Here's a design problem you may find interesting.
Feel free to help by answering or commenting. All input is welcome.
The design goal: an easier tool for users to enter a type of location on a mountain side. The location consists of:
- The elevation band: Alpine, Treeline, or Below-treeline. These aren't absolute elevations; they vary depending on where the trees grow on a particular slope.
- The aspect: North, Northeast, East, etc, the compass direction that someone is facing when they look downhill.
The thought is that two drop-down boxes are more error prone than a graphic data-entry tool. The data is combined with other data and then used in avalanche-risk assessment, to keep mountaineers out of danger. The thought is that users are conscientious about data entry; their own safety depends on it.
The proposed solution is still at the design stage; no code is written. The proposal started with three concentric circles, one for each elevation band, which are divided into eight wedges, one for each compass points. The users will click one of the 24 options that represents the slope on which they're reporting. A similar tool was spotted in another setting, but that doesn't mean it was tested for usability or data quality.
Since this is not a common tool, it was further proposed to add skeuomorphic details: the colours of trees and snow, and the mountain's shadow. The tools dimensions are large—so each target is bigger—in keeping with Fitts' Law. Have a look at the drop-down solution (left) and the proposed tool (right) in this mock-up:
A limitation: the intended users are not available for any testing, because their work is seasonal and currently they're all away. Testing with the general public won't work because they don't know the concepts.
Two questions: Are there other ways to get this input from users? What heuristics should be in play, here, to make the proposed design better?