This sounds like a very interesting idea (to me) if the goal was to use it
as an applications framework, where everything in an application is an
object that can be manipulated using ruby. If that were true, How would
persistent application data (the contents of a word processing file for
example) be stored? Here are some ideas:

1. Store all application data as xml documents. One per object (Since ruby
excels at text manipulation)
2. Store the xml documents in source code control system to do object
revisioning
3. Use ruby as an "explorer" over all these objects that exist in the
system. Instead of a hierarchical files system. Develop a query based system
where folders are only a the results of an object query.
4. Allow any data to be added to any object for later classification. For
example when you receive an e-mail, it becomes an e-mail object. If it
concerns ruby you could tag it with a content-type attribute with a value of
ruby. If it concerns a bug report then you could add to it a task-type
attribute as review-bug-report. So if you had a folder hierarchy like:
	Root 
		Ruby
			Tasks
				Review-Bug-Reports
		Tasks
			Review-Bug-Reports
				Ruby

The email would appear in both of the deepest in folders. Ruby could be used
for the query language.
5. Add in text indexing to every object in the system so that you could find
any object in a hurry by searching for text.
6. By using the SCCS, you could view any object's history and branches off
of that history
7. Sub-class any object so that you duplicate as little effort as possible.
So for example in a word processor if you had a paragraph style (which would
be an object) you liked to use with some changes, then you would sub-class
it to create a new paragraph style with the changes you want in it.  
8. Every object would have an object creation date so you can easily know
which objects were created most recently (a system wide history)
9. Every object could be tagged as a favorite through the object addition
facility.

Oh Well just some thoughts,

Steve Tuckner

-----Original Message-----
From: Michael Neumann [mailto:neumann / s-direktnet.de]
Sent: Friday, February 23, 2001 12:11 PM
To: ruby-talk / ruby-lang.org
Subject: [ruby-talk:11395] Re: trial balloon: Ruby desktop?


On Sat, Feb 24, 2001 at 01:18:44AM +0900, Jon Aseltine wrote:
> Hi,
> 
> I have been thinking about something for some time now, and I wanted to
> get some feedback from the list. Would it be feasible/possible/wise to
> craft a desktop environment using Ruby as its backbone? I am quite
> unhappy with the state of both GNOME and KDE, and I find the idea of a
> Ruby desktop exciting. Some parts need to be fast (some window manager
> stuff, e.g.), but other things, including most application logic, could
> be in pure Ruby, I think.

Great idea. 
IIRC, Andy/Dave worked on a window manager written in Ruby. Am I right?

But KDE and Gnome both come with a GUI toolkit, KDE/QT and GTK.
Which one to choose?
QT cannot be used under Windows, due to the license.
Tk is to slow and ugly.
GTK, Fltk, FOX?

And what's about the component model ?
KDE uses its own, as far as I know, and Gnome uses Corba.
dRuby would be nice.


> Is anyone else interested in this? Given Ruby's many advantages, I think
> something useful could be crafted quickly (a year? is that quick?). I
> also think it would push Ruby to the limit in some ways, and perhaps
> make the language better.

> Any thoughts? Or am I crazy?

Do you consider a framework for applications or "only" an application which
shows a desktop with icons which runs programs.



--
Michael Neumann