Re: KDE Interop [Was: D-BUS background]

[Havoc Pennington <hp redhat com>]

On Sun, Mar 09, 2003 at 10:59:14PM +0300, Vadim Plessky wrote: 
> I am curious to know where GNOME is ahead of KDE.
> I try every major GNOME release and still can't find it useful (as of 2.2 
> version).
> On the other hand, KDE became useful since 2.1 version.
> So, I am really curious to hear your arguments.

So we can have a neverending pointless thread about it? ;-)

There is some stuff in http://www.gnome.org/start/2.2/notes/ if you
want the party line.

> |  The reason "enterprise customers" matter is that "enterprise customer"
> |  basically means "customer who deploys lots of seats and pays lots of
> |  money." And this is thus where you win significant marketshare and
> |  make significant revenue. And that's where you can start having some
> |  momentum.
> 
> I doubt "enterprise customer" in current economical environment would wish to 
> pay lots of money.

Of course everyone wishes to pay $0. However, even in the current
environment, there are plenty of people who are willing to pay money
for services and features that they need.
 
> If you think that GNOME has larger development team - *why* GNOME is (IMHO) 
> behind KDE (in terms of usability for *real world* apps, like mail client , 
> office suit, etc.)?

Thinking of GNOME/KDE as complete application bundles is archaic
thinking.  Apps should work fine under any desktop, and the plan over
the next couple years is to make them not only work but integrate
nicely, whether they are GTK or Qt apps. Nontechnical users do not
choose apps based on what toolkit they link to.

That said, I have no idea what you mean about being behind in apps.
See the poll linked from http://dot.kde.org/1046679118/ for example.
Evolution, AbiWord, Gnumeric, etc.
 
The only way to prove this either way of course is for us to generate
a huge feature matrix and check all the boxes, otherwise this thread
will just go back and forth "yes" "no" "yes" "no" so really there is
no point in following up ;-)

> well, kdelibs+kdebase have the same size as 57 packages required to install 
> Gnome desktop, while:
> a)  GNOME Desktop (panel+terminal+control_center+Nautilus) lacks Web Browser.
> 
> Adding Mozilla results in additional 15MB-20MB of disk space.
> Besides, Mozilla is not very well intergrated with GNOME (to say the truth: 
> not integrated at all), and adding Galeon results in one more package 
> installed, and it is still not 100% integrate dinto GNOME.

The web browser situation is ugly. A sane, properly-supported desktop
operating system needs to choose a single browser engine (khtml or
gecko), if not a single browser shell. Of the options:
 
 - Mozilla doesn't have native UI (Qt/GTK) and is too complex
 - Galeon is too complex
 - Phoenix is not native UI 
 - Epiphany is not finished
 - Konqueror is too complex and is tightly bound to having 
   KDE desktop running

Blah. Nitpicking about GNOME vs. KDE here seems to miss the point
given that we don't have even one fully satisfactory option.  Right
now Epiphany seems most likely to combine simple UI + native widgets +
running fine under any environment + handling most web sites, but
konqueror or other khtml-based browser could certainly get there too
if motivated.

> c) KDE's Control Center is superior to GNOME's one

Um, no. It's full of technical jargon, it's based on tree widgets and
search (both of which are well-known to be hard for nontechnical
users), it contains a ridiculous number of control panels most of
which have subtabs.

But again, following up here is pointless - you'll say "but it works
for me" and I'll say "it violates many well-known UI guidelines and is
too complex" and you'll say "no it isn't" and we'll go around until
the end of time.

I think there's some argument for a shell-type thing in GNOME, but
it'd be more like the OS X or Windows XP one than the KDE one.
  
> What version of GNOME would have usable File Selector?
> (comparable to KDE 3 or Windows 2000 File Selector in terms of functionality 
> and usability)

GTK will have it in GTK 2.4.  With GTK 2.2 libgnomeui is 90%
superfluous, by GTK 2.4 libgnomeui will be 95% superfluous, and by GTK
2.6 should be entirely so. That's approximately 7 months (2.4) and 16
months (2.6) out. In 16 months that means we have a single devel
platform with no silly stuff like GtkToolbar/BonoboUIToolbar or
QProgressBar/KProgress.

This approach is slightly slower in terms of getting all the
integration done, but it's a much better final result.

Of course you're free to disagree, and I encourage you to use whatever
desktop you like. The whole point of my original mail is that these
kinds of architectural issues, accessibility, simple UI, etc. are
important to many customers/vendors while they probably don't matter
to hackers most of the time.

Havoc