Frustrated by Oracle’s delay in releasing the latest version of OpenSolaris, the OpenSolaris Governing Board is growing uneasy over Oracle’s lack of communication regarding the future of the Unix OS code. At least two members of the board have even said they would be open to forking the code base from the Oracle version.
The lack of communication isn’t just something Oracle are guilty of; it seems to be an epidemic in the business world having poor communication not only with partners but also customers – both potential and existing.
I understand that Oracle is going through the Sun business with a fine tooth comb; so why not communicate what is happening? just a broad over view and some sort of communication showing that something is being done and that the project isn’t be ignored by Oracle. Communication, no matter how small, is valuable because it shows the community that there is care being taken.
one Oracle has finished assessing how it is going to move forward with Suns products there will be a lot of news. Oracle doesn’t want to jump the gun and say something it might have to retract. They have a lot of work on their plate with this huge acquisition and they have to be very careful not to give false information of false hope.
Though near complete silence seems unnecessary, i understand why they haven’t said much as they don’t want anything to leak out. But still, it would be nice to hear something, even if its just “were working on it and will let you know as soon as we can”
I had a look on this list:
http://mail.opensolaris.org/pipermail/indiana-discuss/2010-April/
It appears that there are show stopper bugs – so why didn’t they just come out and state the obvious as a news piece on the front page of OpenSolaris.org? the mind boggles some times when there is a distinct lack of communication by people who are in their area of expertise, very smart and talented.
Regarding Oracle; there are rumours that Oracle will keep some features back from OpenSolaris sometime in the future or maybe more of a delayed merging between the work and the code in the tree. It would be disappointing if they took that route because it would go against what Larry Ellison said regarding pirates – that those who don’t pay for software aren’t their customers anyway (in that they weren’t going to be a potential customer because they didn’t want to pay for it in the first place – so nothing lost, nothing gained).
They are a publicly traded company, so they have to be very careful about what they say.
I assume OpenSolaris is not at the top of their priorities, due to the lack of revenue/profit proposition. It sucks, but that is how greedy and successful companies operate. Capitalism and technical excellence/creativity are not usually compatible.
In my (albeit) limited experience talking with people who have run open solaris, is that they were looking for something that was compatible with solaris. That may become more difficult if the forks diverge too much.
Those of us Linux folks that have tried it to kick the tires, haven’t found it to be better enough to convert. Sure it has dtrace and zfs goodness, but its not gnu userland, and the package manager put stuff in weird ( to us ) places.
It goes both ways their are die hard Solaris guys that wouldn’t touch Linux with a 20 foot pole… most of them will probably move to one of the BSDs if they have to.
Edited 2010-04-16 16:46 UTC
I bet most of those “diehards” are running Windows or Mac on desktop.
I’m sure that is true for some and not for others. I know people that use linux and mac also, and I myself sometimes use all of the above (except mac, unless I have no choice).
Could be but they aren’t running it on a server and that was my point… I personally use Linux 95% of the time I’m just saying the people that were *really* using solaris and can’t pay for it now are probably going to move to BSD … and not Linux
And I guess my point was that it’s not terribly interesting what you are running on server (since you actually “use” very little of the OS on server). Those “solaris die-hards” are still primarily windows users, and OpenSolaris probably made very little to change that.
Have you tried the 2009.06 release or a later build, or even looked over their web site? I think you’ll find there are quite a lot of things that have changed, for the desktop/mobile device.
A partial list:
Wireless support for numerous chipsets
Integrated desktop packages (many of these)
FUSE support
Wacom support (may be in the works still)
Virtual TTY
TimeSlider
Improved Audio (Boomer)
Continued Gnome integration and initial KDE support (Korona)
Also if you go on the forums or IRC, you’ll find quite a few people using it for desktop use. Not all of these people are oldschool “diehards”, but some of them are.
I haven’t tried it, nor seriously looked at what it has to offer. It’s firmly in the “I don’t care” bucket, along with *BSD these days.
I’m doing alright without zfs and dtrace. By the time I need them I imagine I can go for btrfs and SystemTap.
I suppose that is fair – you should go with what you need when you need it. Right now I admit I don’t use dtrace, though I want to – but I do use ZFS (and I mean I use the extra features, not just use it as a filesystem). I do wonder if SystemTap will ever get to the level of dtrace based on what some people have said, but I haven’t looked into it lately. Btrfs sounds like it will be comparable in many ways, with a few advantages and disadvantages over ZFS – but again I don’t remember the details of this preliminary comparison.
So then you’re really in no position to comment
I actually consider the relative lack of GNU userland a big plus, what’s keeping me off Opensolaris is the lack of “Quality of Life” stuff vs what I can get with say Fedora or Ubuntu and the relative sluggishness of the system.
Huh? I guess you have kicked those tires that much, if you missed that OpenSolaris has indeed GNU userland.
IT has the GNU tools available yes, but the Sun version is there as well and are preferable to many of the folks who come from a Solaris background
ok, that seems reasonable.
People coming from solaris want it the solaris way. People coming from linux see is as a welcome addition so that they finally have tools that are doing things they like.
Some time ago I talked to someone that was an HPUX guy. When I said.. ok 25+ years old, no human readable flags on ls. grep that’s not recursive…. he then responded but you can easily do it this waY (find, xargs, grep) and then I thought…. ok if someone wants to live in the history — fine.
I tend to think that the gnu tools addition (not a replacement) was a very very wise decision.
I actually come from linux (but have been using opensolaris for 3 years now), and have also been using find, xargs, and grep for my needs. I did not now GNU ls had a recursive option – man pages will surprise you! (or at least me)
You mean find . | grep whatever? Thats how I have always done it… And I use Linux.
and people cat /etc/passwd | grep ….. so?
The tools are generally more efficient. As stated for HPUX: if you have 10 TB storage mounted somewhere and you have to group every three digits to find out how much space you have… how wonderful would it have been if human readable flags were there?
Also find . […] | grep — ok if you do it that way. However there are easier ways. What’s wrong with that?
Serieusly it’s ok if someone doesn’t use the gnu toolkit but I assure you that it makes your life easier.
I’ll admit I’m not a Solaris expert (I’ve mostly used OpenSolaris and NexentaCP). But when I’ve used pure Solaris for Oracle systems I seem to recall having human readable format and search parameters in find:
$find . -name “foobar”
I think sometimes people just assume one toolkit isn’t as good as another because it behaves slightly differently – regardless of whether the toolkits actually have the same level of sophistication in the commands.
Unfortunately this fork project would have many problems since start . Major problem will be lack of sponsors , many companies have to choose . Get OpenSolaris fork without Support or get still upgrated oracle solaris with support . This situation will be painful for osol because corporations will pay oracle .
Only two options can transform this state of affairs :
PORT more tools from/to bsd mainly freebsd or change license to gpl and porting from/to linux .
http://www.theregister.co.uk/2010/04/16/oracle_solaris_nehalem_ex/
Or there is a third option; OpenSolaris becomes the basis of Solaris 11, they finally fund a project to finish off the i11n component of libc which cannot be open sourced (plus the other parts needed to fully emancipate its reliance on closed source components) – much discussion about in the past but nothing has been done in over a year. Turn it into a fully open source project akin to Fedora is to Red Hat and move it forward from there. Unfortunately I don’t see it happening because in the past we have seen some Sun employee’s sabotage open source efforts by erecting road blocks.
Edited 2010-04-17 01:58 UTC
Not quite Oracle is company which don’t like supporting open source their major products are still closed and in this moment their OS project are set for supporting their commercial solutions . Like brtfs .
Several former SUN employees have said in the mail lists, that there are showstopper bugs. Partuicularaly in ZFS and dedup.
I prefer that Oracle releases a bugg free version later, than a buggy version earlier.
Why the fuzz?
Why the fuzz? do you mean, why the lack of communication? you’d need to talk to Oracle management who seem to think that the world are full of mind readers.
You mean that when MS is late with a new version of Office, or late with a patch, then MS has to explain themselves officially why they are late?
Maybe.
I think that lack of communication is because Oracle has a lots of more important stuff right now. I hope that Oracle will be better at communication later when they have fully absorbed with Sun. Otherwise, I agree with the critics. I hope this is only passing.
I agree with the opinions expressed about possibly forking it also. Prospective clients may have to use OpenSolaris as somewhat of an evaluation seeings that Solaris is no longer free for people to try. Oracle is another company that bought out the competition, they will pick through the remaining carcass of Sun Microsystems and take what they want and leave the rest to die. It’s the name of the game.
Oracle is probably just busy taking stock of what they have now. When Red Hat bought Qumranet, they spent 6-9 months pouring over everything to make sure they knew what they had. Nothing worse that putting out source code that Sun licensed from someone else and that slipped through the cracks. I imagine Sun had a crap load of code that will have to be vetted before anything is done with it. At least the Java stack is done already, since they did it for OpenJDK. That and MySQL probably won’t require much work.