On Wednesday 17 January 2007 01:03, John Plocher wrote:
> Stefan Teleman wrote:
> > There is currently no place where "Things that align well with
> > the kernel and/or the core utilities, that demand or require
> > tight integration" can happen for KDE Solaris,
>
> Consider me to be ignorant; you won't be too far off ;-)
>
> The proposal says "collaborate with the KDE e.V. and kde-core-devel
> in order to integrate KDE as an OpenSolaris project", yet the
> solaris.kde.org web page gives the impression that that is *its*
> charter; furthermore, there already is an OpenSolaris Desktop KDE
> community with what looks to be the same charter yet again. So my
> question is "what will this new project do that is different from
> those other two"? Why do we need three communities/projects to do
> this?
solaris.kde.org is a web site hosted by KDE e.V. and dedicated to KDE
on Solaris. As a side note, Solaris is the *only* commercial OS to
have its own web site hosted at KDE. Its existence predates the
existence of OpenSolaris. KDE e.V. also pays for hosting patches and
packages for KDE Solaris, for the maintenance of the boxes which host
these packages and patches, as well as for the bandwith for
downloading these patches and packages. As everyone may or may not be
aware of, KDE is, and has always been licensed under LGPL/GPLv2 (with
a limited number of exceptions in some translation units). KDE's
primary development OS is, and has been since its inception,
GNU/Linux.
KDE is predominant in the EU. It originated in Germany, and it is
closely aligned with Trolltech.
The source code downloadable from ftp.kde.org is, as you say, a
baseline, even for GNU/Linux distributions. Distro-specific patches
exist for SUSE and for RedHat (those are the ones i know of, there
may be more).
Given that GNU/Linux is the "reference implementation" of KDE, it is
virtually impossible, with the minute number of KDE developers
currently involved with KDE Solaris, to provide portable patches,
when these become necessary because of incompatibilities between
Linux, FreeBSD and Solaris, or between compilers, or simply because
better implementations exist. Common, and correct solutions *do*
exist. However, there simply is no time to write, and test such
OS-transparent artifacts, which would work equally well, and equally
transparently, on GNU/Linux, FreeBSD and Solaris, with GCC and
SunStudio. As such, KDE Solaris patches are, and have been relegated
to just that: KDE Solaris Patches floating around on ftp servers.
What this Project will achieve:
[0]. Provide a unified location for the source code of KDE Solaris.
[1]. Given [0], provide a "reference implementation" of KDE Solaris,
similar to the OpenSUSE or Fedora Projects. This reference
implementation would be the result of collaborative development work
under the OpenSolaris KDE Project.
[2]. Given [1], create a foundation for commercial distributions of
KDE Solaris. There are currently in existence 3 commercial
distributions of KDE (that i know of):
- Xandros [ http://www.xandros.com/ ]
- Linspire [ http://www.linspire.com/ ]
- Tomahawk Computers [ http://www.tomahawkcomputers.com ]
This would open the possibility for mixed GNU/Linux and
OpenSolaris-based KDE installations. Currently, this is simply not
possible, and not even considered.
[3]. Also given [1], create a baseline foundation for currently
existing OpenSolaris-based distributions (Belenix and Nexenta). So
far, patches for KDE Solaris have been distributed to Belenix and
Nexenta via email.
[4]. Prevent the automatic dismissal of OpenSolaris as a Desktop OE
based on the sole reason that "it does not support or include KDE"
(it has happened, and it's been documented in some needlessly bad
press).
[5]. Provide those customers who have asked for KDE with what they
have asked for: KDE.
[6]. Provide an OpenSolaris equivalent to what is currently considered
the "best" (whatever that means) GNU/Linux Distro (SUSE Enterprise
Desktop 10, which provides both GNOME and KDE as a choice).
[7]. Provide a unified way of bringing into OpenSolaris those external
dependencies which are common between KDE and GNOME, and which should
be included in OpenSolaris. In spite of several discussions on this
subject, on this very email list, we are in the same spot where we
were two years ago.
[8]. Provide guidance to the developer community outside of
OpenSolaris, if not directly, then by example. If we are to build a
community of developers from outside Sun, and if we are to pique
their interest in OpenSolaris, then a lot of good will and interest
can be obtained just by showing "community spirit". Licenses aren't
that interesting. Source code is.
[9]. Alleviate the perception that "Sun dislikes C++". This language
is here to stay. KDE is one of the largest C++ software projects in
existence.
[10]. Provide a GUI building toolkit (QT) tested and patched for the
latest Studio compilers. AFAIK, Trolltech doesn't even test with
Studio 10 or 11 any longer, the latest documented compiler tested is
Studio 8. This is not a good sign. Maybe their interest in the latest
versions of the Sun compilers can be revived. Maybe this renewed
interest in these compilers will also renew Trolltech's customers'
interest in Solaris as a GUI-based development platform.
--Stefan
--
Stefan Teleman 'Nobody Expects the Spanish Inquisition'
KDE e.V. -Monty Python
[EMAIL PROTECTED]
_______________________________________________
opensolaris-discuss mailing list
[email protected]
