Linux and BSD have thriving communities. OpenSolaris is starting to get interesting.
And now QNX my favorite UNIX like OS has finally brought a community edition back to life.
I’ve been playing a lot with OpenSolaris the last little while. I wonder if there are any lessons to be learned?
As an alternate OS hobbiest, four things make Solaris more interesting than Linux for me these days: ZFS, DTRACE, iSCSI and gunsolaris.org, a port of the Ubuntu packages to Solaris.
Any thoughts on a gnuqnx.org? What effort would it take? There is an awesome packaging system and lots and lots of interesting software (OpenOffice et al).
Also FUSE for QNX 6 seems like low hanging fruit (this way you get ZFS and NTFS-3g for almost free), anyone looked into porting this yet?
Does the QNX license kill any of the above activities? Not really open source, but rather “source available” and all that.
I’m sure there are many lessons we can learn from the OpenSolaris community - they have been at this for a lot longer than we have!
FUSE has been prototyped - I think that finishing off a port of FUSE would be an excellent opensource project for the community. The developer did get as far as getting NTFS-3g mostly working, before he finished his coop term and headed back to school. Hopefully he is interested in continuing this nifty project…
Glad to hear a FUSE port was attempted? Any chance of getting the work that has been done on Foundry27?
Sorry I didn’t mean a new portal or new forums. I meant a repository as easy to use as Ubuntu. gnusolaris.org aka nexenta.org/os is an example of the Debian/Ubuntu packages ported to OpenSolaris.
Without an easy way to get and build the thousands UNIX applications out there the QNX community will sit where it has for the last 7 years (effectively in stasis). (Nexenta has > 12k packages, built by a very small team that leverage off the Ubuntu work).
Alternatively, something like MkLinux or CoLinux, where Linux runs on top of NTO would be a great way to simply “just use” the wealth of Linux software out there and still have the “sendreceivereply” micro-kernel architecture many of us have admired for 20 years.
When QNX was bought by Harmon, I thought that was it, Becker will use it to try and compete with Johnson Controls (and Windows CE) and QNX as a general purpose OS was dead.
This news of the renewed community effort is a dream come true for an old timer like myself who thought that the chance to make a living using QNX (as I once did for more than 10 years) was pretty much over.
Frankly, I don’t see why it should be restricted to GNU only. About two or three years ago I wanted to port some of the BSD userland to QNX. I never got enough time to do that but I still think it’s a nice idea. I guess the only restriction would be the applications’ level of POSIX conformity.
You need a QNX specific bootstrap (and a few other things) to start. Some of QNX employees are working very hard right now, try to get this out of the door, just give it some time.
This project looks very interesting. A way to quickly get > 2000 packages on to NTO.
It is great QNX is having co-ops and full timers work on this stuff, but their resources are extremely limited, and would be better spent working on the core OS, not porting community applications. If QNX is to have a real community, then they are going to need help from us the community.
