Re: Progress bar on startup

[Malcolm Tredinnick <malcolm commsecure com au>]

On Sat, Sep 21, 2002 at 01:03:59PM +1000, Damien Covey wrote:
> Sean Middleditch wrote:
> 
> >This can be done very easily; mozilla/galeon/open-office all use a
> >method to make their start-up times faster.  they have a
> >program/applet/daemon that is linked to their libraries.
> >
> >So, making a gqld (gnome quick load daemon) that is linked against all
> >major gnome libraries, and starting it somewhere in your init process,
> >would solve this problem.  A lot of people would want it off for
> >efficiency reasons if they are rarely in gnome.  Others may want it
> >running to speed up login.  It will slow down boot up time a bit (if
> >it's launched into the background, it at least shouldn't cause a big
> >pause - but boot up will still be slowed).
> >  
> >
> This is the sort of behaviour that I'm talking about.  I think that once 
> a user is greeted with the GDM that they shouldn't have to wait for 
> Gnome to load.  Obviously when someone starts their machine to an X 
> environment they are most likely to be using Gnome.

Slow down a bit here. It is not "obvious" at all. It may be common in a
scenario where you are the only person using the machine. However, in a
multi-user environment, people will be using many different
environments. You have to consider workplaces with basically homogeneous
setups where people log in to different machines, university-style
setups, shared houses with only a single computer (I've shared with two
other people where one person used GNOME, one used KDE and one used no
over-arching desktop environment), diskless workstations, etc. All of
these scenarios may use graphical login via gdm, but you don't want
GNOME preloaded.

Logging in is something that is done comparatively rarely, with respect
to the amount of time you spend using GNOME for other things. Crippling
a whole bunch of use cases for the benefit of the single-user per
machine case sounds less than optimal.

Malcolm