits of a united Europe. It's a genuine error but in the current political climate the European Union really should be aware that a mistake like this is a monumental gift to Eurosceptics.
This is a serious oversight, but borders are also redrawn on purpose. While the US exports its democratic values abroad, changing the geography of electoral districts will have a profound influence on American democracy at home. Redistricting is the — perfectly legal — process by which the boundaries of elective districts are periodically redrawn to maintain equal representation on the basis of population. Sometimes these boundaries are redrawn, not because of population change, but to maximise electoral advantage. Partisan politicians select their constituents rather than the other way around. Leading up to the presidential elections, a BBC report reveals Texan Democrats currently seem to be hit hardest by this practice, referred to as gerrymandering.
With help of a GIS package and maps like the Money Map, it's fairly easy to predict the electoral effect of changing boundaries. The Redistricting section of the Texas Legislative Council provide public access to maps and demographic and election reports for current districts and selected proposed plans through the RedViewer online mapping application. Don't suppose it complies with Federal Web Accessibility Standards, though...
![]()

