Sun Microsystems Inc has maintained that it will be developing its own extensions to the Linux operating system after confirming suggestions that its recently released Sun Linux 5.0 distribution is based almost entirely on Red Hat Inc’s Linux 7.2, writes Matthew Aslett. You can also read an interview with Sun Microsystems Chief Executive Scott McNealy.
Post Red Hat Knock-off, Sun Still Plans Own Linux
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
sure its based on a relatively mature product but starting out with version 5.0 is still retarded. Version racing seems to be the general trend among linux distros..
from the interview:
There’s a lot of work that should be done that isn’t being done. The government could be enforcing the antitrust laws. They could be giving out remedies, not rewards. They could be simplifying the tax code … (and) accounting standards. They could be reducing the size and scope of government.
There’s a lot of common sense coming from this guy (in this interview anyway).
Though, I still don’t get why they’re doing Solaris on Intel. If customers want Linux (or Linux-like functionality), why not simply give Solaris the Linux features that they’re asking for? If customers are saying, “…but we want it on *Intel* hardware”, Sun’s sales people should ask them “why?”.
>If customers are saying, “…but we want it on *Intel* >hardware”, Sun’s sales people should ask them “why?”.
Sun’s sales people don’t need to ask them why. They know the answer already — Intel hardware is cheap.
So,Sun, where are versions 1-4? Did I miss something?
A couple days back, I heard a quote from Bill Joy, saying that open source was bad. Of course, if you read closely, you realize that’s not what he meant at all. He meant that the GPL is bad – after all, he said that ideally, the best license would be “something like the MIT or BSD license”. But still, I thought he was out of his mind.
Now, I read this interview, where Scott McNealy pretty much says the same thing – that the GPL is bad because it’s anticommercial, but interfaces need to be public. And I realize that Bill Joy wasn’t so wrong after all.
The GPL never hurts. If x package is made GPL, that’s always a good thing. But a lot of corporations want to be able to add value to a product; to add something that their competitors don’t have. What’s important is only that the product is based on open specifications. If certain cars only ran on certain roads, everyone would immediately see that as absurd. But Win32 apps only run on Windows, and nobody cares.
What Linux allows us to do is to have a lowest-common-denominator. Everything can be based on one open platform. Everything a given person or group runs on top of it might be proprietary, but it’s still an open base. Someone else could come along and build tools that worked perfectly side-by-side with those proprietary tools. Esesentially, “proprietary” no longer exists; the tool that works only with others from the same vendor is no more. Proprietary only refers to the tool itself, not to its interfaces.
Sun, incidentally, is particularly well-suited to this, because they have Java. Windows and Linux have one very annoying “feature”, namely that applications compiled with one version of a library need that and exactly that version. Maybe between one version and the next, a bunch of new features were added, or some features were taken away that the program didn’t use anyway, but if it’s a different version it won’t work. Two common ways of getting around this on Linux are to keep tons of versions of libraries (awkward), or recompile everything (difficult and doesn’t work with closed-source).
The third possibility is to write a wrapper encapsulating absolutely everything; and then, programs only have access to a very limited number of functions, but that’s all they needed anyway. A web browser in Python needs only to call a few commands, and it fetches a website. At the lowest level it can make the connection itself; at the highest level, it calls a command, and the website’s there. Those commands, for the most part, are quite static, and while the way that the VM actually executes the commands might change, the programs don’t; so all that needs to be recompiled is the VM. Obviously, this is how Java works. It’s not perfect, and often quite slow, but it’s constantly improving.
Additionally, what Java allows you to do is develop a program, in Java, and then run that program on pretty much any operating system that’s ever sold a copy (sorry SkyOS). For servers, that’s irrelevant; you want real-time performance, which means VMs, for now, are out. But for desktops, embedded devices, etc., it’s great. Much better than anything MS has to offer.
sounds like a big cry baby. He should concentrate on making good products rather than whining. Sheesh, this guy sounds like Sun is on the verge of dying if Microsoft isn’t shut down. What a loser. IBM did a magnificent turn around after MS whipped its rear end on the desktop. Sun can turn things around too if their CEO decided to concentrate on building real solutions rather than wasting most of his time whining.
SUN Stock Price: 4.19 (ouch — SUN may end up like SGI)
So,Sun, where are versions 1-4? Did I miss something?
Not really. Sun likes to gratuitously rename and renumber things.
Sun’s SunOS operating system was BSD based. Sun collaborated with AT&T to produce SVR4, and decided to replace SunOS 4.x with the resulting operating system. They figured this was a large enough change to market the product under a different name. However, no one in the computer field wants to buy release one of anything. So they retroactively renamed SunOS 4 to Solaris 1, and thus could market their new operating system as “Solaris 2”.
Similarly, when they felt “version number pressure” from the upcoming release of WinNT 5, they renamed what was supposed to be “Solaris 2.7” to “Solaris 7”. This was such a sudden move that some of their docs weren’t updated in time for the launch. Microsoft’s response was to release Windows NT 5 as “Windows 2000”.
And if you ask the new Solaris 9 operating system what it is via uname -a, it will respond:
“SunOS 5.9”
Yours truly,
Jeffrey Boulier
PS Ok, to be incredibly precise, Sun officially says that “Solaris” means the kernal of the operating system (e.g. SunOS 5.9) plus the user interface, GUI, etc. Colloquially, SunOS means “the BSD Sun operating system (pre 5)”, and Solaris means “The SVR4 SunOS”.
While I can not agree more with Jeffery’s comment on sun’s strange numbering system (java 1.1, java 2…) early version of Sun Linux are Colbalt Linux (which are an old version of RedHat).
aleksandr wrote:
…Now, I read this interview, where Scott McNealy pretty much says the same thing – that the GPL is bad because it’s anticommercial,..
I don’t see anywhere where he says the GPL is bad.
From the interview.
I think it’s a shame the stock market got so overheated a few years ago because it set expectations at levels that are not achievable, not sustainable and not right. Yet employees can’t understand why the options we gave out back then aren’t very valuable. So that’s a problem.
This is crazy. My employer has provided me options 4 times. Once at 48$, once at 66$, once at 28$, and once at 20$. The current stock price is 15$.
I’m not even full vested with the 48$ options. However, I noticed that several senior employees have recently cashed in options at 2$. I believe this to be a completely arbitrary practice, and applaud Coke and other seeking option reform. Talk about Golden Handcuffs.
If certain cars only ran on certain roads, everyone would immediately see that as absurd. But Win32 apps only run on Windows, and nobody cares.
Cars are very much different from computers. What you are saying is that all OS and hardware becomes secondary and the merits and speed of the VM (Java) becomes the deciding point. Seems good…. NOT.
It wouldn’t work. For example, does car accessories for the Merc work in the Porsche? Most likely no. Sure, they use the same roads, but PCs and Macs use the same power lines. Plus in reality, this doesn’t work.
Developers would look for their target platform which their target audience uses mostly, and then they would optimize for that platform. Probably instead of compiling using bytecode, they compile it for the platform. Maybe they would use native APIs. This happens a lot with Java – no, not because of Microsoft cause this happens mostly in the J2EE space.
The GPL never hurts. If x package is made GPL, that’s always a good thing. But a lot of corporations want to be able to add value to a product; to add something that their competitors don’t have.
If Red Hat wrote support into the Linux kernel to have support for 120 processors – the next version of SuSE would have it. If SuSE wrote support into ext3 and the Linux kernel to have a database file system, in a few versions Mandrake would have it. The original goal of GNU/FSF was to get rid of companies whose profits depend on software – whether dirrectly or indirrectly.
If GPL doesn’t hurt, name me ONE company that is making a lot of money out of purely GPLed software? Don’t even dare to say Red Hat – whose profits is considered very low for the amount of effort they placed in. Windows racked in for Microsoft 40 billion bucks. Linux, 11 million bucks.
The third possibility is to write a wrapper encapsulating absolutely everything; and then, programs only have access to a very limited number of functions
Stupid idea. Let’s take Windows for example. IE uses a set of DLLs (which 9 states want to remove). Now, Explorer (the shell) uses many of these DLLs, and same with Windows Explorer with the desktop. Later applications came to use IE’s DLLs. Now, lets see what happens if all that never happened:
1) Windows would consumer much more memory than it is now. It would use much more CPU power too. ALl because of redundant stuff.
2) Developers would spend a lot more time reinventing the whell when they could have save time and cost using the free DLLs.
3) Netscape would be happy, and would bully the web market like what Microsoft is doing now.
But for desktops, embedded devices, etc., it’s great. Much better than anything MS has to offer.
For the desktops, people don’t want to use something slow. ALL the Java apps I have used is slower than it’s closest non-Java competitor. On embedded devices on the other hand, I’m not too sure, never used a Java-compatible embedded device.
I work for Sun in the Cobalt group, so I have some insight into the naming process. The first 4 versions of “Sun Linux” were the base operating systems (also based on RedHat) on the various Cobalt Networks appliances. Granted, it’s still a bit gratuitous and was entirely a marketing decision (I’m not sure anyone asked anyone in Engineering what their thoughts were)… I just know they didn’t want to release a “1.0” version as people instinctively shy away from those, whereas there _is_ quite a bit of background experience in releasing Linux “distributions” in the Cobalt group, so something over 1.0 _is_ justified IMO…