On Fri, 1 Jul 2005, John D Groenveld wrote:
> Unless you're confident that Anil Gadre and company's marketing wonks
> already understand your concerns, please don't stop here.
> You're on a roll.
For the sake of argument, JDS is not one size fits all. Its very nature
is to be minimalistic and consistent, primarily geared towards a
business environment (nothing wrong with that). However, by the time Sun
improves JDS to near CDE speed and consistency, it will be obsolete and
resemble something other than GNOME. JDS 3 is based on GNOME 2.6, and
GNOME 2.10 is already out. At this rate, the latest GNOME will be 4 or
more releases ahead by the time Sun QAs bits of 2.10. Maybe I'm wrong on
that, but that's my outsider perspective.
Sun says it can't afford to do Linux/JDS and that's plenty reasonable.
But it says it can do Solaris/JDS, which means it is still investing a lot
of money/time/effort maintaining a fork of the GNOME codebase. If Sun
truely understood the Open Source movement, it would make sure its
changes made it into the community source in a timely manner, and a
portion of that money/time/effort could go into paying a few people to
help the GNOME project. In other words, make the GNOME project better for
everyone, and put some of the maintenance burden on the community in
exchange. Make it so that what I download tomorrow as GNOME 2.11 would
be the same thing Sun packages as its defacto desktop (i.e integrate the
QA process into the project, allowing Sun to ship the latest version of
GNOME within update releases). If I have some Linux machines running in
my environment, the user experience can be truely identical, regardless of
whether they're accessing a Sun Ray on S10 Sparc/x86 or * Linux on a U20.
This was the idea behind JDS in the first place, but we'll not have that
going forward. In addition to a faster release cycle/more up-to-date
desktop, Sun gains even more community support for being interoperable.
Sounds kind of utopian, doesn't it? I know there are lots of problems
with this, but let me dream.
Now, I know that Sun is contributing heavily to the GNOME project, at
least in terms of usability. They are also paying a guy who is now the
maintainer of GDM, and is supposed to be getting the SRSS specific bits
into the community version. But case in point, it's been over a year
since the GDM thing, and I don't think those bits are available to me yet.
If I want to use GDM on my Sun Rays, I have to use a version that only
works with JDS 2/3 (GNOME 2.4/2.6). As it is, I use WDM which works
remarkably well, but isn't as pretty. Maybe Sun is doing a lot more in
this area that isn't directly visible, so I could be way off base with
What does this have to do with my concern of using Solaris as a desktop
for my Sun Rays? My users expect a nice desktop environment, and being
in a research institution, I have the flexibility to keep things
far more up-to-date than a typical Fortune N company. JDS3 is not bad,
but it's not nearly as up-to-date as the GNOME you'll find in any recent
Linux distro. In addition, there is a fair amount of related software
that isn't included in JDS, and an equally fair amount that compiles
cleanly under Linux but not under Solaris (neither of which are directly
Sun's fault, but I argue that it is a problem for them if they want
general community acceptance. Janus is a poor solution that doesn't
solve the underlying problem). I can download more recent packages of
GNOME and desktop extras from places like Blastwave, but such projects
have limited resources and are also not as up-to-date.
Solaris is a pain in terms of keeping non-system software up-to-date and
installing new versions of applications, when compared to something like
Debian Linux. If I want the latest version of Firefox that was released a
a few days ago, I can type one simple command to install it more or less
automatically. I can come close to this with Blastwave, assuming the
package maintainer for Firefox has had the opportunity to create a new
package. Therein lies the problem.
All of this boils down to community. For the more popular Linux
distributions (Fedora, Debian, Gentoo, etc.), they have well established,
enthusiastic communities with good infrastructure for maintaining third-party
software as value add to the base OS. The Red Hat/Fedora model is
definitely something Sun should pursue in terms of the desktop experience
for commercial Solaris, and the possibility exists to do this if/when the
OpenSolaris community builds. It would be nice to see a project like
Blastwave receive an infusion of Sun corporate backing and new volunteer
effort. From what I've seen Dennis Clarke talking about, he has a plan of
sorts for building the "good infrastructure" necessary to facilitate this.
Given the state of other similar efforts, it has the best chance of
succeeding in my opinion.
Whether or not Anil Gadre and company see this as a worthy concern
remains to be seen. The beauty of it is that with relatively little
investment on Sun's part, they can nurture the seed they've planted with
OpenSolaris and speed the growth of a real community and a real
distribution that rivals anything on the Linux side, with plenty of
kickback for commercial Solaris value add. It lets them keep their
developers/hobbiests happy and uses that base to fuel innovation, while
keeping the good stable bits for the corporate types.