Tuesday, May 12, 2009

The Dangers of Templates and Models

Brady Forrest of O'Reilly Radar interviewed Jack Dangermond, the founder and CEO of ESRI about various GeoWeb related topics in 3 interviews (1, 2 and 3).
The part that really interested me was Brady's question in the first interview section where he asked:
" Do you think that the explosion of web-based mapping has just filled the world with ugly and poorly designed maps?"

Cartographic Templates: Jack agreed and answered that he thought that the answer lay in producing cartographic templates like color ramps and symbology sets for applications such as hydrology. He goes on to compare templates in cartography with the use of templates in PowerPoint, which is an interesting choice of example as I'll discuss. On the one hand, I agree with him. I've called for an improvement in the default icon template in GEarth as this would help low level users produce better maps.

Inappropriate Template use: However, there is also danger in the use of templates in that they get used inappropriately, this has been pointed out in a wicked lampoon of PowerPoint where the Gettysburg address is forced into a PowerPoint presentation. In the case of GEarth, you can imagine someone opening two data sets using different template symbologies and having trouble separating similar looking icons that were never designed to be viewed together.

Not a Complete Solution: Producing templates also doesn't answer a number of presentational problems in GEarth that relate to it being far more interactive than a simple paper map, for example, designing the places column to be clear, coping with large numbers of points and using the timeline.

Models Misused: Jack then goes on to discuss how he thinks standard computer models will become more common with people being able to use the models of others to do expert analysis. Even more than templates being misused this rings alarm bells for me. My masters degree was in Hydrogeology and I still remember our professors continuously warning us about 'taking an off the shelf model and shoveling data into it', time and time again they gave us examples where non-experts have done this and made bad decisions as a result. A simple example was the 'never ending lake', a hydrogeological model had been produced by a non-expert that seemed to suggest an aquifer had abundant water supplies that would last idefinately. When analysed by an expert it turned out the model had a badly programmed lake within it, the level of the lake didn't fall however much water leaked out of it and it was the effect of this that was incorrectly predicting that there would be abundant water available in the future.

Models used Well: Which isn't to say models for the public are always bad, when you put a start and end position into GMaps and ask it to work out a car journey route its running quite a sophisticated network model to give you a result. We just need to be sure that where expertise is required to input data into a model, only experts do it.

Norms and Principles: My own two cents worth of opinion on what would really help the red dot fever problem is that more than models and templates we need to evolve norms of use like tabbed navigation in web pages and design principles such as knowing to keep text in pop up balloons brief and to the point.

No comments: