Re: [Usability] SM UI plan

[Havoc Pennington <hp redhat com>]

Gregory Merchan <merchan baton phys lsu edu> writes: 
> How would the session be chosen in the greeter window without disclosing
> information to an unauthenticated user?

Yeah, that bugs some people. (kdm ignores this issue and does it
anyway, FWIW)
 
> I hope constant autosave wouldn't be necessary. Part of what I mean by
> transparency is that the user never has to save the session; if the user
> does nothing directly affecting the session, then what he has when he logs
> in is exactly what he had when he last logged out - modulo any thing which
> occured between requesting a log out and the actual log out, such as saving
> files from application shutdown prompts.
> 
> Was enabling this the cause of the complaints?

Yes. People want to set things up in some canonical way, and remember
that; they don't want their random clutter remembered all the time.

> > This is probably right for maintaining uniqueness of various desktop
> > pieces (I've proposed it on xdg-list already I think), but an SM
> > property would be much more convenient if you want the SM to warn
> > about missing components.
> 
> Point of pedantry. Whose convenience and at what cost?

The SM's, and no cost I can think of. It also makes this more reliable.
 
> I don't know how we'd do this in a friendly manner without adding to wm-spec
> and placing responsibilities of the session manager into that seems anathema.
> I was just casting my vote for the SM using libwnck and against doing it in
> the window manager.

The main disadvantage of doing it in the SM is that suddenly your SM
is using a lot more RAM/CPU and you have a big hunk of SM complexity
that it didn't currently have. While the WM already has the window
information.

Havoc