User Interface

[Chris Harris]

On Thu, 22 Dec 1994, Jecel Mattos de Assumpcao Jr. wrote:

> I don't think that the current window-based standard will last
> more than five to six years. It does not make the best use of
> limited screen real estate and is not well suited to the future
> generation of Virtual Reality computers.

I'd like to second this.  Therefore, I'll probably be advocating a sort 
of UI that is quite flexible, so we can experiment with all sorts of 
ideas.  Keeping them all (or as many as possible) compatible, or at least 
somewhat so, would be desireable too, but won't always be possible.

> It is a very good start, though. One thing that I think is very
> important is what I call "transparency". I don't like for the
> system to be built one way and the UI in another with a big, fat
> translation layer in between ( Windows/DOS or Motif/Unix are
> examples of what I am talking about ). I believe that the system
> should seem to the user to be what it actually is. You should
> learn deeper versions of the same ideas as you go down levels,
> not totally different ideas.

I'll agree with this, although such "transparency" can sometimes be 
useful.  For example, if I'm dealing with a large spreadsheet 
application, I may not want to see that each cell is really an 
object, and have to send indivual messages to them.  I guess 
hiding this isn't so much "transparency" as simply dividing things 
into components....  Hopefully, the underlying system will be 
powerful, flexible, and easy enough to represent without transparency....

> There are many neat ideas out there. The current Windows users
> might resent any big change, but they'll have to face it sooner
> or later ( some command line die-hards still denounce the evils
> of windows :-)

This is definitely true, and so we ought to support anything that'll come 
along.  Some apps are also better suited for different interfaces.  It'd be 
nice then to support several of them at once.  On one system, for 
example, the object browser might be a 3D interface (similar to the idea 
behind the Merlin interface perhaps), but you could still open up a 
database app in 2D (perhaps on the surface of a cube).

The danger here, of course, is under-standardizing, like they've gone and 
done in X-windows.  I believe that each interface should have standards 
to conform to, but there might be many different interfaces (and 
standards to tie them all together.)

> 
> -- Jecel

Anyone have any ideas for goals of our "UI Toolkit", as we might call it?  
I've mentioned some here, and I've got more.  Just gotta write them down 
in some fashion.