Re: Menubar (and Empty Default Desktop) Proposals for GNOME 2.10

[Ken Harris <ken bitpoetry com>]

This is kind of diverging into 3 or 4 parallel (yet related)
discussions.  So again I have 2 answers.  :-)  (Well, maybe not answers,
but points of view.)


Answer #1: empty desktop

> My original e-mail mentioned the controversial idea of an empty default
> Desktop - but perhaps in the short term a question that I would like
> answered is should we roll a "Places" menu into the gnome-panel menubar.
> My claim is that it's A Workflow Improvement (TM).

An empty desktop doesn't sit well with me.  The whole point of the
spatial metaphor is that files and folders are physical objects ... and
then we go and make everything hidden, and only appear when you choose a
menuitem?  That doesn't sound right.

One of the two principles of the spatial metaphor is "stability" (things
stay where you put them).  I think that making everything invisible by
default and hiding them in a menu weakens this (since menus, by
definition, don't stay open).  The way to make something seem like a
physical object is to make it visible, always, and make its visible
attributes (position, size, color, etc.) under the user's control.

(Also, FWIW, Apple has never shipped an empty desktop, and Microsoft
hasn't in at least 10 years, AFAIK, which hints to me that there may be
a discoverability aspect here.)

> For example: I have a browser open, maximized.

"Maximized" is by definition "I don't want to work with anything else
right now".  I don't see a problem with simply un-maximizing it,
double-clicking Home, etc.  We don't need to duplicate all existing
functionality of everything in the panel, just so people can work with
windows maximized all the time.  (Note that Mac doesn't even have a
maximize.)  (Well, they do, but it's pretty well hidden.)

Perhaps a better fix would be to add the Gtk+ "bookmarks" to the
Nautilus "Places" menu (which already exists), either directly on the
bottom or under a submenu with a catchy name like "Bookmarks" or
"Favorites".


Answer #2: cost of features

> Thus, I don't know what _application_ I should launch, so I have to go
> document-centric.  With the Places menu (and having previously
> bookmarked the Project X folder) it's just 2 clicks to the folder, and
> then I can look for the file by name.

OK, I'll grant that for some set of starting conditions, this menu would
save clicks.  But is it worthwhile?  There are dozens of things I do
every day which take me lots of clicks.  We could make each one a
2-click operation, by making the system menu bar 10 menus wide and 20
menuitems tall, but that would obviously be a bad idea.  The (useful)
question is: does the benefit of this feature outweigh its cost?  (How
often do people need to do this?  Does it confuse people?  Does it
weaken/break metaphors we want to strengthen?  etc.)

Here's another way to look at it: In what cases do you need to open a
folder?  For me, 90-95% of the time, it's the folder containing the
document I already have open.  The solution to that, then, is to make it
easy to open the folder that contains the document you're looking at --
coincidentally (ha) the Mac has a rather elegant way to do this:
command-click the title.  (I tried to implement this: see
http://bugzilla.gnome.org/show_bug.cgi?id=152697 )

See where I'm going with this?

We have an operation that's somewhat hard to do (opening folders).  The
most direct solution (which appeals to lots of programmers) is to simply
add a menu.  But (to use Havoc's terminology), adding new interface
elements has a price.

That price might be acceptable if we really need this function, and all
of the other options have been exhausted.  But they haven't.  There are
many other ways (I've mentioned two) in which the existing feature set
can be made more usable, with little or no associated cost.  <Insert
your Antoine de Saint-Exupery quote here.>

When we've implemented the nice little touches that other systems do to
be more usable, and we've come up with all that we can think of on our
own, and it still doesn't flow, then I'll be OK with considering shotgun
spot-fixes just to make it work.  But I really don't think we're there
yet.


- Ken