Re: window-in-window MDI

[Aaron Lehmann <aaronl vitelus com>]

On Sat, 18 Mar 2000, Chris Jones wrote:

> Specifically, what is actually so wrong with it?

I realize that many of the arguments so far have not been very
constructive, so I will provide a few ideas that I think have merit:

1) Bloat. 90% of users will not want anything to do with WIW, or just
prefer something else. Regardless of their MDI mode selections, this code
will REMAIN IN MEMEORY WHENEVER A GNOME APP IS RUNNING. My understanding
is that it may be paged out to disk but the whole library must remain in
some form of RAM if it is being used. Comments from somone more
knowledgeble are welcomed.

2) Bad image. Is choice the best attitude towards design? I think this
needs some rational discussion rather than just "choice is good". If you
allow a user to select every single aspect of GNOME, down to what buttons
look like, how dialogs are arranged, and what pitch the keyboard bell
should sound at, this is likely to overwhelm newbie users. I hope that is
not what you want. I am not opposed to choice, but who is GNOME catering
to? If the goal is to get people to switch from windows or macos, days of
going through configuration windows is probably not something that a
newbie user would like to do, but would probably feel inclined to do, not
knowing that the defaults are great for everyone except hackers who really
want extreme customization. I am not sure what the best compromise is
here, I'd welcome opinions. But I don't think it is right to have a system
where if someone has an idea for gnome they can include it as an option
regardless of whether it is a good idea or not. Apple and Microsoft
introduce new UI concepts often. And if the new concept is bad, the press
is going to say "This new feature sucks." WIW is universally agreed upon
as a bad idea among people who have researched usability, and I think I
can sum up why:

	1) Complexity. Rather than just having windows, you have a
heirarchy of windows. It's more confusing to a newbie that way (and I
imagine that most people interested in WIW would be newbies)
	2) It wastes screen space. Makes it less convienent
	
3) Users are not the best people to ask for UI design. Just because a user
asks for something does not mean they would actually use the finished
product. I don't think that it's worth the critisism and bloat that will
result if every idea that a user sugests and a hacker codes up, regardless
of the quality or usability, is included.

4) WIW makes a bad role model. If programmers have there settings set up
in a way that WIW MDI is used, they may model their apps to look like/work
well in that MDI interface, forgetting that it is a _choice_. Yes, idiocy
is not in our control, but we shouldn't give these people ideas ;-).

Hope this helps. If I think of anything else I'll post again.


Aaron Lehmann