Gary A. Ross wrote:
I absolutely agree. Innovation at the right areas is critical.
Resources are tight everywhere, and trying to figure out where the
biggest "bang for the buck" is difficult at best. But, where does JDS
really differ from the other Graphic Environments out there?
JDS contributions aren't always obvious and much of it must be
contributed back to the community and therefore it is challenging to
maintain this differentiation.
- QA/Bug fixes
- i18n
- Documentation
- non-GNU/Linux platform, thin client and scalability bugs
- A11Y
- Enterprise supportability features such as APOC integration.
- Branding
The fact that JDS works within the constraints imposed by Sun's
enterprise class OpenSolaris distribution is easy to overlook. What
most differentiates JDS from other open desktops may be that it works on
Sun's Solaris without replacing the package manager, installer, and core
libraries.
I would like to see a much more concerted effort on things like basic
things like HAL, than a new feature in JDS.
There is work going on to integrate HAL, but that wouldn't really
differentiate JDS, it would just integrate HAL with Solaris rather than
with GNU/Linux.
That said, the current JDS doesn't live up to SuSE, or Ubuntu.
Opinions on specifics here would be useful. It's important for the
desktop community to be able to distinguish bugs and rfes we can work on
from those which need to be passed to other opensolaris communities
(e.g. packaging, kernel hardware support, installation, patch
management, compiler...) And some RFEs should be passed up to 3rd party
developers so that we might convince them that OpenSolaris is a viable
development and deployment platform.
Focus on what makes JDS a value to the user, and just (Yeah, I know
...) port the rest of the apps from the community. If there isn't much
value, why try and force functionality into JDS where it isn't needed?
Move those resources to where it would make the biggest impact.
Porting and helping prevent communities from going in platform specific
or unsupportable directions takes quite a bit of time. But assuming
that we had enough resources to make this part "easy", what I would like
to see is a desktop which exposes more of the features of OpenSolaris in
such a way that it hides the underlying complexity. For example:
Better integration of Sun Ray and Smart Card presence.
Shouldn't utremove tell desktop components that you're away so
that they can go into a mode where they aren't consuming resources to
update the (invisible!) screen. Thin clients and embedded devices would
be much happier if your desktop wasn't consuming so many resources when
you aren't present. Shouldn't the instant messenger mark you as away if
you remove your card?
Better integration of ZFS.
When I open a folder I see the current content, but I can
rotate to view layers of the same files from yesterday's snapshot, last
week's snapshot, last year's snapshot... Wouldn't this be a good use
of 3D?
Better integration of zones.
I'm not sure how this should be exposed from the desktop. If
I have an application which is hard coded to a particular kernel (e.g.
adobe acroread, skype...) should the mime handler look for a zone which
can run the application?
ACLs/Trusted Solaris integration (some internal work is already going on
here!)
I'm sure there are other good ideas for exposing Solaris features
through the desktop, but these are just a few off the top of my head.
Glynn Foster wrote:
Gary A. Ross wrote:
Thinking about it more, do you want the Solaris Desktop to be just
another "me too" desktop, just a GUI with all of the same modules
every one else has? I would think that you might want to
differentiate it more, integrating with something like the Looking
Glass project. I realize that I trivialize the porting process a
bit, so no flames please, and a lot of times it's not as easy as it
looks, but for the most part, the basic coding is already done.
People in this community live "on the edge", most of us being
developers. Would it be possible to use this as a platform to
experiment with alternatives, instead of the same thing?
Absolutely - we totally want to encourage innovation. However, we want
to encourage innovation in the right places. I don't think it's in
anyone's best interests to fork off development from the open source
projects we're currently involved in - we simply just can't sustain the
resources needed to do a good job. I'd much rather see that happen
upstream, though having a vibrant Solaris desktop community would be
awesome too.
Glynn
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org
