The Tempest

[David Mohring <heretic ihug co nz>]

After lurking upon this list for a while, I think it it time to
define ( if somewhat weirdly ) some concepts...

A collection of application does not a make office suite, they have to be 
integrated to the point where they share :- 
 cross application scripting ( OpenOffice does this very well ),
 compound document inclusion ( bonobo ),
 reasonably common "look and feel" GUI. 

IMO There will be four distinct classes gnome office suites that could
co-exist for some time to come ... 

Dramatis Projectus ( Guess who learnt latin from Road-Runner cartoons ) 

One host ... 

Prospero : the right Duke of Milan.
	The Gnome foundation's environment
	i.e. Gnome services, existing and future
	Including New office gnome services such as
		Virtual Enterprise Interface and Peer server
		( a user can serve many masters - see later email ).
		Cross Application Scripting

And four classes of gnome office "suite" projects ...

Spirits :Iris,Ceres,Juno etc. 
	e.g.. Gnumeric, Dia, Evolution, Gimp 
	Existing native gnome applications - together as a suite.

Sycorax : a Hag.
	e.g.. OpenOffice Koffice XEmacs-Gtk Abiword Gnucash
	Cross platform with a gnome-fied port.
	for each existing office suit and app, not native to gnome,
	get each to work within gnome and accept and serve bonobo'd 
	objects.

Caliban : a savage and deformed Slave (Son of Sycorax).
	e.g. like gmc ( the gnome-fied Midnight commander )
	A new gnome foundation project. 
	Rip source code from existing Hags and Spirits, and patch it
	together.

Ariel   : an airy Spirit.
	e.g. like nautilus 
	A more ethereal new gnome foundation project.
	Taking existing source from applications only as a guideline,
	design a new suite from scratch.


Each has pros and cons for the programmer and user.

Spirits 
	Pros
		Mostly working now
		Native to gnome 
			- smaller memory footprint
			- better performance.
		Mostly share gnome look and feel.
	Cons
		Lack functionality of Open ( and MS ) Office
		Not so integrated as a suite.
		 		
Sycorax
	Pros
		Mostly working now 
		Cross platform ( aside from Gnome ) and therefore
			larger user base 
	Cons
		May lack common look and feel.
		Multi-Platform Abstraction layer.
			- larger memory footprint.
			- performance hit.
		Little code sharing between each vendors product
			- duplication of common services 
			- foreign vendor bonobo inclusion = lots more memory.
			- many interfaces to deal with
		Not under "direct" gnome foundation control.

Caliban 
	Pros
		Under "direct" gnome foundation control 
		Could take OpenOffice and replace VCL with 
			C++ wrapped Gnome-Gtk
		Could start hacking straight away.
		Would perform better under gnome than OpenOffice. 
	Cons
		Would fork the entire OpenOffice project.
		Would the performance gain be worth the hassle?
		Could end up like "gmc".
 		
Ariel
	Pros
		Under "direct" gnome foundation control.
		Could be targeted at next generation of Gnome
		Could use gcc3.X using new standard C++ with templates
			(G++ is now using a new C++ ABI that represents 
			 classes more compactly, uses shorter mangled names, 
			 and is optimized for higher run-time performance).
		Could use new refactoring tool ( more on this later ). 
		Replace application with components :-
			Greater code sharing 
			Compound View 
			Task orientation
			Enterprise integration
			Unified Scripting interface
		Would truly be "the" gnome office.
	Cons
		Would be starting from scratch again!
		Would be duplicating the role and work of OpenOffice.


[ Please cut out above this line if just answering the following]
Which then begs the questions ...

Would it be too wasteful to develop the above projects in parallel?

Is there a real demand for a native gnome suite?

Which is worth pursuing in the short/long term?
	Spirits
	Sycorax 
	Caliban 
	Ariel


"Be collected:
 No more amazement: tell your piteous heart
 There's no harm done." Prospro, Act I, Scene II , The Tempest

David Mohring - With apologies old Will and anyone that might take offence.