> > Sometimes it seems to me that Chandler is a difficult application to > understand because it's too timid to be an "opinionated" application > (one that tells you how to work), but too opinionated to be a "tool" > application (one that simply gives you a uniformly-constructed model > and expects you to shape it to your own opinions). > > For example, calling it the "dashboard" instead of something like > "current work" or "for review", hides the opinion that this is where > your stuff is supposed to go. It's almost like the program is > ashamed of telling you what to do... but does it anyway. You just > have to understand the code words. :) > > For an example in the other direction, look at Hank Williams' > comments, which are trying to drag the application to the "tool" end > of the spectrum, where the user is given a uniform model and allowed > to make their own decisions about how things should work. >
This is an interesting analysis. But I would be hard pressed to think of any successful consumer (non-enterprise) application (pim or otherwise) that does not provide a uniform clear predictable model that the user can hold in her head. So you are right that I do think this is critical. As an example of why a clear user model is necessary, I challenge anyone at OSAF except maybe Mimi to accurately tell me what happens when the user goes to the "View" menu and selects "Calendar". Since I know no one can do this, I will then move on to explain that not being able to do this is not the fault of the user. It is not because she just needs more training or to get comfortable. The software design must create a clear model in the user's head of what the application is doing. In short, it must be predictable. Don Norman calls these expectations affordances (http://www.jnd.org/dn.mss/affordances_and.html) In this essay, he lays out baseline design principles that are indeed elemental. In short there is *no* debate in the UX community about the critical nature of a clean understandable user model that allows the user to predict behavior. The arguments that are being made that this is "just one way to do it" or that it is not necessary to present a consistent predictable model, or that it is "futile" to try do this and be "usable and useful" is far afield of all of the research on this topic. There is no gray area here. Not even one little bit. None at all. Hank P.S. For those of you who may not have seen it in my earlier email, this is also discussed by one of the great UI thinkers, Bruce Tognazzini here http://www.asktog.com/columns/069ScottAdamsMeltdown.html. I really hope you can read it. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ Open Source Applications Foundation "Design" mailing list http://lists.osafoundation.org/mailman/listinfo/design
