Re: GNOME and .NET

[yakk-gnome-hackers yakk net]

On Sat, Feb 02, 2002 at 09:06:20PM +0000, Michael Meeks wrote:
> Hi Ian
> 
> On Fri, 2002-02-01 at 20:30, Ian McKellar wrote:
> > The failure of the Bonobo experiment should make us cautious about
> > adopting Microsoft technologies too quickly in the future.
> 
> 	Please do explain the failure of the Bonobo experiment.

I knew that would provoke a response from someone - well, I knew it would
provoke a response from you :)

I just keep hearing about applications that are having bonobo removed from
them in the port from the gnome1 to gnome2 platforms. Bonobo seems to make
APIs harder rather than simpler (take a look at the proposed API for
Jacob's bonobo-based file selector and tell me thats simple). Bonobo seems
to add slowness to applications (silly stuff like base64ing every ui
string all the time seemed to be at the top of Nautilus profiles last time
I looked). Bonobo is to blame (according to applications developers I talk
to) for strange UI inconsistencies (notice how the keybindings in the
menus in Evolution's compose dialog are wrong if you've customized the
keybindings for the gtkhtml component through the capplet). All this makes
me think that Bonobo isn't working out for GNOME.

The other aspect of the failure seems to be personal/political. Bonobo
doesn't use GNOME function / object naming conventions. Bonobo doesn't
(last time I checked) use standard GtkObject/GObject ref/unref functions.
Bonobo based UIs don't support the standard GNOME way of assigning
accelerator keys to menu items. There was the Not Invented Here
bonobo-conf fiasco.

I don't know if that makes it a complete failure but it seems that theres
some serious problems with the technology at least and perhaps other
problems too that are stopping people adopting it.

Perhaps as a "core gnome module maintainer" I should give an interview to
The Register about my opinions on Bonobo's future in the platform, but I
thought I should bring it up in an appropriate forum first.

Ian