Fedora Core 7 - Compiz vs Beryl

[alan]

On Fri, 5 Jan 2007, Peter Gordon wrote:

> On Thu, 2007-01-04 at 21:20 -0700, Trever L. Adams wrote:
>> And as of the last version, compiz had a horrid configuration (read
>> almost none) for the different plugins, etc.
>
> Nearly if not all of it is through GConf. While no graphical preferences
> capplet or similar exists for it yet, the keys are listed in detail in
> gconf-editor and modifying them gets you the results you need/want.

Saying it is in gconf is like saying to a Windows user that "You can just 
make changes in the registry to configure it".  It is a non-obvious 
configuration option.

>> That needs to be fixed. I like the GREAT configurability of Beryl.
>
> That configurability comes at the cost of a disastrous UI and many cases
> of Not-Invented-Here syndrome (using flat files for storage instead of
> GConf, using its own themes instead of those of Metacity/Cairo/GTK, et.
> al). That, and the fact that the people of Beryl is purportedly much
> less strict about the quality and consistency of code accepted.

I am not certain why gconf is a "better" solution.  It makes moving users 
from one version to another difficult.  (More than once, I have had all 
the configuration data tied up in the gconf blob.  Not easy to say "pull 
all the configuation for app x".)  Gconf is just repeating the mistakes of 
the Windows registry without the extra obfuscation of uid keys.

Compiz themes are different from Metacity themes.  They just added to 
them.

As for the code quality, I have not investigated that (yet).

I do know that I have been able to get Beryl to work consistantly. With 
Compiz, I am lucky to get the window decorations to come up most of the 
time.

> I've got three words for that: No thank you.

To each his own.  I would prefer to see both.  But looking at both 
projects, the interesting and fun stuff all seems to be going on in Beryl, 
not in Compiz.

-- 
"Invoking the supernatural can explain anything, and hence explains nothing."
                   - University of Utah bioengineering professor Gregory Clark