>You bring up some interesting points. I suggest you look at "About Face" by Alan Cooper, who also questions the whole idea of "saving." His main point is that the concept is an artifact of the file system, where there are two copies of a document: one in memory and another on disk. This doesn't fit the user's model, so he thinks the idea should be abandoned.
>

Thanks for all of your insights into this thread. It has been very good. I am going to get a copy of About Face. It sounds very worthwhile.


I was reading some comments from D.A. Norman. He was talking about people centred views and machine centred views. We tend to create interfaces that are machine centred and not people centred, which tends to cause problems. His discussion related to having locked your car keys in your car. The concept here is that leaving your keys in your car is a detail that you shouldn't have to worry about. Maybe you were thinking about more important things than keys. The problem here is a design fault on the part of the car, not your stupidity!

Some new cars now beep if the keys are left in the ignition, or don't allow the door to lock or even require the key to lock the door from the outside. People centred design changes.

He had an interesting matrix:

For the machine-centred view
People are Machines are

Vague Precise
Disorganized Orderly
Distractible Undistractible
Emotional Unemotional
Illogical Logical

For the people-centred view
People are Machines are

Creative Dumb
Compliant Rigid
Attentive to change Insensitive to change
Resourceful Unimaginative
Able to make flexible Constrained to make
decisions based on consistent decisions
context

My focus in this discussion is to question if what we are doing and the models we are using are the best ones. Can we come up with better ones? The notion of saving is just the seed for discussion. I liked Craig's comment about doing automatic saving. My question in a way was 'does saving need to be user task?' There was a time that I always, unthinkingly, provided an edit button to place the user in edit mode. Lots of applications are like. But I have come to realize that in a good design it is unnecessary.

I find that users are sometimes very tolerant of systems. If you give them a new system that improves on the old, they don't complain much, even if a more elegant solution could have made their work load so much easier. I think our systems should be partners with the user in seeing that information is collected in the best possible way that satisfies not just the database/system requirements, but also the users needs.

The complexity of development will be reduced as we understand the requirements more and more. We will never produce better systems if we think we are producing them now.