Re: [dwm] Again client titles and nmaster indicator

From: Anselm R. Garbe <>
Date: Mon, 15 Jan 2007 14:11:53 +0100

On Sun, Jan 14, 2007 at 08:59:57PM +0100, Antoni Grzymala wrote:
> Tako rzecze Giorgio Lando (w e-mailu datowanym 2007-01-14, 13:38):
> > Oh, I understand only now that it will be still possible to read the title of the
> > focused window in the status bar, since only the titles of the unfocussed
> > windows will be removed. Then my complaint above about screen hardstatus is
> > largely ungrounded and I do not complain anymore.
> That means I have to switch focus 9 times if I have 10 windows and want
> to read their titles. Instead of just looking. This is *MAJOR*
> unexplainable crap.

No. As I already pointed out, in most cases you don't identify a
window by its title. Instead, you identify it by its content.

If you have 10 windows open, you don't need to read all titles
to lookup the window you want to pay attention next, you simply
preselect relevant candidates based on the window contents
first. And if there are any doubts you only check the titles of
those candidates next - mostly you have no doubts at all and
focus the correct candidate in the first step.

The reason for this is, because the window contents are much
more descriptive and recognizable than an abstract window title
telling you 'antoni_AT_localhost: /tmp'.
Much more important than such a title are the terminal contents?
Which commands have been invoked last? Which file contents are
presented in the editor buffer?

When I develop source code, I don't identify a window by its
file name, I identify it by the content. I easily see depending
on the function name to which file such a function belongs.
In C this decision is much harder than in OOP languages for
example, because in OOP file names normally correspond to the
specific class which is edited.

> I cannot understand why a config value cannot be added to inhibit
> drawing of the titlebars if someone doesn't like them.

Because this is against the dwm philosophy and it introduces bad
style. Once you add compile-time options using CPP for a
specific feature, it won't last long that people request such
options for other things as well. Hence I prefer strict decisions
to remove a feature which is not essentially, like client title

The software philosophy I follow is, to create usable software
which only provides the essential functionality, which has been
identified during its evolution. That might be different to
other philosophies like that found in mutt - there features have
been added during its life-time and never removed. That's why
developing software like mutt is so depressive.

> @arg: I understand this is your own window manager, but then why do you
> consult your decisions? You do as you like anyway...

I consult because I have been proven wrong quite often.
But sometimes I notice that my decision is better for the future
of some software, even if there are complains.


 Anselm R. Garbe >< >< GPG key: 0D73F361
Received on Mon Jan 15 2007 - 14:11:53 UTC

This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 14:34:56 UTC