What IGS thinks about the Linux Desktop

[Aaron J. Seigo]

i'm moving this to what i think most consider to be a more appropriate list 
for this discussion?

On Sunday 23 November 2003 07:04, Matthias Ettrich wrote:
> > so you're saying that using non-KDE apps in KDE isn't as good as using
> > KDE apps in KDE? holy stating the obvious, bat man!
>
> The discussion is about the Business Desktop. The main choice here is the
> applications. Applications first, desktop second. If an enterprise
> standarizes on OOo, Mozilla and Evolution, there is little benefit of using
> KDE. In fact - as several people pointed out on this list - , there are
> real disadvantages over XD2.

to be perfectly honest, if a company uses OOo, Mozilla and Evolution there is 
very little benefit in using Linux as a desktop as it becomes quite apparent 
that in such a configuration Windows XP and MacOS X are both lighter and more 
coherent environments.

you, and the others, are perfectly correct IMHO that:

 o we need to eat our own dogfood. this way we can know what we need to 
improve and to provide the incentive to improve it [1]

 o deciding to go with a varied bunch of apps written with 3 different 
NIH-syndrome-driven non-KDE library stacks removes much of the strategic 
benefit that arises from KDE as a desktop environment.

in fact i'd go one step further: choosing 3 different libraries counteracts 
much of the strategic benefit that arises from choosing *Linux* as a desktop. 
why do you think Microsoft is thumping on their "integrated application 
stack" drum so loudly these days? because in this case they are absolutely 
right.

KOffice (or a properly KDE-ified OOo.. haha) + Konqueror + KMail stands a far 
better chance of ever getting close to the MS trio than OOo+Moz+Evo ever 
will. KDE _is_ the best thing the Linux desktop has going for it, whether 
people realize that or not. and i'd say that those who use Linux DO realize 
it because it's the most widely used desktop environment on Linux. it's the 
marketing folk and new corporate adopters that the industry is in danger of 
misleading to everyone's detriment.

> I really don't want to start a flamewar with you

me neither, of course =)

> , but if that is stating the obvious, why do people shout out
> loudly if businesses go for XD2 rather than accepting a valid business
> decisions?

because as an industry, most of the Linux desktop companies have their heads 
crammed snugly up their posteriors, IMHO. they are making some amazingly poor 
decisions and are following up on those decisions by marketing the Linux 
desktop in directions that aren't strategically sound.

KDE as an Open Source project DOES get it, probably because we're driven 
largely by grass roots requirements: a user has needs and we strive to meet 
them. your own initial KDE announcement[2] really highlighted that focus 
well. the Linux desktop industry hasn't arrived at the same point of 
understanding that the various truly open desktop projects (e.g. KDE) have; 
this is a gap i wish we could help them close, but i'm not sure how to.

one of the best and perhaps only things we can do is to make KDE ever more 
deployment-ready. we might want to start compiling a list of what that would 
mean somewhere. 

> > > Many KDE users don't even use KDE, they use xterms, emacs ghostview,
> > > xchat and  mutt - and of course Ooo and Mozilla.
> >
> > two words: bull shit.
>
> The sentence is part of my summary of Dirk's mail. My mistake was to write
> "users", it should have been "developers". Dirk was pointing out that we
> should eat our own shit.

ah.. yes, i agree with you then..

> Maybe it's just us "old" guys, I don't know. I use emacs rather than
> kdevelop, Harri reads mail with Netscape, Simon uses xmms, XChat is seen
> quite often, and a wide usage of OOo would not surprise me.

it isn't just the old guys... those who have been around often use the apps 
they've used for years; that's understandable (but not good, IMHO, if you're 
a KDE developer). there have been areas of functionality KDE simply hasn't 
covered to date, removing the choice altogether (JuK and Kopete covered two 
such areas for me). the third big reason is how distros choose default 
applications, which may have made sense 4 years ago but make less and less 
sense with each passing release.

> > > If that is KDE's culture, that's what it is.
> >
> > it may be how you experience KDE, and that's fine. it isn't how everyone
> > does. in fact, i'd say that mode of operation being the common viewpoint
> > is on the way out. good riddance.
>
> I want to believe you, I really do. The hostility I sensed when e.g. Stefan
> started working with the GStreamer guys or when Havoc talks about
> FreeDesktop.org tells me a different story.  

well, we aren't always perfect, that's true[3]. there will always be 
individuals who lean towards the conservative; and it will always be easier 
to garner support for the exportation of KDE technologies than for the 
replacement of KDE techniologies via the importation of external 
technologies. this is just human nature, and the two examples you provide are 
good examples of that. but thankfully those aren't the only examples:
 
  o KDE's accessability will almost certainly happen in cooperation with the 
previously rather GNOME-centric ATK toolkits. this is due to a few 
individual's pragmatic and enduring involvement with the Free Standard's 
Group accessability workgroup.

 o DBUS: inspired by DCOP, probably DCOP's eventual successor and enjoys 
involvement from KDE people. but DBUS is also set to be used quite widely 
outside of KDE and has developers from other projects involved.
 
 o KDE's adoption of various FD.o "standards", which is not a KDE-centric 
organization. some have been rather invasive such as the .desktop reorg for 
vfolders

 o JuK uses the KDE GStreamer bindings

i'm sure there are others, but you get the point. KDE does not suffer widely 
from NIH syndrome, as one might be led to think if they focus only on a few 
negative examples.

> I hope you are right, or that you will be.

i suppose there are a limited number of ways to make sure that i am =)

> So what has been done that made it easier for other development communities
> to make use of the features that the KDE desktop system provides?

not enough ... then again, we've only really recently arrived at a point where 
the features that the KDE desktop system provides actually exist in a 
relatively mature state...

> Is there e.g. a little C-API that - via DCOP and a desktop process - offers
> things like our file dialog, network transparency, desktop communication
> and access to the print system? Starting kprinter is not really sufficient,
> I want to be able to customize the print dialog (e.g. with a bit of ui-XML)
> and get a socket back where I can write my postscript into.

this would indeed be very useful and very cool. 

some of this will likely be a lot easier / more realistic to accomplish once 
the move to DBUS takes place and once Qt4 is out and about. perhaps this 
should be one of the primary goals going into KDE4, then: exposing KDE's 
capabilities to non-KDE applications.

things others mentioned elsewhere such as the mimetypes would be "fixed" by 
haivng a FD.o standard.



[1]  personally, i run KDE from CVS as my day-to-day desktop both from my home 
office and my office downtown. i keep KDE 3.1 around for disasters, but CVS 
is generally good enough for day to day usage as far as i'm concerned.

[2] http://lists.trolltech.com/qt-interest/1996-10/thread00134-0.html

[3]  i had issues with Havoc's FD.o talk at first because it was framed in a 
rather context-innapropriate manner. when the point of the talk was 
clarified, i had no problems with it. for me at least, it wasn't NIH, it was 
an appreciation for the context.

-- 
Aaron J. Seigo
GPG Fingerprint: 8B8B 2209 0C6F 7C47 B1EA  EE75 D6B7 2EB1 A7F1 DB43
"Never weary, never dispair, never give in" - Winston Churchill