Re: [Usability] spatial nautilus options (was spatial nautilus concerns)

[David Feldman <mailing-lists interfacethis com>]

>  The main thing that you
> one must remember when dealing with a spatial model, is that there
> cannot be an ``up'' button, because there is no ``up'' to go to in a
> spatial model.

But doesn't that seem a bit contrived? I mean, there's clearly an "up" 
in the filesystem. I think one thing that concerns me about a pure 
spatial model is that it attempts to hide the filesystem structure in 
some ways. This forces more of a barrier between novice and expert 
users. If a model exists that can make the filesystem structure plain 
and comfortable for users, that seems like it'd be better because it 
gives users a mental model of their filesystems that's accurate. This 
is one of my concerns with Windows: The Start Menu and desktop icons 
are fairly easy to navigate but once you hit the filesystem itself it's 
not only fairly difficult to get around, but completely unrelated to 
the Start Menu and desktop icons.

> So for those of you that argue that the default behaviour, in a
> spatial environment, should be to keep everything in one window, are
> not craving a spatial model at all, but a navigational one, which has
> already been pointed out by several other people in this thread.

But this is what I was getting at before: I don't think it's accurate 
to delineate a "spatial model" or a "navigational model" as two 
separate options. It's a continuum with a number of features that can 
be implemented either way. Open-in-same-window is an aspect of the 
navigational model but applying it in an otherwise spatial context 
doesn't have to invalidate the other benefits of that spatial context.

I haven't stated it outright but it's probably pretty clear that that's 
what I favor. I think the principle underlying spatial model is a good 
one. I also think there are some drawbacks (like tons of windows 
onscreen), and some may outweigh certain aspects of the spatial model. 
So, it seems like a carefully-considered hybrid is probably the way to 
go. I think Apple's got it mostly right. Try logging into a Mac OS X 
box without tweaking any preferences and you'll see it _almost_ follows 
the spatial model but that it's pretty easy to avoid the screen clutter 
without knowing modifier keys.

> I disagree with the person you replied to here.  I don't believe that
> the spatial desktop metaphor is about not thinking about how to get to
> your file; it's about knowing where your file is contained, but not
> always thinking about where the container is contained.  The file is
> in the container, and that's that.  <snip>

As I said earlier, I think we should be cautious with our metaphors on 
both sides, because (a) in the end, the computer desktop is _not_ the 
same as the physical desktop and at times it may make sense for ease of 
use and/or efficiency to do something that doesn't reflect the physical 
world; and (b) we can all come up with metaphors that support our 
arguments. We don't know for sure that any one analogy is correct, 
though they can seem convincing. The simple fact that we have so many 
analogies to support both sides should make that clear.

--Dave