When it comes to open sourcing Solaris and Java, patents and politics are leading Sun toward a change of heart. The question is which open source licence should govern the building of projects out of the company’s technology crown jewels. The open source Solaris project began with a Community Development and Distribution License, and open source Java employs version 2 of the General Public License. Now, though, Sun likes the idea of governing both projects with the upcoming GPL version 3, chief executive Jonathan Schwartz said in a speech and an interview at the company’s analyst summit in San Francisco on Tuesday.
Sun Looks to GPL v3 for Java, Solaris
47 Comments
Intel 3945 A/B/G is unsupported on Solaris.
I never said it was. Just as my Broadcom Wireless chipset doesn’t currently work under Ubuntu Linux though they are working on that.
However, it is likely the 3945 A/B/G will work with the ndiswrapper that is available for OpenSolaris.
However, expect further announcements related to Intel hardware since SUN has just announced a new partnership with Intel.
These things take time, and it isn’t because of SUN alone. Hardware manufacturers are very picky about the details they give about hardware, especially wireless chipsets.
v3 of the GPL seems way to scary for my taste. I think I would be better off with closed source. To many restrictions that are supposed to “Restrict Evil” use of GPL Code like DRM restrictions and Patent restrictions. While I am not a big fan of DRM and Patents I don’t like adding tangental political motives to the license.
-
2007-02-10 12:35 am
Ringheims AutoThen be happy with Windows Vista. But dont get scared when you encounter any of that DRM stuff
-
2007-02-10 2:01 am
b3timmonsv3 of the GPL seems way to scary for my taste. I think I would be better off with closed source. To many restrictions that are supposed to “Restrict Evil” use of GPL Code like DRM restrictions and Patent restrictions.
Nonsense. No one has or will ever even have to agree to the GPL in order to use or modify the covered code. So right away you have more rights than closed source, for which you typically must accept a EULA before use. So how could closed source possibly be better? Only upon distribution of modifications involving the covered GPL code, must one accept the license.
While I am not a big fan of DRM and Patents I don’t like adding tangental political motives to the license.
DRM and software patents allow free software to be made effectively nonfree. Given that the Free Software Foundation is writing v3 to protect free software, how can those two threats be considered tangential?
Moreover, if political motivation itself is your dislike, then you cannot agree with any version of the GPL at all, since such motivation has always been essential to the GPL’s existence. Read the preamble to v2, for example.
What was your reasoning that led to your conclusion about tangential political motives?
I recommend learning something about the GPL, preferably from, you know, its authors: http://www.gnu.org
-
2007-02-10 6:55 am
BrandybuckBut GPL *v3* may go beyond mere distribution. There is a very high likelihood that it will include restrictions on remote execution. It’s meant to close the “webapp loophole”. If a user executes the code remotely, the GPL v3 may be triggered. If this covers the VM, then this can have a profound impact on Java developers.
-
2007-02-10 1:03 pm
Ford PrefectSounds FUD to me. I have never read something like this ni the GPL 3 drafts. Can you please point it out?
-
2007-02-10 2:35 pm
b3timmonsBut GPL *v3* may go beyond mere distribution. There is a very high likelihood that it will include restrictions on remote execution. It’s meant to close the “webapp loophole”. If a user executes the code remotely, the GPL v3 may be triggered. If this covers the VM, then this can have a profound impact on Java developers.
Please refrain from repeating obsolete and discredited speculation. How about just reading the current draft?
Anyone wanting to use the GPLv2 to address the “loophole” can already do so by using a modification of v2, the Affero GPL, as described in
http://www.gnu.org/philosophy/license-list.html
v3 is no different than v2 here in that it cannot be used verbatim to close the loophole. The matter is simply left for the individual developers to decide.
-
2007-02-10 9:02 am
rayinerThe DRM and patent clauses aren’t solely based on political motives. In the context of the license, they are direct attacks on various mechanisms that allow companies to use GPL’ed software without actually contributing changes back in a usable way. By leveraging DRM or patents, you can distribute the source at a technical level, but prevent anybody from actually using it in practice. The Tivo example is the classical one: Tivo distributes their modified GPL’ed code, but since the machine will only run signed binaries, you can’t modify that code, recompile it, and use it. It largely negates the whole point of open source.
This discussion would be incomplete without the statement issued by the Open Solaris Governing Board after the discussion of this amongst teh community.
Quoted from
http://www.opensolaris.org/jive/thread.jspa?threadID=23699&tstart=0
CAB/OGB Position Paper # 20070207 version 0.6
Topic: Should OpenSolaris be dual licensed via CDDL and GPLv3
Published by: OpenSolaris CAB/OGB current members:
Casper Dik, Al Hopper, Roy Fielding, Simon Phipps, Rich Teer
Background: Over the past week or so a heated and passionate debate
has taken place on the opensolaris-discuss mailing list [1] relating
to the possibility of dual-licensing OpenSolaris, and in particular
with GPLv3 as the 2nd license, in addition to CDDL, under which
OpenSolaris is currently licensed. The CAB/OGB (henceforth referred
to as simply the OGB) has observed this discussion carefully and
individual OGB members have been active participants.
Now that this discussion is winding down, the OGB, as the elected
representatives of the OpenSolaris project, will render an OGB
statement of position and provide guidance to the community to bring
closure to this discussion and to determine the communities near term
future licensing direction.
Discussion summaries (the pros and the cons):
The Pro dual licensing community members assert that dual licensing
will:
– increase developer mindshare
– attract active developers currently working on other FOSS projects
– promote code exchange across FOSS boundries
– end the constant anti-CDDL campaign waged by GPL* license
proponents
The Anti dual licensing community members assert that dual licensing
will:
– increase licensing complexity and futher complicate this already
legally complex licensing landscape
– lead to endless continued debates related to various “what if” code
inclusion/exclusion scenarios
– allow a one-way code fork by acquiring the OpenSolaris body of
code, manipulating, removing or modifying the (eventual, but
currently unknown) GPLv3 license terms in a way that prevents or
impedes the changes being propagated back to OpenSolaris.
– *not* entice or attract GPL* proponent FOSS developers, who want to
ensure that other Operating Systems (they actively work on) flourish,
to OpenSolaris.
—————
The OGB, having heard arguments from both sides, concludes:
o Discussing GPLv3 is pre-mature as the license does not exist at
this time.
o That there is little, if any, benefit to dual-licensing OpenSolaris
with CDDL and the yet to be approved/upcoming GPLv3 license – aside
from possible short term good press for the project.
o There are significant downsides to dual licensing, including, but
not limited to, license complexity, confusion and the possibility of
long term bad press from any exception language that such a license
would inevitably require.
o GPL* licensing OpenSolaris would be yielding to a small vocal
minority of FOSS developers who use the lack of GPL licensing, purely
as a means of fostering FUD towards OpenSolaris and who will, in all
likelyhood, find some other workable mechanism to continue to foster
FUD towards the project.
o There are higher priority action items to be completed in order to
build developer mindshare and that this opinion is held by a large
number of current contributors and acts as a barrier to other
potential contributors.
————-
The OGB having carefully weighed the available options concludes and
decrees that:
o any option related to GPLv3 dual licensing be re-assessed no sooner
than 6 months after the GPLv3 has been published and approved.
o Further discussions related to any form of dual licensing be
postponed until after the GPLv3 has been published and approved and
should take place on the OGB discussion forum only.
o Further discussion on GPL* is merely a diversion and distraction
that should be discouraged, so as to allow the community to
concentrate on the higher priority action items – especially those
that will improve developer mindshare.
——————-
[1] begin here – it will take approximately 5 hours to read all the
related threads:
http://mail.opensolaris.org/pipermail/opensolaris-discuss/2007-Janu…
MOTION To adopt the CAB/OGB Position Paper # 20070207 version 0.6 by
Al Hopper seconded by Rich Teer. Motion carried unanimously (In
favor: Rich Teer, Casper Dik, Row Fielding, Al Hopper. Absent Simon
Phipps).
-
2007-02-10 4:27 pm
b3timmonsThanks for that statement. What surprises me about it is the perceived degree of anti-OpenSolaris or anti-CDDL sentiment. Do that many people know or care enough about these things to be against them?
For example, many wrongly assume that the CDDL is not a free software license. OTOH, OpenSolaris under CDDL is rightly seen as just another vehicle for Trusted Computing — business as usual. This keeps it off the radar screen of anyone not wanting their work to be made nonfree. While the Linux kernel under GPLv2 is also such a vehicle, it just so happens to be in the center of the radar screen. Any assumption otherwise is just unsafe for Sun.
Delaying the decision until late this year also puts it very close to, among other diversions, the likely first release of the kernel component from the Coyotos project. That Coyotos will not be ready for production is beside the point — the point being mindshare. Projects such as Coyotos are attractive on a number of levels, especially with ever better VM technologies for experimentation. However innovative Sun’s Redshift may be, it will be hardly anyone’s dogfood, the value of which Sun underestimates at its peril.
OpenSolaris is very much indeed up against something, but FUD is certainly not it. FUD is not at all required to be in awe of the scope of experimentation that GNU/Linux and VM technologies encourage.
Edited 2007-02-10 16:42
-
2007-02-11 1:23 am
binarycrusaderOTOH, OpenSolaris under CDDL is rightly seen as just another vehicle for Trusted Computing — business as usual
Huh? How do you even begin to assert that? Just because the CDDL does not have a anti-DRM clause does not mean thatit is encourages or acts as a vehicle for Trusted Computing. By that logic, every single Open Source license ever created that doesn’t have such a clause is a vehicle for Trusted Computing.
Stop the FUD.
Delaying the decision until late this year also puts it very close to, among other diversions, the likely first release of the kernel component from the Coyotos project. That Coyotos will not be ready for production is beside the point — the point being mindshare. Projects such as Coyotos are attractive on a number of levels, especially with ever better VM technologies for experimentation. However innovative Sun’s Redshift may be, it will be hardly anyone’s dogfood, the value of which Sun underestimates at its peril.
Perhaps you failed to understand that SUN has nothing to do with the decision to delay the consideration of the GPLv3. The community does. *We* (the community) decided that it was inappropriate at this time, and the CAB as our representatives agreed and unanimously supported us.
The license will not be changed by SUN, it will be changed by a decision of the community. If SUN was just going to change it, they could have done so without *asking* us first. However, they asked us, and we responded.
So take your FUD elsewhere please.
-
2007-02-11 2:14 am
b3timmonsBy that logic, every single Open Source license ever created that doesn’t have such a clause is a vehicle for Trusted Computing.
Bingo, hence my description, which I also attributed to Linux you will note. Moreover, I know about the delegated decision making.
Thus, this is not FUD directed against Sun or its community at all; rather it simply expresses values held by any advocate of free software.
Indeed, I and many others are full of hope about Sun, hope that must, however, be tempered by that natural hope for alternative technologies. What some of you guys must get over is always seeing this kind of hope play out as FUD and believing that your technology by itself will carry the day. That word you use — “community” — is the big wildcard — make or break.
Again, none of this sentiment requires any negative feelings toward Sun, merely an acknowledgement of the myriad of opportunities that GNU/Linux has had without Sun for some time. This results from what Microsoft identified as its main target way back in the Halloween Documents — not companies but a process, a process that Sun must reconcile itself with somehow. It already has done it a bit with OpenOffice, so I know you guys can do it, too.
Show me the error of my ways, if you please.
Edited 2007-02-11 02:18
-
2007-02-11 7:20 am
elsewhereThe Pro dual licensing community members assert that dual licensing
will:
– increase developer mindshare
– attract active developers currently working on other FOSS projects
– promote code exchange across FOSS boundries
– end the constant anti-CDDL campaign waged by GPL* license
proponents
This is the part I don’t get. If developers stayed away from contributing because of CDDL, then why will they contribute because it’s CDDL + GPL? More so, I personally think that one of the biggest advantages of the linux kernel development model is that there is no requirement for copyright assignation, this is a much easier pill for commercial companies, as well as independent developers, to swallow.
I really don’t understand the argument that co-licensing under GPL v3 will somehow encourage further development contributions and co-operation.
The Anti dual licensing community members assert that dual licensing
will:
– increase licensing complexity and futher complicate this already
legally complex licensing landscape
– lead to endless continued debates related to various “what if” code
inclusion/exclusion scenarios
I don’t think this is a big an issue as it seems; there’s no license confusion unless people want to confuse the licensing. Other projects manage GPL + non-GPL type dual-licensing.
The issue will be convincing developers supporting license A and developers supporting license B to agree to contribute their work under license A + B. This isn’t a problem restricted to openSolaris.
– allow a one-way code fork by acquiring the OpenSolaris body of
code, manipulating, removing or modifying the (eventual, but
currently unknown) GPLv3 license terms in a way that prevents or
impedes the changes being propagated back to OpenSolaris.
This is almost guaranteed to happen. For all the support being offered by the FSF community, at the end of the day their agenda conflicts with both the CDDL and Sun’s reserved right to relicense openSolaris. It’s a solid mathematical probability that the day after openSolaris is GPL’d, it will be forked and renamed HURD 1.0 to become the basis of GNU, given the refusal of the linux devs to convert to v3.
– *not* entice or attract GPL* proponent FOSS developers, who want to
ensure that other Operating Systems (they actively work on) flourish,
to OpenSolaris.
I disagree with the part about the developers not wanting other Operating Systems to flourish. I don’t think the developers are as polarized as the community likes to imply. I think the inhibitor to attracting wider support will be the CDDL dual-licensing, and the fact that Sun requires copyright assignation. From the licensing aspect, developers prefer CDDL or GPL, or either. For pro-GPL developers, “either” is not an option, so really there is no gain; they’ve always had the option to dual-license their own code if they wanted.
And as for the copyright issue, linux didn’t achieve it’s current level of development solely by attracting grass roots developers; it attained it by attracting companies like IBM and HP to contribute code, something they would likely not have done if they had to assign copyright. Sun is in the same business these companies are so they should understand this.
At the end of the day, I can’t help thinking that the talk about v3 licensing is merely publicity generation to score points with the FLOSS community. Aside from the FSF gaining a new kernel to hijack, there’s no significant tangible gain to either Sun or the GPL community from this move.
And frankly, if the FSF community decides to hijack openSolaris with a fork to v3 for their own purposes, linux actually comes out ahead. It’s one less distraction for the developers to deal with…
Another important point is that GPL v3 will be compatible with Apache License v2.0 and the Eclipse public license (EPL).
This would allow several exiting technologies (e.g. OSGi, trapestry, SWT…) to be included in the JVM.
-
2007-02-10 1:16 pm
Simply put, can GPL 3 code be put in to a GPL 2 licensed source set? If they GPL 3 OpenSolaris, will that enable code sharing with the current Linux kernel sources, or is the OpenSolaris code excluded much like it is today because of CDDL?
-
2007-02-10 4:58 pm
b3timmonsSince v3 and v2 are incompatible, the Linux kernel will not benefit. OTOH, about 1/3 of Linux kernel code is under v2+ and thus is eligible for incorporating into OpenSolaris under v3. (OpenSolaris under CDDL seems more problematic here, however.)
-
2007-02-11 7:36 am
elsewhereOTOH, about 1/3 of Linux kernel code is under v2+ and thus is eligible for incorporating into OpenSolaris under v3.
It’s not that simple. First off, I don’t know where you got 1/3 from, but a quick grep of the linux source makes me suspect it’s significantly less, way significantly less. Besides that, don’t forget that the GPL extends to derivative works; therefore code in the kernel, even if it’s licensed with the “v2 or later” in the boilerplate, will likely be bound under the kernel’s v2-only license. It would really depend on the nature of the code. Regardless, Sun will never permit linux code into openSolaris for the sole reason of preventing the risk of license taint, and the fact that they require copyright assignment.
For all intents and purposes, despite openSolaris potentially licensing under v3, both linux and openSolaris will remain seperate, autonomous islands with no potential for cross-collaboration. Which is kind of sad, but then they likely wouldn’t be choosing v3 otherwise.
What really matters – what actually ends up happening. Sun has been pretty good about doing what they say, lately, but you never can tell.
I’m mixed on the GPLization of Solaris/Java. I’m not a big fan of the GPL for personal reasons (no need airing them out) – but trying to think in terms of what will be good for Sun in the long-run – GPL might be the best choice. Instant +1000 karma with the OSS (gnu/gpl) crowd. Also makes porting easier, from linux – which is kind of scary imo. I don’t want the hacked up/halfway working device drivers from hell ported over. :p
Since Sun will still maintain control over the direction of Solaris/Java (simply because they have more paid developers working on these two projects than anybody else) … so that should avoid the mass-importation of bad-code from the GPL folks.
All in all, I’m for it – but I’ll wait until it happens to celebrate.![;)](https://www.osnews.com/images/emo/wink.gif)
Of course, the FSF will drive v3 adoption. glibc, gcc, and the rest of the development stack, along with hundreds of FSF packages, will automatically move to v3. Moreover, a huge fraction of free software in general is licensed v2+ and so could move to v3.
Sun is more than just one of many firms engaged in the v3 drafting process; with the freeing of Java, Sun will be the biggest corporate contributor to FOSS — the corporate leader. Other firms will take inspiration from Sun’s example — Sun is not only an important company, but also one with a young CEO who has recently been associated with a turnaround and innovative technologies.
So the GPL will win big here, but how will Sun win? Let’s face it: Sun is competing for GNU/Linux users and developers. With enough of this mindshare, Sun will win. Clearly, Sun needs to differentiate OpenSolaris, and licensing under GPLv3 is an obvious help. Of course, this will be but a part of the cooperation with the community. There will be inevitable dilemmas to consider.
For example, while Sun cheers the GPLv3, Sun also emphasizes that it is a license addition to, not a replacement of, the CDDL. So if OpenSolaris supports “Trusted” Computing, how is contributing to it any different than the Tivoized Linux case? How would a free software developer be any more motivated here? (Note that this goes beyond the case of, say, the dual-licensed MySQL, since Treacherous Computing is relevant to just kernels here.)
In this context, I would not be surprised to see a TC-less fork of OpenSolaris under just the GPLv3. While this may not be ideal for Sun, it would undoubtedly be better than the current CDDL-only situation. Indeed, such a worst-case scenario does not at all have to be that bad for Sun if you think about it!
I am hopeful that Sun is serious about cooperation and, to reiterate, is emerging as the FOSS corporate leader. Not only have they chosen well with v3, but their continuing dialog with free software leaders is very much appreciated.
Edited 2007-02-10 03:46
As I’ve noted before, Sun only chose GPLv3 because it really doesn’t do anything in combination with CDDL. It’s purely about street cred and mindshare, as you suggest.
Why? Because the CDDL, believe it or not, is the more permissive of the two licenses, in every facet that I’ve managed to think about. There’s never a case (that I’m aware of) where the GPLv3 says yes and the CDDL says no. It’s always the other way around.
So for all you GPL zealots around here, consider OpenSolaris to be effectively licensed under the CDDL. This means that OpenSolaris is weak copyleft, allowing new source files to be licensed under non-CDDL/GPLv3 licenses. So license propagation is granular at the function level for non-OOP languages and at the class level for OOP languages.
As for TC, I imagine that this is a feature reserved for flagship Solaris. However, this has more to do with validation and marketing than with licensing. The CDDL would permit TC for OpenSolaris as long as the closed parts don’t touch CDDL functions or classes, and if they do, Sun owns the copyright, so they can just relicense those parts.
The licensing choices made by Sun, combined with those made by the Linux kernel community, create a copyright wall separating the two projects as far as the kernels are concerned. No code may be shared between the kernels. The two projects will compete for developers and ultimately marketshare on the basis of the technical merit of their kernels, as well as the developer-friendly CDDL vs. the user-friendly GPLv2.
May the best kernel with the best religious affiliation win!
I don’t want the hacked up/halfway working device drivers from hell ported over. :p
I assume that is a tongue and cheek remark because Solaris desperately needs alot of those Linux drivers, and most importantly, ALSA, to bring it kicking and screaming into 2007. Right now, the current status in regards to hardware and software support for Solaris right now sits at around 1996 compared to Linux and Windows.
on the server side, linux is far behind Solaris.
Depends on what you define as ‘far behind’ – as a server operating system, that is, sitting a small corner which no end user interacts with directly then sure, its great.
However, when it comes to a server being used as a desktop hosting device, as with the case of Sunray or as an operating system on a non-Sun workstation, it is absolutely terrible.
Opensource applications don’t compile, weird inconsistant locations for various opensource files, paths not properly setup because of a so-called fiction that ‘give flexibility’ when in reality it is more ‘we’re too lazy to set it up’ syndrome.
Sun know the problems with Solaris, thats not the problem, all operating systems have their problems, I’m running Windows Vista right now, haven’t seen any problems but thats not to say there won’t be any nasties pop up in the future.
The problem is that Sun is doing nothing to fix the problem, they’ve been fixated on cutting staff rather than working out *why* their profit hasn’t increase; instead of using the employees as whipping boys, how about addressing product deficiencies – fix the problems and the customers will come.
Its the old story, focus on the product and profits will follow naturally – too bad management seem to be focused more on cost cutting that product improving and inturn, profit generating.
Are you going to start this “things don’t compile” nonsense again? Which applications don’t compile for you? I look at the list of applications available at SunFreeware, BlastWave and Solaris4you and what I have compiled both at work and home and I keep thinking “where is the problem”?
The PATH issue is complete rubbish, if you want something to work modify the PATH based on the shell you are using. I have had to do this with Solaris, SCO Unix, AIX, HP-UX and RedHat Enterprise Linux. I have even had to tweak a Windows path from time to time, this is not an OS issue, this is something the user should be able to do.
And what is broken? How about some specifics, have you submitted any bugs to OpenSolaris.org? In the past I have had no problems with Sun fxing bugs, it might not come out in the next release, but they do get fixed.
Are you going to start this “things don’t compile” nonsense again? Which applications don’t compile for you? I look at the list of applications available at SunFreeware, BlastWave and Solaris4you and what I have compiled both at work and home and I keep thinking “where is the problem”?
Grab the source off KDE’s ftp, KDE 3.5.6 and try to compile it using an out of the box installation of Solaris 11/06 without any modifications to its default settings.
The day you can do that with a default installation of Solaris, then it has made progress.
The PATH issue is complete rubbish, if you want something to work modify the PATH based on the shell you are using. I have had to do this with Solaris, SCO Unix, AIX, HP-UX and RedHat Enterprise Linux. I have even had to tweak a Windows path from time to time, this is not an OS issue, this is something the user should be able to do.
Until you have conflicts between the GNU version of iconv and the default install with Solaris, or the fact that all the gnu utils have g in front of them, or the fact that there seems to be links left right and centre setup by Sun for libraries.
Its a friggin nightmare. All I should need to do is download and compile components as required; not try to work out why things keep breaking when compiling!
And what is broken? How about some specifics, have you submitted any bugs to OpenSolaris.org? In the past I have had no problems with Sun fxing bugs, it might not come out in the next release, but they do get fixed.
The day I can download KDE + Qt and everything compiles out of the box, then it has made progress, until then, Solaris for all intensive purposes is a waste of time; the Sun compiler is broken to the point it can’t compile Qt without the whole thing dying, the gcc included with Solaris is old, outdated, and like their own compiler, buggy.
Again, they *know* what needs to be fixed; they just can’t be bothered fixing them.
> Its a friggin nightmare. All I should need to do is
> download and compile components as required; not try to
> work out why things keep breaking when compiling!
One of the goals currently under way is to change Solaris so such applications can be compiled out of the box. Simply, ./configure and way you go. No funny settings, no setting environment variables, no hacking on configure.in files.
Is there an ETA on when this is going to addressed? if there was anything that was holding OpenSolaris back from widespread mainstream desktop adoption, its the weirdness of the system and inability to compile something without major TLC.
It’s starting right now with Solaris Express. New open source things are being integrated into /usr/bin or their expected location from the start and the existing pieces in funny places such as /usr/sfw are being moved too. Here’s a recent code review request to get bison and GNU’s m4 moved from /usr/sfw/bin to /usr/bin
http://www.opensolaris.org/jive/thread.jspa?threadID=23797&tstart=0
Another recent one is the integration of GNU coreutils into OpenSolaris in /usr/bin
http://www.opensolaris.org/jive/thread.jspa?threadID=20199&tstart=0
In the coming weeks, there should be a bunch of other such cases including uprevs of many of these components.
So I assume that if I download a source file off the net, and it fails to compile with OpenSolaris, it is an issue with OpenSolaris that needs to be addressed rather than the software itself – if so, it would be great as it would allow one to run the latest version of KDE without needing to jump through 1/2 dozen hoops.
Its a friggin nightmare. All I should need to do is download and compile components as required; not try to work out why things keep breaking when compiling!
The only nightmare is people’s non-portable software that makes non-standard assumptions.
The day I can download KDE + Qt and everything compiles out of the box, then it has made progress, until then, Solaris for all intensive purposes is a waste of time; the Sun compiler is broken to the point it can’t compile Qt without the whole thing dying, the gcc included with Solaris is old, outdated, and like their own compiler, buggy.
The sun compiler is NOT broken. Why is it that people make the incorrect *assumption* that if X item doesn’t compile on Y platform, that the problem is with Y platform and not X item?
Do you realise how much code out there is NON portable and is written with just Linux in mind?
Do you realise how many developers don’t actually follow POSIX standards?
If you did, you wouldn’t be so quick to blame SUN for something that isn’t their problem.
For the record, KDE and Qt work and run just fine on Solaris.
See this site:
http://solaris.kde.org/
Now go spread your fud elsewhere thanks
Oh, not to sound cynical, but how delicious an out of date version of KDE for download – I might as well be running an old out of date version of Linux or Windows 9x.
Oh, not to sound cynical, but how delicious an out of date version of KDE for download – I might as well be running an old out of date version of Linux or Windows 9x.
Oh heaven forbid, it’s a few point releases behind for the official KDE.org release! It’s a disaster!
This is because the package maintainers for KDE on Solaris have opted to focus on KDE 4.0.
If you want KDE then why don’t you download it from Blastwave? If you want to build it yourself fine, after reading the instructions I don’t see any real show stoppers other than making sure your library paths are updated (man crle).
And what exactly do you mean “with default settings”? If you mean that you do not set up a user environment (which includes a PATH statement) then I would not be surprised if the compile fails! It is not Sun’s fault that it doesn’t work, so stop blaming them!
I think the problem is you do not know how Solaris works, and instead of learning how to use it, you bitch and moan about how messed up Solaris is because something doesn’t “work out of the box” like it does in Linux. This is not the first time I have read something like this from a Linux user who is not familiar with Solaris and did not take the time to learn. Solaris is not Linux, and I would not expect it to work like Linux, just as I would not expect certain things I do in Solaris to work on Linux. If you are not willing to learn how to use Solaris then maybe you should go back to using Linux.
If you want KDE then why don’t you download it from Blastwave? If you want to build it yourself fine, after reading the instructions I don’t see any real show stoppers other than making sure your library paths are updated (man crle).
Out of date and broken KDE on Blastwave.
As for the path, it should be setup by default; too bad duplicate libraries are detected and all hell breaks loose during the compiling – specifying which source to prefer? good luck trying to find documentation that actually addresses the issue.
And what exactly do you mean “with default settings”? If you mean that you do not set up a user environment (which includes a PATH statement) then I would not be surprised if the compile fails! It is not Sun’s fault that it doesn’t work, so stop blaming them!
Yes it is; is it too hard for their to setup their library/header paths so that one doesn’t have to go through a hit-and-miss experimentation?
I think the problem is you do not know how Solaris works, and instead of learning how to use it, you bitch and moan about how messed up Solaris is because something doesn’t “work out of the box” like it does in Linux. This is not the first time I have read something like this from a Linux user who is not familiar with Solaris and did not take the time to learn. Solaris is not Linux, and I would not expect it to work like Linux, just as I would not expect certain things I do in Solaris to work on Linux. If you are not willing to learn how to use Solaris then maybe you should go back to using Linux.
My problem is that there has been no effort to improve the usability of Solaris so that people who might know what to do but are not inclined to spend time tweaking, can focus on the more important things rather than setting up mundane crap that should have already been setup by default.
> Opensource applications don’t compile, weird
> inconsistant locations for various opensource files,
> paths not properly setup because of a so-called fiction
> that ‘give flexibility’ when in reality it is more
> ‘we’re too lazy to set it up’ syndrome.
Actually, the reason these open source files were put in a funny location was concern about them changing in an incompatible fashion and the unsuspecting user ending up being burned if/when the such a change took place. In short, we thought an opt-in policy by changing one’s path was the right approach.
Times have changed and we learned the error of our ways. New open source commands are going into /usr/bin in general and the ones in /usr/sfw (that funny location) are being moved into /usr/bin as well.
It’s taken time, yes, but these changes have been comind and continue to come. If you look at Solaris Express, you can see a lot of improvement over the current Solaris 10 and in six months, the improvement will be even more apparent.
And that is probably why you don’t see those “hacked up/halfway working device drivers from hell” in Solaris/OpenSolaris, because the people resonsible for Solaris/OpenSolaris want a quality product that is up to the same standards as the Solaris/OpenSolaris code base.
Solaris is not trying to be Windows or Linux, and a number of people (myself included) have Solaris and OpenSolaris working on their PC’s without all of the Linux “features”. It might not be the latest and greatest, but what works works well, which to me is alot more important than supporting every device on the planet half-assed.
Solaris is not trying to be Windows or Linux, and a number of people (myself included) have Solaris and OpenSolaris working on their PC’s without all of the Linux “features”. It might not be the latest and greatest, but what works works well, which to me is alot more important than supporting every device on the planet half-assed.
If this keeps being the general tune around Solaris, Linux has nothing to fear. People like to consolidate, what runs on the server runs on the desktop, runs at home too. Look at the success Microsoft has had with that model.
What I use for a server is far different than what I use for a desktop. Since I cannot play Unreal Tournament on Solaris (and will probably never be able to) I have a Windows machine for that purpose.
Geeks will use Linux, Solaris, AIX, HP-UX, BSD or whatever for both their servers and desktops. The average Joe will use Windows because (1) that is what they are familiar with and (2) what they use at work.
I really don’t think Solaris has much to worry about from Linux or Windows, since Sun’s target market isn’t the home desktop machine. I just don’t know where you guys get that Sun is even trying to capture that market.
I agree with you a certain extent. I develop AIX, which clearly has a particular market in mind and does a better job than any other OS for this market. However, as developers we run into situations where we want to have some nice development features like modern desktop environments or code editors. AIX doesn’t offer this, and therefore nobody would develop on it if they had a choice.
Solaris is much (much) better in these respects, but it still doesn’t have all of the features that make Linux the ultimate development workstation. If people have a choice, they’ll develop on whatever system is most convenient, and right now that means Linux. Solaris might have some great features for certain sectors of the server market, but if it wants to gain developer mindshare from Linux, it needs features like better multimedia support. Developers, developers, developers.
And that is probably why you don’t see those “hacked up/halfway working device drivers from hell” in Solaris/OpenSolaris, because the people resonsible for Solaris/OpenSolaris want a quality product that is up to the same standards as the Solaris/OpenSolaris code base.
Solaris is not trying to be Windows or Linux, and a number of people (myself included) have Solaris and OpenSolaris working on their PC’s without all of the Linux “features”. It might not be the latest and greatest, but what works works well, which to me is alot more important than supporting every device on the planet half-assed.
Oh yes, I really hate it when my Realtek sound card is supported out of the box; I want to be forced to put up with no sound.
I really hate it that my wireless card is supported – OpenSolaris is so cool in that Sun programmers promise a driver for it for the last year, claiming it to be in development, but it never appears with in that year.
Yeah, I have to admit, I must be a closet Solaris fan because gosh darn it, I just hate it when all my hardware is supported, because its just a nusence being able to use my computer for more than just a glorified type writer.
Have you tried getting drivers from these sources:
http://solaris-x86.org/software/drivers/
http://www.bolthole.com/solaris/x86-laptops.html
Intel 3945 A/B/G is unsupported by OpenSolaris/Solaris – Sun has said they’re working on an NDIS but a year later and still no result; when questioned on porting wpi from OpenBSD to Solaris, they gave the community the run around.
All I see is a desperate company pissed that they can’t get a whole heap of free labour; you’d think that since OpenSolaris has been made available for over a year, the issues relating to the crap hardware support would have been addressed, but no – Sun is too busy firing programmers left, right and centre whilst forgetting its the programmers who manufacture the operating system, not magical fairies at the bottom of the garden.
> Intel 3945 A/B/G is unsupported by OpenSolaris/Solaris
> – Sun has said they’re working on an NDIS but a year
> later and still no result; when questioned on porting
> wpi from OpenBSD to Solaris, they gave the community
> the run around.
Work has progressed on the NDIS wrapper. You can download these from
http://opensolaris.org/os/community/laptop/wireless/
for now – over time, they’ll be part of Solaris proper. In addition, I would expect the new embrace of OpenSolaris but Intel will likely lead to improvments in the area of wireless and graphics drivers.
Intel 3945 A/B/G is unsupported by OpenSolaris/Solaris – Sun has said they’re working on an NDIS but a year later and still no result; when questioned on porting wpi from OpenBSD to Solaris, they gave the community the run around.
Actually there is a result. Go look at the community process and discussions. There is no “run around.” There is this little thing called “due legal research” and making sure that they have the legal rights to distribute, etc. That process takes far longer than anyone would like, but SUN isn’t in control of that.
All I see is a desperate company pissed that they can’t get a whole heap of free labour; you’d think that since OpenSolaris has been made available for over a year, the issues relating to the crap hardware support would have been addressed, but no – Sun is too busy firing programmers left, right and centre whilst forgetting its the programmers who manufacture the operating system, not magical fairies at the bottom of the garden.
Did someone dump some poison in your cereal or something? Seriously. Your accussations are nothing but pure, flat-out wild lies and speculation.
Your claims of “crap hardware support” are vague and lack any real meaning. If you don’t like the hardware support, try listing specifically what doesn’t work. I would never expect Solaris to support all the hardware that Linux does, and I hope it doesn’t. We don’t need the baggage. However, just about any reasonable modern x86 system should work just fine. I have run Solaris on multiple x86 systems of my own, from laptops to deskstops without any serious issues.
As far as firing programmers? You are aware that SUN has more engineers working on the Solaris operating system than any other *company* has working on theirs right? SUN has about 1,000 engineers working on Solaris. So much for your firing left and right theory…
I’m not even going to comment on your delusions regarding free labour.
Oh yes, I really hate it when my Realtek sound card is supported out of the box; I want to be forced to put up with no sound.
1) Out of the box sound support has been added for *many* chipsets recently. Check recent builds.
2) Windows doesn’t have “out of the box” sound support for a lot cards, not even Vista! So don’t play that game. Instead, you can get free, professional quality audio drivers (for personal use) from http://www.opensound.com/
I really hate it that my wireless card is supported – OpenSolaris is so cool in that Sun programmers promise a driver for it for the last year, claiming it to be in development, but it never appears with in that year.
Sorry, apparently you don’t understand how our community works. Please see the wireless project where wireless drivers for some devices are currently available and more are under development:
http://www.opensolaris.org/os/community/laptop/wireless/
Yeah, I have to admit, I must be a closet Solaris fan because gosh darn it, I just hate it when all my hardware is supported, because its just a nusence being able to use my computer for more than just a glorified type writer.
Really, spare us the FUD. On both my laptop and my desktop, under Solaris, I have the following working fully:
1) Networking
2) Sound
3) 3D Graphics
4) Removable USB External Drives (just plug them in and they appear mounted on my desktop!)
5) Keyboard, Mouse, SATA and ATA drives
Really, your information is way out of date.
Edited 2007-02-11 01:27
1) Out of the box sound support has been added for *many* chipsets recently. Check recent builds.
2) Windows doesn’t have “out of the box” sound support for a lot cards, not even Vista! So don’t play that game. Instead, you can get free, professional quality audio drivers (for personal use) from http://www.opensound.com/
Yes, I understand that, but past experiences I’ve had nothing but problems; if it does work now, I retract me statement.
Sorry, apparently you don’t understand how our community works. Please see the wireless project where wireless drivers for some devices are currently available and more are under development: