Linked by Thom Holwerda on Thu 9th Aug 2007 17:28 UTC, submitted by vondur
Linux "Don't expect to see key features of OpenSolaris showing up in the Linux kernel," said a top Linux maintainer. At his LinuxWorld opening keynote, Andrew Morton made it very clear that the appointment of former OSDL CTO and Debian co-founder Ian Murdock to Sun's OS platforms organization will not translate into a merging between the open source version of Solaris Unix with Linux. He didn't mince words. "It's a great shame that OpenSolaris still exists. They should have killed it," said Morton, addressing one attendee's question about the possibility of Solaris' most notable features being integrated into the kernel. "It's a disappointment and a mistake by Sun." Morton said none of those features - Zones, ZFS, DTrace - will end up in the Linux kernel because Sun refuses to adopt the GPL.
Thread beginning with comment 262238
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[7]: Possible interpretations
by kaiwai on Fri 10th Aug 2007 03:06 UTC in reply to "RE[6]: Possible interpretations"
kaiwai
Member since:
2005-07-06

I'd forgotten about the Sun Ray - that's a cute little piece of hardware. I didn't really realise they still sold them; the original revision was already quite cool so I assume the latest ones are pretty nifty!

I guess they kind of do need StarOffice in order to sell those (although I guess the "software to sell hardware" argument doesn't necessarily require them to have a separate proprietary version of OO.org - still there are decent arguments for doing that).


I'm running StarOffice 8 - the difference is this; support, you get updates long after the 'latest and greatest' has been released from OpenOffice.org, fonts, templates and clipart are also included in the bundle.

*nod* they do have some nice tech out there. I really like the direction they're taking with the T series SPARCs. There haven't been (to my knowledge) many or perhaps any business-oriented CPUs for a while i.e. ones that are optimised for transaction processing at the expense of single-threaded / scientific processing. Even if they can't match the fabrication technologies of the other big CPU vendors, I hope to see them outperform them by exploiting an underserved niche.


They've also had to adjust to much lower margins; I remember around 10 years ago their margins were like Apples are today, but eventually the crunch occurs.

Their new 'traditional' SPARC's are being jointly developed with Fujitsu, SPARC64 which as awesome performance. Sun should thank itself that Fujitsu focused mainly on Japan because Sun would have gone long ago had Fujitsu taken their products global and outside the niche of ultra high end servers.

Understood. OpenSolaris is a nice piece of technology. I've not used it myself, but I do very much envy some of the features already available (Zones and ZFS in particular).


For me I used Linux for a few months, them moved to FreeBSD, where I remained until 2002, then moved to Mac, then back Solaris. Its sad to say bit it is Linux vocal 'fanboys' which put me off, along with the arrogance of the developers who refuse to acknowledge when design mistakes are made, when development methods need to be taken up a level in terms of professionalism.

Sun have done an amazing job of taking what I understood to be a so-so UNIX and making it rather awesome. They've also taken what was becoming an "also ran" processor architecture and found a niche where it can really rock. If they can just bring all these individual awesome projects together in a strong combined strategy, they deserve to be very successful.


Solaris to me like a mamuska dolls they have in Russia, you look in and seen there are more and more; from the outside it looks like 'another UNIX' but it isn't until you really use it then you realise just how truly unique it actually is.

I mean, I hear uneducated twitts here compare zones to UML - ignoring the giant overhead that comes with UML, or KernelTap when compared to Dtrace - what it tells me, these fanboys are scared and worried. They're these new 'ubuntu boys' who grabbed onto Linux in the last 2 years thinking that because they can pipe a few commands and use vi, it makes them a UNIX expert. You can now see why I'm happy to have left the Linux community long ago before the riffraf arrived.

Reply Parent Bookmark Score: 2

Mark Williamson Member since:
2005-07-06

Mmmmm long term support is valuable, for an office suite - as for everything else.

I've been using Linux pretty much full time since 2001. In the past it was often pretty painful, but nowadays it does have a pleasing level of "just works" to it. Of course, ones mileage can still vary rather a lot on that score.

I too find that the kernel devs attitudes are infuriating sometimes, they can be arrogant, dogmatic and hypocritical. NIH also seems to occur more than you'd want. That said, I think there have been times when it *looked* like NIH but actually their decisions turned out to have some merit. It's hard to judge... in fairness, similar levels of arrogance are probably going on in other projects involving strong leaders and talented developers. But you don't always see it when those people can talk via private e-mail or face-to-face...

I do get the impression that the Linux guys are /much/ stricter about what they check in than they used to be. Stuff that's not "just so" often gets quite a rough ride until it's either ready or dropped. Sometimes I question the results of the process but it does seem to work "not too bad" overall (for now). I've generally considered Andrew Morton to be one of the most level-headed and generally sane of the kernel devs, which was why I immediately thought perhaps he'd been quoted out of context.

I mean, I hear uneducated twitts here compare zones to UML - ignoring the giant overhead that comes with UML, or KernelTap when compared to Dtrace - what it tells me, these fanboys are scared and worried. They're these new 'ubuntu boys' who grabbed onto Linux in the last 2 years thinking that because they can pipe a few commands and use vi, it makes them a UNIX expert. You can now see why I'm happy to have left the Linux community long ago before the riffraf arrived.


There are lots of fanboys around, it's true. I think most people are an irrational fanboy for something at some stage in their life - the trouble comes when people don't think hard enough. I don't have a problem with people who are just new, or enthusiastic, just with guys who are rude / unreasonable because they get too caught up in it.

UML has some advantages over Zones-type technology but also plenty of disadvantages. IMO one would ideally want a platform combining at least machine-level virtualization (i.e. a hypervisor or similar) and a zones-style OS virtualization. They're both useful things. I'm actually quite excited about the containers work going on in Linux now. It'll provide Zones-type functionality but the implementation for mainline seems to be proceeding in a nice, clean, generic way.

As for SystemTap, from what I've read it may well one day equal DTrace but it's not there yet. I think it'd be cool to have a dtrace-alike for Linux, but I'm not sure how much I'd use it personally.

Overall though, I sometimes lament the lack of success of systems like Plan 9. Oftentimes I find myself looking at the way Unixes do things and shouting "Why!?!". No matter how much I enjoy the broad functionality and compatibility of modern Unix, I still kinda wish for a completely clean system built from the ground-up on solid principles. Never mind, on Linux I get to watch DVDs instead of thinking about things like that! ;-)

Reply Parent Bookmark Score: 2