Re: Decorations (again)

[Paul Warren <pdw ferret lmh ox ac uk>]

On Fri, 12 Nov 1999, Michael ROGERS wrote:

> >I suggest that what we really need to do is get a version of this spec
> >finalised and get some implementations of it.  We're never going to get it
> >100% perfect first time, so I reckon we need to get something released and
> >get some feedback.
> 
> I agree. I think the only major outstanding issue is pager/tasklist
> interaction. My feelings are as follows:
> 
> The simplest place to implement a pager/tasklist is as part of the window 
> manager (possibly an optional module). Most window managers provide one.

Agreed.

> Users do not switch window managers often, so a tasklist which remains the
> same when the user switches window managers is not a big advantage.

True.

> Window managers which can operate outside the DE do not *need* to communicate
> with the DE's pager/tasklist, since they already provide equivalents of their
> own, usually better looking. (Although they might still *want* to use the
> DE's version).
> 
> Windows managers which are specifically designed for the DE (ie KWM, future
> Gnome window managers) could use a CORBA interface to export window management
> functions.
> 
> Therefore I think an *optional* CORBA interface is the best way to communicate
> with an external pager/tasklist. 

This has always struck me as reasonable - WMs aiming to work with DEs
should have no problem with including CORBA support - like any other DE
app.

> However, if I'm the only one who still has a
> problem with this, I'm happy to go with the current proposal if it means we
> will get the spec out the door.

Scrapping the existing method does mean a lot of work all round.  WMs that
don't already have a pager (sawmill springs to mind) will need to acquire
one.  Someone needs to come up with a CORBA interface, although I suspect
that that can be left out of the intial spec.

However, I have always thought that separating the WM from its pager
wasn't a great idea, and that trying to make a one-size-fits-all universal
pager will always lead to feature/usability shortcomings.

The only problem is that any WM may require several different flavours of
pager - eg. an applet for Gnome, a standalone variant, a KDE variant.  

Anyone else have an opinion on this?

Paul