On Mon, Sep 11, 2006 at 01:46:57PM +0200, Anselm R. Garbe wrote:
> On Mon, Sep 11, 2006 at 07:33:49AM -0400, Doug Bell wrote:
> > Anselm R. Garbe wrote:
> > > I propose following strategy. I'd like to wait until October, if
> > > Kris will re-appear till then, everything is fine. If not, we
> > > need another maintainer for wmii. I won't do the job again,
> > So you refuse to maintain wmii and there is no visible replacement
> > maintainer, but you are still directing it's development? Do you
> > expect this to work?
> As long as wmii is hosted on my server it has to work ;)
>
> > > because from my POV dwm is the way to go.
> > I think that some of dwm's simplifications are worthwhile. For example,
> > I never thought that the 9P support in wmii was worth the downsides of
> > slower operation and so many race conditions.
> >
> > But I disagree with the decision to eliminate run-time configuration
> > files in dwm. Yes, I could edit "config.h". But why should I have to
> > recompile and re-install every time I change a setting? Also, this
> > scheme does not fit in well with multi-user systems.
> >
> > I've written configuration file parsers before. It isn't that tough.
> > Please re-consider this decision.
> >
> > I think that the right direction lies somewhere in the middle ground
> > between wmii and dwm.
> This is UNIX, and it has tradition to customize software in its
> source code. dwm is dedicated to point out this tradition.
>
> Is it more time-consuming to edit a config.h file and
> recompile the whole source code in 3s than learning yet another
> config file syntax and editing a config file? I think the 3s
> overhead are not an issue. Besides this, they keep out some
> useless complexity from the source code (eg parsing shortcuts,
> rules etc.), the program would only process one time at the
> startup.
>
> Also, you should settle with a configuration which can be used
> for quite a long time. If you really can't stand the way as it's
> done, it looks trivial to add yet another 200 LOC as a patch to
> write a config file parser or whatever. But I don't plan to do
> that (if you or someone else really considers this, the
> simpliest way would be writing the configuration to
> standard input and defining a special prefix for status
> text, e.g. 'status:').
>
> But the main idea behind dwm is, that you can do source
> modifications easily to let the wm fit your needs. You can
> change how it arranges windows, or add extra special functions
> and bind keys to it, without learning yet another programming
> language. That has been the intention, and it's also the
> intention for other upcoming tools like st.
> At a bare minimum, there is config.h for the most common
> modifications.
The main target group of dwm are developers and advanced users
anyways. Normal users will stick to fluxbox or kde.
Regards,
-- Anselm R. Garbe ><>< www.ebrag.de ><>< GPG key: 0D73F361Received on Mon Sep 11 2006 - 13:49:47 UTC
This archive was generated by hypermail 2.2.0 : Sun Jul 13 2008 - 16:13:37 UTC