OpenSource Upgrades


2010-06-24 23:35




mapserver, opinion, web development


openlayers, opensource, upgrades





image Moving to new versions in OpenSource has in the past been like playing Russian roulette. Just now clicking on “Upgrade to WordPress 3.0” I thought would lead to a greater understanding of the database schema, template inconsistencies, and plug-in idiosynchrocies of WordPress as I met an inevitable “Upgrade Failed. Please see log for details” message.

But no! Five seconds later, all was successful, and everything looked as before. There have been numerous bug fixes and improvements, but after a brief look through the list I didn’t see any “killer” new features.

In the world of GIS however there are a number of important new versions coming up this summer.


OpenLayers 3.0 development has started. Key features and ideology are listed here. There is also a wiki page. There is a move towards separating the core functions from the UI. I get the feeling OpenLayers could be moving from a homage to the Google Maps API , into a full JavaScript / browser geospatial framework.

The core should focus on the following:

  • layer management - tile handling & vector rendering

  • map & layer related event handling

  • server/client communication - generating requests and parsing responses

  • geometry digitization & editing

A separate OpenLayers UI lib should focus on the following:

  • rendering of buttons and other widgets

  • presenting map information that is not rendered as features (layer properties, feature attributes)

  • gathering user input for map configuration, queries, etc.


On the server side of things, it seems GeoServer has recently had more traction than the older MapServer. However there will soon be (sometime in July if I understand correctly) a version 6. This has a number of new items I’ve recently required for a new project, and they are now tantalisingly within reach. These include:

  • XML mapfile consumption: run-time compilation (XML/XSLT => .map) or embedded XML parser - it will then be possible to easily generate and manipulate MapFiles from a database, or script (I’m sure these already exist, but moving to XML will mean all sorts of powerful libraries in any language can be used to do this)

  • Output formats: KML and GeoJSON ?” excellent news! I’ve tried WFS, and I don’t like it. Its nowhere near as ubiquitous as KML (a now open standard) , or as easy to manipulate as JSON.

  • Tile caching support ?” whilst it is slightly annoying that having got to grips with *TileCache* and Windows it looks like it will be integrated into MapServer, one less system component to deal with is always a good thing.



Add Comment