Re: [tlug] Running without Gnome/KDE/xfce/whatever. (was: Ubuntu 16.04-LTS Japanese Text Input)

["Stephen J. Turnbull" <stephen@example.com>]

Curt Sampson writes:

 > 1. What does one use for a session manager?

Many independently developed window managers include session
management facilities (ie, they support the manager side of the "save
yourself, we're going down!" protocol).  Do you need more?

 > Given that I can start up everything I need in fvwm, I'm not sure I
 > really need a session manager, but it sounds like if I'm going to go
 > this route, I have to have one.

You don't need one if your apps normally save their own state, and you
don't care if the running apps persist across reboots.

 > I could be extremely happy with systemd (it's a very well done
 > piece of software); is anybody using that for user sessions, yet?

I'd never heard of it -- after all, it's really designed for operating
system services rather than user applications.  But I don't need user
level session management -- *I* am my session manager.

 > 2. I notice that my /etc/alternatives/x-window-manager is linked to
 >    fvwm2. I'd not thought about this, or how it happened, but I
 >    wonder what that means if I'm on a system I share with others
 >    who do not also share my...predilication...for certain types of
 >    user interfaces.

You'll need a personal ~/.xinit or ~/.Xsession, and you'll need to
hard-code your preference for such applications (or you could create a
~/.local/etc/alternatives, but that seems like overkill), that's all.

 > On 2016-04-27 21:14 +0200 (Wed), Attila Kinali wrote:
 > 
 > > Instead authors make many assumptions about how your system looks
 > > like and lable it as user error when it those assumptions do not
 > > hold. :-(
 > 
 > When did the majority of software developers _ever_ work without these
 > sorts of assumptions?

Never; as you point out, most are wage slaves, with the emphasis on
the bondage part.  However the Unix philosophy pushes in the direction
of minimal assumptions, as does the Zen of Python (for two examples
among many).  Perhaps you recall this graphic:

http://www.visualcomplexity.com/vc/project.cfm?id=392

At the time, that was an aspect of open source we were very proud of.
Then came KDE and GNOME and welded-shut hoods to protect the wee users'
widdle fingas from the moving parts, and the complexity got stashed
away in UIs that hackers don't need....  The original flow of open
source was in the direction of the Unix philosophy, as much for
pragmatic reasons (when Stallman was the only free software hacker
left at MIT AI Lab and so was motivated to start GNU, he had to make
each utility flexible as possible to maximize its usefulness in the
next task) as philosophical ones. 

 > Most software developers and their managers, like most humans, are
 > inherently hugely optimistic, and so thinking about failure modes
 > is not something that they do well.

Who needs to *think*?  Just observe your own pratfalls and listen to
your users. ;-)  The big problem (outside of the fields where 6-sigma
reliability is a real requirement rather than a buzzword) is that they
*ignore* observations of failure modes when the problem can be pushed
onto "user training", not that they fail to anticipate them.