The Economist on The Rise of Linux

nickco3 writes "The Economist is telling the business world that Linux is a worthy adversay to Windows and Unix. It is free, runs on almost any hardware, and generally more secure than Windows As result it is dividing the industry into winners that offer Linux (e.g. IBM and HP), and losers that don't, (e.g. Microsoft). Sun is probably doomed."

Comments

This made me confused.

[termos]

Sun is probably doomed
According to this article SUN is the one who's not doomed.

Re:This made me confused.

[jc42]

... SUN is the one who's not doomed.

Right. Note first that Solaris is highly POSIX-compliant, as is linux. This means that most software ports from one to the other with few if any problems, as long as you haven't used the private extensions of either. This isn't a conjecture; I and many others have tested this with our software. Portability between linux and Solaris is easy, almost as easy as between linux and *BSD.

And note also that Sun is actively supporting several linux distros. There was some confused news recently about Sun supposedly dropping their linux. What they actually did was drop the attempt to "rebrand" RedHat linux as Sun linux. This was mostly because customers got confused. And some of them wanted RedHat explicitly. But Sun seems to be going strongly into the linux support business, letting someone else supply the POSIX-compliant platform that runs on their hardware and has all of their software goodies available as options.

There is a strong contrast with Microsoft here: Microsoft has been moving strongly to a "total experience" platform which doesn't allow any software that isn't on their approved list. So if you're a software developer, you are facing a market in which you can only sell to Microsoft, on their terms. If you try selling retail, you'll find that your software constantly breaks, until you sign the rights over to Microsoft.

Sun, on the other hand, has a strong history of supporting independent software developers by sharing information about the innards of their systems while not requiring onerous licensing of any sort. As either a software developer or an IT manager, it's obvious which would be the wiser purchase. Why would anyone with half a brain go with a secretive, monopolistic computer system when there's another available that is open and cooperative?

And for a final note, we might observe that Sun has in the past objected to being called "SUN", since that refers to the Stanford University Network that they grew out of. They are officially "Sun", which isn't an acronym for anything. In todays environment of rabid copyright and trademark enforcement, it's important to get such things right. ;-)

Re:This made me confused.

[snero3]

I have to agree with you here. We use Solaris as a our primary OS and linux as as our secondary OS (It runs things like email gateways while solaris runs DB's). What we have found is that most software we run on one platform will run just as happily on the other with a little recompile.

I don't think sun is as doom as everyone thinks but I do think they are in trouble and need a little creative thinking to see them through

Re:This made me confused.

[cpeterso]

Solaris is highly POSIX-compliant, as is linux. This means that most software ports from one to the other with few if any problems, as long as you haven't used the private extensions of either. This isn't a conjecture; I and many others have tested this with our software. Portability between linux and Solaris is easy, almost as easy as between linux and *BSD.


I think this is particularly interesting as Sun says that future versions of Solaris will be LSB compliant. I think this is an interesting spin opportunity. Sun is vulnerable to Linux on the low-end, but if they "embrace and extend" Linux on Solaris, they can support Linux but spin it as "training wheels" for Solaris.

Re:This made me confused.

[Moses Lawn]

Microsoft has been moving strongly to a "total experience" platform which doesn't allow any software that isn't on their approved list. So if you're a software developer, you are facing a market in which you can only sell to Microsoft, on their terms. If you try selling retail, you'll find that your software constantly breaks, until you sign the rights over to Microsoft.


I'm sorry, but what the *hell* are you talking about? The only thing I can imagine you're talking about is the "approved for Windows NN" logo campaign, which has been around since Windows 95 came out. The logo is completely optional (unless you want to sell your product in Office Depot or wherever it is) and does nothing but show that you comply with and support a bunch of Microsoft's latest pet technology. It does not force you to "sign over the rights" to anything, not even your soul.

Are you really trying to say that Microsoft is trying to control what software you can sell for Windows, and that you need to sell it through them? If you are, you'd better come up with a few shreds of proof.

If your software constantly breaks, then your code is broken and you'd better go fix it. That's about the size of it. Microsoft is a nasty corporation, but they're not stupid. More people developing for Windows means more copies of Windows sold. They like that.

Sheesh. The things that get "+5 Insightful" these days...

Re:This made me confused.

[Penguin Follower]

I'm sorry, but what the *hell* are you talking about? The only thing I can imagine you're talking about is the "approved for Windows NN" logo campaign, which has been around since Windows 95 came out. The logo is completely optional (unless you want to sell your product in Office Depot or wherever it is) and does nothing but show that you comply with and support a bunch of Microsoft's latest pet technology. It does not force you to "sign over the rights" to anything, not even your soul.

Are you really trying to say that Microsoft is trying to control what software you can sell for Windows, and that you need to sell it through them? If you are, you'd better come up with a few shreds of proof.

I think he is refering to something I read once (which may have just been rumor). I remember something about all software that runs on Windows in the future needed to be approved by and digitally signed by MS. I don't remember where I saw that, so I don't have a link. Peronally, I agree with you that MS isn't so stupid as to kill their market like that. After all, there are only 2 things that keep MS where it's at: 1) Applications, 2) "User-Friendliness".

Re:This made me confused.

[Newcastle22]

I think the paranoia about MS "forcing" their software onto users comes from Palladium.

d

Re:This made me confused.

[jc42]

I'm sorry, but what the *hell* are you talking about? ... Are you really trying to say that Microsoft is trying to control what software you can sell for Windows, and that you need to sell it through them?

Yes, that's exactly what I'm saying. And it's well known in some developer circles.

For example, I have a couple of friends who have been working on high-quality "component-style" music software. They are getting rather depressed these days.

The reason is simple: Their software more and more can not run on Windows unless they license it through Microsoft. The reason is that, if any component of MS Media Player wakes up and runs, their unapproved music programs stop working. Often, they need to be re-installed.

This behavior was documented in some of the earliest reviews of MP several years ago. It has gotten worse, as MP gets better at recognizing unapproved music software. The way to get off the hit list is to license your music software through Microsoft.

(Actually, there probably isn't an explicit "hit list". Rather, there appears to be an "approved list". If a program isn't on the list, it may have a lot of problems getting at some things that are locked by some component of MP. Some details of the MP implementation are hidden inside the binary, of course.)

It's well understood among music-software developers that Microsoft fully intends to control this market, similarly to how the recording industry has historically controlled the distribution chain. But Microsoft has the advantage that it controls the player on the major "desktop". The result will be similar to what has happened with commercial music: Aside from a handful of big-name acts, nobody but the controlling corporation(s) will be able to make money in the business.

In the US, at least, it's now obvious that this is perfectly legal. The outcome of the recent Microsoft case means that it applies to software, too. The future for Windows developers is that you will be able to sell your softare only under license to Microsoft, under terms of their license. "If you don't like the license, you don't have to sign it." And your software will die mysteriously at random times.

This is much of the explanation for the recent stories of venture capitalists dropping support for Windows development. They've figured out that they probably can't make a profit in this market any more. And it is also behind the recent stories of independent Windows software developers one by one switching over to linux or OSX. If you have a dream of getting rich from sales of your great new app, it's best if you can sell to the customers. If you have to go through a single licenser who can dictate terms, you'll never get rich no matter how big your sales might be.

I've never done any significant Windows development. But I have quite a lot of friends in this line of work. They are all showing signs of depression, and are all seriously looking into other computer platforms where their software is supported by the OS, and not controlled by its owner.

Re:This made me confused.

[Moses Lawn]

The reason is simple: Their software more and more can not run on Windows unless they license it through Microsoft. The reason is that, if any component of MS Media Player wakes up and runs, their unapproved music programs stop working. Often, they need to be re-installed.

This behavior was documented in some of the earliest reviews of MP several years ago. It has gotten worse, as MP gets better at recognizing unapproved music software. The way to get off the hit list is to license your music software through Microsoft.

(Actually, there probably isn't an explicit "hit list". Rather, there appears to be an "approved list". If a program isn't on the list, it may have a lot of problems getting at some things that are locked by some component of MP. Some details of the MP implementation are hidden inside the binary, of course.)


All right, now that makes your point more clear. You'll understand why I thought you were a little paranoid, reading lines like "your code won't work unless you sign the rights over to Microsoft". Yeah, this is shitty behavior, much like the business with digital cameras and Microsoft's photo manager software pushing third-party stuff out of the way. Of course, Microsoft makes no bones about the fact that it wants to control the digital-media-path-blah for the world, so I'm not at all surprised something like this would happen.

I wonder, though, how much of this is evil intent and how much is just the arrogant "of course our code is going to run things" attitude. It sounds a lot like their usual business of not telling anyone how to write their code so as to avoid this mess unless they either dig really really deep, or "partner" up with Microsoft (typically a dom/dub relationship, and you *know* who's the submissive). My ~10 years of Windows programming experience has convinced me that once Microsoft had decided to have some new bit of technology, if you want to keep your hair and sanity, you had better just accept the fact that you're going to do it their way or else.

I sympathise with your friends' frustration. I haven't done consumer software for a few years, since the Win98 days, so I've been able to ignore a lot of this. I'd hate to have to develop for XP. However, I'm not sure what other platform is left. Apple is just as well known for subsuming applications in with the OS, and you can't make money writing for Linux, so what can you do?

Re:This made me confused.

[jc42]

I'm not sure what other platform is left. Apple is just as well known for subsuming applications in with the OS, and you can't make money writing for Linux, so what can you do?

Well, in the tax forms I just filled in, I reported about $120,000 of income last year. All of it was for software development on linux.

It's just one data point, of course. But I've been working with a team that's helping a Big Corporation convert their old IMB-mainframe system over to an online flock of linux servers. Rumors that this is happening (and Larry Ellison's recent prediction) are not just rumors.

And, contrary to the usual image, we're Americans working in the US for a foreign corporation.

So much for the media's image of what's going on.

A growing "talking point" in much of the world now is "Do you want your computer systems and company data controlled by a big American corporation?" This isn't just a rhetorical question. A lot of people in the world are getting very worried about American corporate power. The fun thing is that there are a lot of us Americans who are quite happy to help them out of their mess.

There's no irony here; we don't like those big corporations any better than they do. ;-)

Re:This made me confused.

[Moses Lawn]

I said:

I'm not sure what other platform is left. Apple is just as well known for subsuming applications in with the OS, and you can't make money writing for Linux, so what can you do?


You said:

Well, in the tax forms I just filled in, I reported about $120,000 of income last year. All of it was for software development on linux.


I was speaking from the point of view of the desktop software developer writing music/multimedia apps for the home user, which I think you were referring to in your original post (the one I jumped on). There, you can't sell Linux software. It's *very* good to know that one can make good money doing OSS development outside of driver and kernel work. There's hope for me after all...

I'm with you completely re the fears of Big American Corporations. It kind of goes along with the rest of the world's fears of Big American Government. The fears are mostly right - BAC and BAG are about the same thing these days. I personally think things would work much better with a lot more diversity - many small organisms are healthier for an environment than a few very large ones, which tend to choke off everything else (kind of like algae) [Insert other biological and Darwinian metaphors here as necessary].

I'm afraid Microsoft and such are just making the situation worse, not that they'd ever believe (or admit) it. I'm not sure if this will eventually be a good thing or not.

Re:This made me confused.

[princeofweasels]

No wonder your confused your speaking from the point of the desktop software developer writing applications for home users who get those applications for free from M$. Well, percieved to be free, oddly enough office seems to have jumped several fold in price. Do everyone a favor including yourself and help out with Open office, xmmsjb or any other open source desktop app. gl and have fun. pow OSS will get you through times of no money better then money will get you through times of no OSS.

Re:This made me confused.

[Moses Lawn]

I'm not confused, it was the person who started this thread...

When I was doing active Windows development - the early/mid 1990's - this stuff didn't come free. It wasn't until Windows95 that Microsoft decided to stuff a bunch of formerly commercial and shareware utilities in with the OS. Even then, if you wanted more than bare bones, you had to buy something.

Even today, you don't get all that much, aside from the basic media stuff. Office covers basic productivity needs, but people still need educational software, interesting and useful utilities, screensavers, and I notice Eudora still sells pretty well, in spite of Outlook Express.

One of these days I'm going to start banging on some OSS code. A side effect of having done commercial consumer software is a heightened awareness of "fit and finish", if you want to call it that. Thanks for the encouragement.

as much as i like the

[minus_273]

economist.. i dont think MS is really a loser. Nor do i believe that anything else will be the dominant desktop OS. As much as we like to bash MS, they do do somethings right. Enough so that they will keep their customers. Heck sometimes they even make technological improvemsnts in their OS that others are just discovering. Look at all the research that went into Async IO in the open source world only to realise MS had done something better for nearly 10 years.

Re:as much as i like the

[AmberWaves]

Could you amplify just what MS did for the last ten years in the Async IO area ?? Just wondering.

Re:as much as i like the

[pVoid]

Ahhh... Amen to you brother.

We've all heard of the arguments about how some of MS' tech teams (for the kernel for example) are actually quite decent... so there's no need to rehash that.

I want to add a point though to your post, and that is that, in this day and age, anyone who utters "winner vs loser", black and white should be tarred, feathered and sent walking through the city center with a sign that says buffoon on it.

The fact that such a bold (and stupid) statement comes from the economist is actually surprising to me... but I guess such are the times.

PS. Btw, Async IO in NT *does* rule... all of the threading mechanisms in NT rule, IMHO.

Re:as much as i like the

[Anonymous Coward]

Actually, the article, if you read it, did not say that Microsoft was losing. Rather, it said that M$ would come out less of a winner.

The poster of the root article needs to read the article. Doh.

Re:as much as i like the

[pVoid]

Windows NT's entire IO system has always been asynchronous if you want it to. And it's architecturally very sound and useful.

You aren't going to get an answer here on /. that will be sufficient for you (and if you expect one, you are just a troller/flamer)... the best way to know is to use it a bit.

A good example of why I love it so much is that *every* object in the Kernel is synchronizable. You can wait on a file just like you can wait on a thread, process, or mutex. Once you get used to the architecture, it's very sweet.

Re:as much as i like the

[NetMasta10bt]

MS is really doing less and less right. They are increasingly getting more greedy with their licenses. There security is to be laughed at. And every three years they release a rehash of the same old product.

Eventually, (and soon), Linux will have the usability that 51% of the population will find acceptable, and it will landslide from there. There will be no reason to pay for Windows XY or Office YX. Software companies will start producing the little annoying things that exWindows users like for Linux, and everyone will be happy.

Just wait until MS starts writing software for Linux. That will be a funny day.

Re:as much as i like the

[tzanger]

PS. Btw, Async IO in NT *does* rule... all of the threading mechanisms in NT rule, IMHO.

I thought NT's threading mechanism was inferior to Linux's "processes (pretty much) == threads" methodology. That and the "every thread has its own private message queue" on NT drives people batty. To each their own, I suppose.

Re:as much as i like the

[pVoid]

"every thread has its own private message queue"

Threads do not have message queues foo. Windows have message queues... And the queue is associated with the thread that created it, but that's about it.

PS. A window is neither a process, nor a thread. It's a window.

PPS. I wouldn't trust your declaration of inferiority based on the fact that you didn't know the above to be true...

Re:as much as i like the

[geekoid]

remember its econimics. so if you make 1 billion one year, and set a goal to make 1.1billion the next year, yet only make 1 billion, you're a loser. the fact that you make a billion doesn't seem to matter.

Re:as much as i like the

[hughk]

As you know, the NT kernel is in a way a descendent of Digital's VMS. VMS in turned was based on a 16-bit operating system, RSX-11M. The interesting thing is that in the old days, propietary operating systems weren't particularly closed. RSX-11M was built from source code, so you always got the exec in source. Until VMS 4.5, the source listings were delivered on microfiche with the system. There were some very good books available on the internals of VMS too.

Personally, I think that the NT kernel seems quite ok. The trouble is that it is closed (so I can't easily understand what is going on) and it is surrounded by layers of crap. I agree that the Linux people could have learned a lot from the synchronisation objects in NT, but they could also have leaned a lot from older systems like VMS, which were extremely well documented. However, better asynch support is coming soon.

However, the next big area for Linux is resource locking (needed for clusters). NT does it, but not very well. VMS did it better (and still does, from last time I used it).

Re:as much as i like the

[RoLi]

Actually, Microsoft is a prime example of what they are now critizising: They never invented something, always reimplemented or copied. (The GUI was on Apple, Amiga and Unix years before, everybody except Apple had preemptive multitasking, etc. etc.)

Microsoft currently has a stronghold on the desktop (> 90%), but is declining everywhere else (on servers and embedded systems. But also the desktop stronghold gets cracks, many entities have already proven that a MS-free office is possible.

To put it in other words: Microsoft has not much to win, but a lot to lose.

Re:as much as i like the

[dusanv]

Two huge things still need to happen before that. X needs to get fixed and I think Keith Packard's move is going in that direction (I want my the copy and paste and my bloody five mouse button to just work). And the more important and difficult one is Wine. They are also progressing at snail's pace (it is a difficult task I give you that). People just won't move unless most of their software works on the new platform.

Re:as much as i like the

[rutledjw]

You aren't going to get an answer here on /. that will be sufficient for you (and if you expect one, you are just a troller/flamer)

And why not? There are oftentimes intelligent reponses. Sure, there's plenty of garbage, but there's also intelligent info posted with links to back it up. You've provided neither and missed an opporuntity to educate someone who asked a legitimate question.

the best way to know is to use it a bit.

Granted, I think that's true in general. HOWEVER, there are people who don't have the time to go play with such things. Example: I'm involved in server-side programming and would NEVER get into that level of code (kernel). Providing a link to someone like me would be quite useful and informative.

Re:as much as i like the

[dusanv]

You can't do async I/O on pipes which is *really* annoying (as in Win32 'CreatePipe'). No async I/O for Win9x (except for parallel and serial ports) family so you have to stay away from it unless you want to write separate code for Win9x & NT. Also the whole OVERLAPPED crap is just a lot clunkier than UNIX async I/O. And don't get me started with the crazy 'WaitForMultipleObjects'. Returns the index of the first signaled object and you have to test the rest of them. 'select' is a lot smoother than that. Don't get me wrong, making everything sychronizable is a right way to go except Win32 sometimes feels so half-assed (I can go on and on especially about the sockets).

Re:as much as i like the

[oconnorcjo]

And every three years they release a rehash of the same old product.

Actually the fact that they FINALLY stopped the win 3.1 line that was dos+win3.1/95/98/ME and gone completely over to a NT core is a HUGE IMPROVMENT! I still don't like Microsoft but they are providing real OS as standard.

Re:as much as i like the

[bhtooefr]

What was that about MS writing software for Linux? Try this. Anyway, I thought it was Windows F(uck the) U(ser). And their anti-piracy measures don't work. I'm downloading FP and Publisher XP now, and I've cracked Office XP Trial!

Re:as much as i like the

[leonbev]

Only a Slashdot economist could consider a company that owns 90% of the OS market and has over $40 billion in the bank a "loser".

Even the article itself doesn't think that Microsoft is in any serious trouble. And I quote:

"The most likely outcome is that customers will face a choice between Linux, which is cheap and cheerful, and Windows, which offers more bells and whistles, is tightly integrated with other Microsoft products and is easier for unskilled staff to use, but costs more. In short, Microsoft will be not so much a loser from Linux as less of a winner. In the server market at least, Linux is providing Microsoft with some much-needed competition."

Re:as much as i like the

[spitzak]

NT does have a process == threads idea as well.

I do agree about the stupid design of the message queues. All the messages should go to the first thread by default, no matter who creates the windows. Instead I have to be careful to create all windows in the same thread. And they MUST get rid of the assinine direct calls to WinProc, this has been a huge source of headaches and bugs (because if you destroy or change some object you must make sure any data that WinProc might look at is updated and consistent before calling ANY windows procedure, including things like WaitForObject! GAH!) This could easily be fixed by having anything that it wants to send to WinProc be sent through the GetMessage/DispatchMessage loop. Here X actually got things right and Windows loses.

Re:as much as i like the

[rmohr02]

And their anti-piracy measures don't work. I'm downloading FP and Publisher XP now, and I've cracked Office XP Trial!

Are you proud of yourself for pirating software? Personally, I don't pirate software--there's generally a better (or at least acceptable) open source alternative. That said, I hope Microsoft keeps with the "product activation" and makes it harder and harder to crack their software, so that people like you will be forced to actually pay for MS Office, or go with a free alternative.

Re:as much as i like the

[NetMasta10bt]

Hey I think he's actually on to something. Most people in the Windows world download pirated copies of software. There is some amount of illicit temptation about it. Like it makes you cool to have Office XP trial.

What we need are 'Leet' versions of GNOME and KDE, where you have to enter in a license key or get a crack for it, that way they can tell their friends what an uberhacker they are.

Re:as much as i like the

[VZ]

> Threads do not have message queues

Have you ever programmed under Win32? This is simply false. Read the MSDN docs for PostThreadMessage() function:

The PostThreadMessage function posts a message to the message queue of the specified thread

...

Messages sent by PostThreadMessage are not associated with a window.

Re:as much as i like the

[Ataru]

CreatePipe creates an anonymous pipe, and you can't ask for it to be overlapped. But it just does a CreateNamedPipe with a unique name (and you can do overlapped on named pipes). You could do that yourself, although I admit it's a bit strange the way the API is laid out.
I quote from the documentation for WaitForMultipleObjects:
"The function modifies the state of some types of synchronization objects. Modification occurs only for the object or objects whose signaled state caused the function to return. For example, the count of a semaphore object is decreased by one. When bWaitAll is FALSE, and multiple objects are in the signaled state, the function chooses one of the objects to satisfy the wait; the states of the objects not selected are unaffected." [emphasis mine].
You're not supposed to test objects individually, you just peel one off at a time and go back for another WFMO.
I'm off to read about the UNIX stuff.

Re:as much as i like the

[hackstraw]

MS is really doing less and less right.

However, they have done right things, and they certainly can rest on their laurels for a while.
This just popped into my head. Linux has not had big upgrades. With M$, there was DOS, WIN 3.1, nt 5.3, win 95, win 2000, win xp. And since I've been using linux in 94, the system is pretty much the same. I can't think of a feature from the releases of RH4->9 that I could explain to a non geek. Maybe, kde/gnome (basically a win95 ripoff), maybe usb support, but there has not been big bold releases that were landmarks to the end user.

Re:as much as i like the

[Captain Pedantic]

It would be more like a slashdot economist to consider such a company a "looser"

Re:as much as i like the

[fitten]

Yeah, in the past, I've basically put a while loop around the WFMO and just loop. Inside the loop is a multiplexor based on the event signaled if you need it. The code is very simple to do this.

I've use the Asynchronous I/O with windows sockets and got very nice results. It was very hard to get nearly the same results on Linux back then (1998-ish) but things may have changed since then on the Linux side. I haven't had to try to get the same levels of performance out of subsystems since those projects because the things I've worked on since then haven't required it.

In my opinion, the POSIX thread/synchronization stuff is bad. It think it's ugly and clunky. I had a friend who was on the POSIX committee for it and I told him the same.

Re:as much as i like the

[Narcissus]

Funny you should mention that: I've been noticing quite a few Linux/Mandrake/Red Hat serial number/crack files on Overnet recently. I downloaded one out of curiousity and found that it was just an executable that installed your typical spyware etc.

Re:as much as i like the

[Jason Earl]

The real difference is that "Linux" isn't a publicly traded company like Microsoft. Even if Linux does no more than halt Microsoft's earnings growth then MSFT is in for some serious problems. Microsoft currently has a very high Price/Earnings ratio. If Wall Street were to believe that Microsoft was no longer going to grow they would lose a great deal of their current market value.

Microsoft the company would still be packing billions away in cash, but Microsoft investors would be losing their shirts.

Re:as much as i like the

[dusanv]

I am afraid you are mistaken about 'WaitForMultipleObjects'. This is from the manual page (discusses the return value):

If bWaitAll is FALSE, the return value minus WAIT_OBJECT_0 indicates the lpHandles array index of the object that satisfied the wait. If more than one object became signalled during the call, this is the array index of the signalled object with the smallest index value of all the signalled objects

It returns the smallest index and you have to test the objects with higher indexes. As for 'CreatePipe' there is no reason it can't be overlapped except that the same call is supported on Win9x which doesn't do async I/O (the 'CreatePipe' is implemented using 'CreateNamedPipe' on NT which as you noted can be overlapped).

Cheers

Re:as much as i like the

[Ataru]

There's no need to be afraid! I am not mistaken.
The whole point of WFMO is to block your thread until something happens. You call it, then it returns when one (or all, iff bWaitAll) of the objects is signalled. Then you do something with that object. Then you call WFMO again. If more than one object was signalled, then WFMO will return immediately with the next index. You don't have to test the other objects, because WFMO does it for you. You are not using it correctly. What you are doing is called 'polling'. Polling is bad, mmkay, that is why WF[S|M]O exists.
Yes there is a reason why a pipe created with CreatePipe can't be overlapped, because it just doesn't take the required parameters. Win9x sucks, you will get no argument from me there.

Re:as much as i like the

[HuguesT]

You are quite right of course,

The flip side of this comment is that Linux was just as *good* in 1994 as it is now, except for a few minor enhancements such as support for SMP, non-x86 CPUs (including mainframes), 3D hardware, multimedia, clusters (Beowulf) and more. On the appearance thing we went from twm + fixed font to KDE/Gnome which really does not make much of a difference.

Of cource in 1994 we had a striving free software movement which has changed all that much, still producing a usable C/C++ compiler and what basically is an editor, and now we only have a few extra tens of thousands of open-source projects which don't have much of an impact, really. Except maybe for that Office rip-off and a rehash of that old browser, remember Netscape?

Meanwhile on the windows front, we have had 2 *complete* redesign of the GUI. Woah (3.1 -> 95 -> XP). And now we have a real kernel which does not crash all that much. An now Windows works on SMP systems, if you pay more, but only works on Intel CPUs, sorry (used to be that it worked on Alphas and PowerPCs, but no longer).

Technically at the kernel level Windows has been playing catch up, not the other way around. DOS, win3.x and 9x were all pieces of c*@p which only survived because of the installed based and Microsoft monopoly practices. Now win2k and XP are technically OK, but please don't rewrite history.

What you are complaining about is that Linux never had a DOS moment. And this is bad why?

Re:as much as i like the

[Just Another Perl Ha]

...and under Microsoft's definition of "win", that's tantimount to losing... :-)

Re:as much as i like the

[fitten]

Yes... It does say that. But it also says further down in the Remarks section:

The function modifies the state of some types of synchronization objects. Modification occurs only for the object or objects whose signaled state caused the function to return. For example, the count of a semaphore object is decreased by one. When bWaitAll is FALSE, and multiple objects are in the signaled state, the function chooses one of the objects to satisfy the wait; the states of the objects not selected are unaffected.

So, you can do it either way - use the WFMO to wait for one, then poll the rest every time you get control back, or you can just put it in a loop. You will pay basically the same penalty either way as you'll have to call WFSO inside a loop inside the WFMO to get each event. IMO, it's best to just use the simple WFMO in a loop alone.

Polling typically sacrifices cycles for latency. Async I/O sacrifices latency for cycles (for doing other things). Each has its advantages/disadvantages and problems that work well for each method. For example, if you have a tightly coupled problem that is data dependent, you have nothing else to do but burn cycles waiting on data. However, if you have a much more coarse grained app (like some client/server or master/slave) then the async can work very well.

Re:as much as i like the

[gmby]

Just wait until MS starts writing software for Linux. That will be a funny day.

Your a little late.
That day has come-and-gone already.

http://slashdot.org/article.pl?sid=03/04/11/1452 24 3&mode=thread&tid=109&tid=188&tid= 106

Re:as much as i like the

[blrptrpl]

A thread doesn't have a message queue until a window is created on that thread. True, all windows created on a particular thread share a single message queue, but, strictly speaking, the message queue is not an intrinsic feature of a Win32 thread.

Re:as much as i like the

[pVoid]

I will remain confident that you don't know what you're talking about. Threads can have message queues associated with them, but they are not part of a thread.

Here's an excerpt from the MSDN definition of PostThreadMessage... foo ...

Remarks

The thread to which the message is posted must have created a message queue, or else the call to PostThreadMessage fails

Fuck I don't like flaming people about stupid stuff like this, but you are flamebating me...

It's ignorant asses like you who think the multitasking capabilities of Windows are limited to message queues. You're not competent... don't go spreading FUD while you're at it.

Re:as much as i like the

[BalkanBoy]

Just wait until MS starts writing software for Linux. That will be a funny day

April 1st was 12 days ago... Please, don't throw around these convenient innuendos :).

Re:as much as i like the

[pVoid]

Last you programmed in win32, you put all of the rest of us good win32 programmers to shame then.

Read my answer to all of you ignorant flamebaiters, and then shut up.

Re:as much as i like the

[pVoid]

Read away...

I will admit that my anti-anti-microsoft rash got irritated a bit, and I might have been a bit flamy...

Re:as much as i like the

[pVoid]

Here's an even more explicit description of what we talk about here...

Notice how the name of the section goes:

MSDN Home > MSDN Library > User Interface Design and Development > Windows User Interface > Windowing > Messages and Message Queues >

Do you see "Thread" anywhere in there?

Re:as much as i like the

[NamShubCMX]

And they would probably add the evil bit to Linux too...

What, this has been played out already? :P

Re:as much as i like the

[dtfinch]

It's pretty sad the lengths that people go to protect software from piracy. I've know of a company that spent $7000 on a piece of software, then downloaded a pirated copy of that exact software that they've already purchased, same version and everything, because the anti-piracy measures were too much of a pain to deal with. They weren't trying to do anything that would violate their license, there were just a lot of annoying circumstances where it mistakenly thought that they were, and dozens of hours of lost work as a result.

I myself have been annoyed a lot with anti-piracy measures in products I've legitimately purchased for use on my own computer, such as Office XP.

As for pirating software, I see it as disrespectful, selfish, and dishonest. If it's not worth the price, you don't need the software. If you need the software, maybe it's worth the price. If you pirate, it's unfair to the people who pay, and you're benefitting from someone else's work without giving them the compensation they ask in return. Sure, their work is done only once and then costs nothing to copy, but you're still making it unfair for others and getting something for free without completing your part of the bargain. Piracy is to blame for all the stupid, annoying, and/or scary anti-piracy laws and measures we're seeing today. If you don't want to pay someone's prices, you can decide to just not deal with them.

It's annoying to see all these kiddies who say "f**k micro$oft", but would rather use cracked microsoft products than one of the cheap or free alternatives, admitting that they think microsoft products are better, but they would rather insult them than compensate them for the benefit they derive from the software which they seem to secretly love so much.

Re:as much as i like the

[tuffy]

If you're going to use Microsoft software, you should do so without paying. Less money for Microsoft to spend doing harm. That's not to say illegally copying Windows or Office is right, but it is less wrong than paying for it.

Pirating Microsoft software is much like stealing from your crack dealer. It might make you feel better for a little while, but you're still giving them market share, you're still using their tools to get your tasks done, you're still likely saving your data in their formats and you're not going to put them out of business, ever.

If you don't want to buy an OS, get an OS you don't have to buy. I think that's the best way to go.

Re:as much as i like the

[4of12]

Pirating Microsoft software is much like stealing from your crack dealer.

How right you are.

They've got goons from the BSA to protect their interests, too!

Reminds me of the recent ad campaign, too. I can see the revised version....(voice of 6th grader)

"Mom, Dad, someone's gonna offer me Office XP!

Don't like it?

Too bad!

You're gonna have to ask me where, what, when, who, and where's the license key?

Re:as much as i like the

[ratboy666]

Here's a troll...

Why do I have separate close calls for sockets under Windows? And what about select() -- can't seem to pass in NON-sockets. I haven't tried passing in a socket to (say) WaitSingleObject(), but the types are wrong, so it pro'lly won't even compile right. And this is CONSISTENT? Foo on you! Sockets don't even work as an IPC mechanism, which is a real pain... and there isn't even a socketpair() function.

Now, Linux (default) doesn't do message queues, and the pthread mutex were process specific, and didn't work 'twixt process (I should be able to mmap() in memory, and use the same mutex in two different processes... with the addition of a parameter!) Natch' Solaris supports ALL of that.

So, my OS ranking:

1 - Solaris
2 - Linux
3 - Windows

Interesting, because my code supports stuff in the EXACT opposite order. And, I have portability kludges all over the place. Piss on this!

Take Windows and shove it. Then, merge Linux and Solaris. Yeah... that's the plan.

Ratboy666

Re:as much as i like the

[bhtooefr]

I cannot afford M$ Office, but have to use it (my parents have taken a 100% anti-open source stance) on my desktop. They're not paying several hundred dollars for it - and they need it too.

Re:as much as i like the

[bhtooefr]

Sometimes, you HAVE TO get Office XP, and CAN'T afford it. My parents have decided that all open source software is evil, and won't touch it with a ten foot pole (I've got a little hidden on the hard drive, though).

Sun is NOT probably doomed

[Ubergrendle]

Sun is a hardware company; as much as they like to trumpet their 'value add' components and services, their bread winning business is SPARC. Java, a great language, generates little revenue and is a tool that drives need to purchase their hardware. Remember iPlanet? Approaching 0 on the latest web surveys as Apache dominates that space.

IBM and HP are making smart moves adopting Linux business models. As Linux matures and benefits from a gazillion different implementations, AIX and HPUX will begin to look less and less desirable.

Getting back to Sun, Solaris is not a revenue piece for them either. There was alot of complaining in the Slashdot crowd and Sun's commitment to Solaris on Intel has waned, but really, would you like to be running Solaris instead of Linux or Debian? Thought not...

Re:Sun is NOT probably doomed

[lurgyman]

Yes, but for Sun, the software was as much a way of selling their hardware as anything else.

Re:Sun is NOT probably doomed

[Jah-Wren Ryel]

Exactly. The only thing that differentiates Sun hardware is that Solaris runs on it, and thus the multitude of Solaris hosted applications runs on it. Take away Solaris and Sun doesn't lead in anything. Sparc performance has always been and still is a laggard compared to the rest of the industry. Pricing at the low-end is wiped by PCs, and features at the high-end, where the profit margins are still fat, are wiped by IBM and HP in the commercial sector and SGI in the technical markets.

Unless Sun figures out a way to turbo-charge the Sparc architecture, they won't have any competitive advantage in the near future.

Re:Sun is NOT probably doomed

[NetMasta10bt]

You keep telling yourself that. Sun has already said that it will look at AMD's Opteron 64bit processor, and for good reason. It's a smart move if they plan on being competitive pricewise. But this effectively puts them into the same market with dell, ibm, hp, etc, to which they can all undercut eachother all day long.

Sun is not doomed, but it will be a shadow of the company it used to be.

Re:Sun is NOT probably doomed

[christophersaul]

Solaris is indeed an important differentiator, but the hardware's also extremely important. How many Intel systems scale as well? None. SPARC's never been the leader in cpu benchmarks, that's true, but Sun have consistently produced balanced systems that perform well when running applications rather than cpu benchmarks. If just fast cpus were important, we'd all be using Alphas.

Sun's market share in Unix systems grew last year, where HP and IBM sank. HP and IBM certainly do not 'wipe' Sun at this level, either technically or financially. The low end Sun systems are competitively priced and the storage is excellent, so what's the problem exactly?

I'm confused as to why everyone on Slashdot concentrates on irrelevant cpu benchmarks and is so down on Sun at the moment.

Re:Sun is NOT probably doomed

[Temkin]

Sparc performance has always been and still is a laggard compared to the rest of the industry.

Not true. When the SPARC first came out in 1987, Sun enjoyed roughly a 4 to 1 performance advantage. Most PC's at the time were 16-bit CISC, and running between 4.77 and 12 Mhz. The Sun 4/110 was 32-bit RISC, and running at 20Mhz. The 386 came out shortly afterwards, and ran at a blistering 16Mhz. Intel PC's didn't take the lead until the Pentium rolled out. Even then, Sun's are optimized for different uses. A PC is like an F-16. It's really good at taking one person somewhere really fast. A Sun server is more like an airliner. It's really good at taking 300 people somewhere reasonably fast, and effeciently.

Temkin

Re:Sun is NOT probably doomed

[aphor]

I have to disagree.

The only thing that differentiates Sun hardware is that Solaris runs on it, and thus the multitude of Solaris hosted applications runs on it. Take away Solaris and Sun doesn't lead in anything.

What Sun sells is a scalable platform. It is a *hardware* target for developers to *write software* for. It is not *primarily* an Integrated OS/Hardware stack as you suggest. It is trivial to port human-readable Linux code to Solaris, and thus the success of Linux as a development platform does not preclude anyone's ability to implement that software on Solaris. If you can write good, MT code for Linux, then you can test it on a small sub-$1000 UltraSparc system (Running Linux, FreeBSD, OpenBSD, NetBSD... or even Solaris), or maybe a *little* bigger system if you're paranoid about testing CPU concurrency. Once compiled, that software's performance can be made to scale with the harware it runs on in a near linear way. Code for 2 CPUs and 1GB RAM, but the unaltered software will support 32x the load on a 64CPU 32GB RAM box, without re-engineering anything. Instead of worrying about the MPP (Beowulf) style clustering architecture and optimizing the software for that type of system (if it can even be done), Sun provides a stable ABI across all tiers of servers.

"Take away Solaris and Sun doesn't lead in anything." Well, that's true as far as you know anyways. It's a hardware performance niche that Sun occupies.

Sparc performance has always been and still is a laggard compared to the rest of the industry. Pricing at the low-end is wiped by PCs, and features at the high-end, where the profit margins are still fat, are wiped by IBM and HP in the commercial sector and SGI in the technical markets.

This is all very much subject to debate. First of all, you seem to have very narrow definitions of "performance" (which seems like you mean price/performance in comparison to PCs), and "features". IBM can't provide a platform with the scalability potential of UltraSparc. HP doesn't even have a hardware platform any more! Compaq killed Alpha, and HP killed PA-RISC, and even since buying Compaq, HP can't guarantee a niche for Itanium as they are *more* vulnerable to AMD Opteron competition (high-end to low-end) than Sun is to low-end-PCs in their workstation market. This situation is made MORE significant as Linux gains credibility as an OS. Sun is sitting pretty in the niche they currently occupy.

Re:Sun is NOT probably doomed

[Jah-Wren Ryel]

I don't think you have been paying much attention to the high-end of the market. IBM, HP and SGI all provide better performance for the dollar. You can stuff a few more sparc cpus into a single system image machine than you can with the others (well, not SGI they still scale way past Sun). But in terms of most workloads - commercial and technical, the total maximum perofrmance with the leading edge hardware of all the vendors with each vendor's maximum cpu configuration, Sun trails the pack. In terms of flexibile system configuration, sun is also behind HP and IBM - both offer flexible hardware partitioning with fault isolation, Sun has yet to provide a system as robust, although I hear they are working on it. In the server consolidation market, fault isolation is key, without it Sun isn't even a player.

Also, Sun's scalable ABI is no big deal, IBM, HP and SGI all implement the equivalent on their systems and have done so for the last few year. But, unless you are running embarrasingly parallel code, the kind of thing that scales very well on a cluster-of-workstations like Beowulf, you sure aren't going to see linear scaling on a single system-image type of machine - sparc or otherwise.

Re:Sun is NOT probably doomed

[Jah-Wren Ryel]

Intel is not the only competitor, especially not x86 intel. Like I said the first time around, HP, IBM and SGI all scale at least as well as Sun, in some cases better and their building blocks are more powerful to start with. Plus, if you are still dismissing itanic as an under-performer, that would be a mistake. The 2nd generation has floating-point performance that leads the pack and it is reasonable to expect the 3rd generation to continue leading with the FP and bring the integer performance up to parity as well. Sure the chips may require more cooling infrastructure, but that's not a huge factor in system cost, especially in the smaller 4-8 configurations.

As for Sun's market share growing. You might want to check you sources. If you listen to any vendor you will hear how their market share is growing. The problem is that they all pick subsets of the overall market that make them look best. For example, it is common to ignore linux when talking about the Unix market since Linux doesn't have the licensed Unix trademark, or it is harder to measure due to the free nature, or measuring it makes all the other vendors look bad, etc. More important is revenue and Sun has been losing revenue, significantly more revenue than the competition.

Re:Sun is NOT probably doomed

[Jah-Wren Ryel]

Yeah, and that 4/110 was significantly more expensive than PCs - compared to the equivalently priced systems from other vendors that lead was nowhere near as significant. There were plenty of machines that cost more and were much faster than the 4/110 too. They did have a decent price/perf lead with the SparcStation 1, for a while, and it was cool to have one on my desk.

As for the fighter-jet versus jetliner analogy, the same can be said of all the other vendors - SGI, IBM and HP all have systems that cost more than a PC and can support bigger workloads, just like Sun. It is those systems which Sun competes with on that level, not PCs.

On the low end of the range, please don't try to say that Sun's is a better architecture than a PC, ever looked in a low-end sun workstation? It is a PC - VGA card, IDE disk, 32-bit PCI, USB. Just with slower cpu.

Re:Sun is NOT probably doomed

[abulafia]

Just to chime in, Sun provides consistency. You know what you're deploying works, and will work for as long as you care.

Sure, you can do something faster. Or cheaper. Or more elegant.

Sun works, and that's why they're there. Much like Cisco - the hardware doesn't really win any prizes, but the combination of a box, IOS, and support earns the premium they charge.

I think Sun needs to embrace linux and slowly move away from Solaris - It is a great OS, but the industry is moving away from the premises that caused them to create it.

Sun has a wonderful chance to live in the commodity-OS world - they need to learn that hardware exists independent of OS, and service is where you make the buckets of cash.

Re:Sun is NOT probably doomed

[JJahn]

I wouldn't mind using alphas.

Re:Sun is NOT probably doomed

[pilybaby]

... wider bus, lights out management, proper console connection, Solaris stack and future plans for dual core processors and N1. Performance wise, at the low end Sun systems might struggle against x86 but they come with a whole lot more. This make them extreamly nice from a sys admin point of view that makes them excellent systems.

Re:Sun is NOT probably doomed

[voixderaison]

"Getting back to Sun, Solaris is not a revenue piece for them either. There was alot of complaining in the Slashdot crowd and Sun's commitment to Solaris on Intel has waned, but really, would you like to be running Solaris instead of Linux or Debian? Thought not..."

Solaris for Intel should be considered an advertising expense by Sun, and possibly the most important such expense. It doesn't really matter if nobody else runs it at all for anything, as long as there are UNIX administrators who want to be able to learn and experiment and run Solaris at home to sharpen their skills. Sun needs these people, because they are the people who become IT managers and buy Sun instead of IBM or HP.

The day Solaris for Intel goes away is the day the community of Solaris-aware UNIX administrators begins to atrophy. If you see this coming, you'll want to sell your SUNW in lots over a period of months, rather than all at once the the day before this happens, so as not to arouse suspicion.

Re:Sun is NOT probably doomed

[aphor]

All that fault isolation crap costs 1.5 - 2 times what you would pay for two non-fault-isolating internally-redundant servers. Very few actually need contunuous-operation guarantees. If it was all that great, we'd all be running HPUX on Stratus PA-RISC systems. SGI is in a tiny-tiny niche occupied by their Cray legacy clients, and also in a niche for render-farms and technical compute farms that are highly (but not as you put it "embarassingly") paralell for software designed from the ground up for N-way concurrency.

The rest of the world wants a fat database server that will be alternately Disk-IO/Network-IO/Memory bound. BTW, Sun has the best, most mature, most stable 64bit addressing architecture to help the transition from a 32bit address space. You dont gain or lose your lead in something like that over one or two generations of hardware. It takes 5-10 years to gain and assert that kind of dominance. Will Sun be so laggard for 5-10 years that it justifies switching now? The world will wait and see, thus Sun holds the brass ring. 64bit needs haven't even been widely expressed, but as storage, and bandwidth gets bigger and faster, UltraSparc will continue to be the standard by which others are judged.

I'm ranting. I have to stop now. Please try to pick the good stuff out for criticism and ignore the ranting crap as much as possible.

Re:Sun is NOT probably doomed

[Kragg]

Remember iPlanet? Approaching 0 on the latest web surveys as Apache dominates that space.

iPlanet was rebranded by Sun as the Sun ONE line - not just the webserver but the appserver, webserver, mail.. sorry 'Messaging' server, and also includes their new IM server talked about here.

Admittedly still at 1% in the webserver space (from here), but you can be sure that that 1% is SME and greater - paying customers. Sun do just fine out many of the large chunks of business software like these that they sell (naturally together with their hardware).

Re:Sun is NOT probably doomed

[Jah-Wren Ryel]

TPC - Sun has 5-6 entries across the non-clustered top-10 type results across all tests and about 2 or 3 are #1, all the other slots 50-60 are dominated by IBM, HP and DELL with a smattering of Unisys.

Just because your shop is buying 15K's doesn't mean much because A) It's anecdotal and B) My premise was that Solaris is currently a differentiator, but that's it - everything else they do is done equally well or better by their competitors.

Re:Sun is NOT probably doomed

[Jah-Wren Ryel]

Instead of waiting, perhaps you would like to read what I wrote again. When I mentioned partitioning with hardware fault isolation as lacking on Sun but present on HP and IBM that was your clue as to how Sun lags. Look it up, read the docs, you'll see that everything is as I said.

Re:Sun is NOT probably doomed

[Jah-Wren Ryel]

I'm not quite sure what your point is except - "Sun has the best most mature, most stable 64-bit addressing architecture to help the transition from a 32-bit address space." Perhaps you can quantify what makes it the best, as far as I can tell, HP and IBM have the same level of robustness and functionality in terms of 64-bitness.

In terms of performance for big-fat databases, TPC-H shows Sun's latest 72-way 15k results as 7% faster than HP's 64-way results from 7 months prior. The 72-way results are also a little unrealistic given that configuring a 15K to use 72 cpus rather than just 64 cpus really cuts down on the available i/o interfaces, the 72-way config was designed for techincal workloads, not commercial - obviously it can be used that way, but that may not be a common . HP's next hardware rev is due RSN (the configuration that Sun used won't be available until the end of this month either) and will, assuming the comp.arch discussions are accurate, be dual-core which should result in significantly more than even a 15% increase.

As my original premise was that Solaris, and thus the catalog of Solaris-hosted applications are the only thing that distinguishes Sun from the other competitors in the marketplace, I think these numbers support that pretty well.

Re:Sun is NOT probably doomed

[aphor]

Ultrasparc places in the low, mid, and high-end server ranges on TPC-H, but IBM and HP will sell you a PC to compete with Sun on the low end, and IBM doesn't even place in the 3TB high-end systems. HP will sell you a PA-RISC system to compete with SunFire 15K. With IBM and HP, you have to rewrite software from the ground up to scale from low to high end of their Sun-competitive server offerings. It's not the same. Apples and oranges.

Oh, BTW: TPC-H results

Re:Sun is NOT probably doomed

[christophersaul]

Not sure about SGI - but they are irrelevant in business computing anyway, but IBM and HP do not scale as well as Sun. They do not have the linear scalability of UltraSPARC and do not have 106+ cpus in one system. The 128 cpu Superdome is two systems connected together by a highspeed interconnect. Fujitsu have a 128 cpu box however, which Solaris can take full use of. HPUX and AIX can't.

As for Itanium, heat is an extremely important factor as it means that you're going to have to build larger systems that consume more power so that you can cool them. If that means your 4 way takes up twice as many RUs as the competition, as well as sucking up a huge amount more power, that's a factor. What OS were you wanting to run anyway? There's no proven track record for Itanium systems and hardly any apps.

Sources such as Gartner and IDC are fine for me regarding Sun's market share. Sun's revenues also increased - the loss was due to costs of acquisitions and Sun actually made an operating profit. As for Linux market share, you're right it is hard to measure. But since it runs on generally on 1-2 way x86 systems, you track the revenue of how those Intel based systems are selling to get an idea of the hardware market.

Re:Sun is NOT probably doomed

[christophersaul]

Partitioning with hardware fault isolation lacking on Sun? What docs show this? Since Sun leads with its dynamic system domains, your comment sounds unjustified.

News Flash: Linux still not ready for the desktop

[lpret]

The most likely outcome is that customers will face a choice between Linux, which is cheap and cheerful, and Windows, which offers more bells and whistles, is tightly integrated with other Microsoft products and is easier for unskilled staff to use, but costs more. In short, Microsoft will be not so much a loser from Linux as less of a winner. In the server market at least, Linux is providing Microsoft with some much-needed competition.

I think this sums up the consensus of the article -- Linux is coming, but not to the desktop.

Sun isn't doomed

[Sir_Bill_William_Jen]

Sun will be around for quite a some time, because they offer specific product for one market on which they have quite a monopoly.

Linux will probably take some share from Microsoft but It won't be so huge that it will be used more than Windows. Microsoft is too smart to let Linux take hold of them. They will do everything to stop Linux from taking their share, whether it be through Advertizement, Deals with companies, or Palladium.

However Unix might fade away, BSD and Linux provide great alternatives for Unix... But they won't be completely gone.

Re:Sun isn't doomed

[El_Ge_Ex]

Sun will be around for quite a some time, because they offer specific product for one market on which they have quite a monopoly.

What market is that?

1. Servers? (DELL,IBM,HP)

2. Java? (it's free, and IBM has been doing a better job with it lately.)

3. Workstations? (Anything with Linux, Opteron :) )

4. OS!?! (hehe, had to mention it ;) )

So? What market does Sun own again???

Re:News Flash: Linux still not ready for the deskt

[Sir_Bill_William_Jen]

I think that Linux will take away some Desktop Share from Microsoft, however it won't be so significant.

Linux is ready for the Desktop, but is the Desktop ready for Linux?

the conditional tense

[10am-bedtime]

my crack team of critical media analysts have identified the usage of conditional tense in this article. all the "could be", "might be", "would be", etc were boiled down to the simple executive summary: NEXT!

no, i'm not talking about the old cubes from cupertino, just that this class of article has basis in speculation and not in economic law, as the the title of the magazine would have you believe. thus, you can safely skip it (and the rest of this post...)

the nice thing about economic law is that, regardless of interpretation by economists, manipulation by local power regimes, and so forth, zero cost is still zero cost is still zero cost is still zero cost. so what if usgov and usloth prop each other up; the rest of the world can understand this economic law, live peacably within it, and mutually avoid irrelevance through this understanding.

it's easy to be irrelevant (take it from an expert) the trick is how to do it and still have a good time w/ the other ones and zeroes on this rabid rock.

Re:the conditional tense

[voixderaison]

the nice thing about economic law is that, regardless of interpretation by economists, manipulation by local power regimes, and so forth, zero cost is still zero cost is still zero cost is still zero cost.

"Free" doesn't have the same, finite meaning to an economist that it does to an accountant, or a student seeking beer.

Economists sometimes consider "external" costs. These are costs to the economy of an activity which may not result in a ledger entry (say, a direct cash outlay) under current accounting conventions, yet which may be valid costs to "the economy" as a whole. They may be costs actually borne by some third party not directly engaged in the given transaction, as in the case of pollution cleanup costs borne by the tax payer, or less easily measurable such as cancers caused by the pollution. Another commonly sited example would be sales of timber by the Federal government to a lumber company, which then cuts down an acre of forest. The public receive some amount of cash which may not reflect the total value of the acre of forest to the public, since the current accounting practices don't consider that forests are finite and that it may take hundreds of years to grow that acre back, or even if it can be grown back at all (some can't). Economists think about all the other things that can be done with a forest, and the value of those things, and the cost of putting the forest back on that acre.

If we borrow the Economist's perspective on "not really free" we can consider the open source world with respect to Linux, and the opportunity cost of all the time and energy flowing into Linux. Might Linux be a largely redundant "free UNIX-like operating system"? Would the open source movement be further ahead if all that energy had been put into BSD? Hard to estimate. Forward progress and innovation happen in both forks of the happy UNIX family, with advances shared to a significant degree. Clearly, however, there is also significant redundant effort. Maybe OS technology would advance faster with all that effort focused on a single open source operating system project.

On the other hand, there may be some upper limit on coordination of people (for example) that would prevent such increases in the pace of technological advance. Maybe the ideas just don't happen any faster. Even so, that energy could have been applied to some other open source effort, say a common object layer, or a next generation, high performance display architecture to replace X11, and then common user level applications built on top of that stuff, or maybe ease of use could be improved. These things have real value, and the extent to which they happen slower, or do not happen at all would be a real cost -- an opportunity cost -- to an economist. (Assuming of course that any economist thought about the contribution of open source to the GDP, which perhaps by now some do.)

Yet another way to spin "free" is that businesses and governments don't consider Linux to be "free" just because they can download an image of the installable OS. In fact, the licensing cost of the OS is not always even a significant line item on the cost sheet. This kind of "not really free" considers all the costs that go into a business activity. From this angle, consider the impact of platform choice on other costs like support, maintenance, development, migration to new platform, and downtime.

Free isn't always clear.

Sun increase market share against x86 last year

[ChrisRijk]

In the "low-end" of the market (systems costing under $100,000) Sun increased their revenue market share last year (2002 over 2001). Ie they grew faster than the industry average. Almost no Intel, Windows or Linux based systems are sold for more than $100,000.

So if Sun was being hurt by Linux (or x86 based systems in general) why did their market share increase?

It wasn't anything to do with the LX50 (dual P3 Solaris x86 / Linux) systems they launched late last year - they only shipped a bit over $1m worth by the end of the year.

Re:Sun increase market share against x86 last year

[Master Bait]

What happens with Sun depends on price commoditization of the 4-way and 8-way space with the forthcoming AMD Opterons. If that really happens, and Sun doesn't adapt, then that segment could go away for them.

"Open Systems for Open Minds" ...

[Big Jojo]

Sun's just being forced to go back to its roots, which were running a commodity Open Source OS (SunOS started as BSD) on commodity hardware (m68k at first, then SPARC except for a couple years when they also sold 80386 hardware, as they started Solaris). That should be a healthy thing, long term, though they have to get rid of a lot of closed-system attitudes. Like the ones that have crippled so much Free Java work.

If Sun had kept true to their roots, they'd have been running Linux on x86 from day one ... instead, they wanted to keep one founder (Andy Bechtolstein), who wanted to design a RISC chip (became SPARC). So Sun sold out SunOS in favor of Solaris/SVr4, so they could switch to non-commodity hardware. Well I've got news for you: Andy's long gone, and SPARC was never that hot. And the customer lock-in is going away ... customers always wanted the open systems approach, even when Scott McNealy refused to play that game.

Re:"Open Systems for Open Minds" ...

[Anonymous Coward]

"they wanted to keep one founder (Andy Bechtolstein), who wanted to design a RISC chip (became SPARC). So Sun sold out SunOS in favor of Solaris/SVr4, so they could switch to non-commodity hardware"

Everything in this is deadly accurate except for Bechtolsheim's name, the part about him designing SPARC, SunOS not being a part of Solaris, the need to change operating system flavors to change chips, why Sun went from Moto to SPARC, and the comment about commodity vs. non-commodity.

Without going too deeply into the other problems with the above, the reasons that Sun went from SunOS 4.x, which was based on BSD 4.x, to SunOS 5.x, which was based on SVR4, were manyfold:

1. There was a perceived need to redesign the kernel in any event. The 4.x kernel was single-threaded and showing its age; Sun wanted to deliver a more rationally-designed, modular, multithreaded kernel.

2. Many Sun customers were asking for a standards-blessed UNIX. At the time, *the* standard for UNIX was System V (I'm not saying it was a wonderful standard, just that it was the only de jure UNIX standard). AT&T still owned the UNIX brand, and the BSD fork was seen as less "official." For example, KMart had just thrown AIX out of a major deal because it wasn't System V compliant.

3. There was an opportunity for Sun to make significant improvements to System V, and boy did it need it. SVR3 was missing a lot of the goodness present in BSD, and AT&T was inviting Sun to colloborate on SVR4 and merge the two diverging paths back together, to the benefit of both.

By the way, Sun made a significant gaffe at this point that had a major effect on the future of UNIX. When the Sun/colloboration came to pass, Sun promptly alerted the world that this would become the only UNIX that would ever matter, and that they would be one of the major drivers. This so annoyed/scared the other commercial UNIX vendors that they promptly formed OSF [aka "Oppose Sun Forever"] to create Yet Another UNIX. So much for unification. In Sun's defense, this is pretty much exactly what IBM had just said about AIX a couple of years earlier, and what HP is now saying about Itanium. Doesn't make it right, it just doesn't make it rare... .

Bottom line: the development of SunOS 5.x, aka Solaris 2.x, had nothing to do with SPARC at all. When the first Sun SPARC systems came out in 1989, they ran only SunOS 4.x, the BSD based derivative; in fact, IIRC SunOS 4.0 was delivered specifically to support SPARC. SunOS 5.0/Solaris 2.0 didn't come out until three years later, in 1992.

Re:"Open Systems for Open Minds" ...

[cgori]

Wow. I don't know what fantasy world you live in, but it sure must be fun. You did start with a valid point, that original Suns used all-commodity everything (Motorola/BSD/VME/Ethernet), but then you faded fast.

Let's get some things straight. Andy Bechtolsheim (note correct spelling) designed the first Motorola-based Sun1 as a side project to his PhD thesis at Stanford, before he met Scott and Vinod. So I think he knew a little about the tradeoffs they were making in the systems. He's generally considered the most technically savvy of the 3/4 founders (Bill Joy would be the "unofficial 4th," and he would still be 2nd behind Andy in technical ability).

Andy also had the idea for the first SPARC-based system chip, but didn't design the processor himself. At the time it was probably a reasonable project to green-light. This was an era of 16-20-25MHz 80386s. RISC performance was potentially considerably better than commodity CISC (Moto/Intel), so while it would be expensive and difficult, it wasn't such a bad idea. The first SPARC chips were considerably better than what you could buy commodity-wise. On the IO front, SBus blew the doors off commercially available IO buses at the time, including VME (this is the ugly era of EISA and VLB, remember -- 33MHz PCI is maybe just coming onto the scene in 1992)

The switch to Solaris (as someone else points out) had nothing to do with the switch to SPARC. SunOS 4.1.3 was on every Sparc1+ and Sparc2 I used in all my time at school. Solaris really got rolling with real momentum sometime in 1994-5, and that only because Sun knew they had to end-of-life SunOS4 to force everyone to switch. Everyone really liked the BSD-isms, but SVR4 clearly has scaled up damn well.

And as far as Andy being "long gone" -- he left in 1996 after ~14 years with Sun, having had a hand in every top-selling product Sun made in that timeframe. I hate those short-timers who only spend a decade and a half on a company.

Re:"Open Systems for Open Minds" ...

[Big Jojo]

The switch to Solaris (as someone else points out) had nothing to do with the switch to SPARC.

False. Sun didn't have the money to invest in SPARC engineering ("productizing the <university/> RISC machine", if you will) until it got a boatload of $$$ from ATT as part of its work to merge SunOS features into SVR4. That kind of hardware work isn't cheap, especially when it includes transitioning existing product lines (Sun3, Sun386i) and customers. How many companies can start new chip designs today? Even Intel can't afford it, look at Itanic! Even though it was cheaper back then, it wasn't in the least bit "cheap". Co-development funding was needed.

If ATT hadn't wanted to kill of the main BSD variant, Sun would likely still have been an "open systems" company. And if you don't recall that, you weren't at the company when that was all happening, or in the industry watching the dynamics of SVr4 and the "Oppose Sun Forever/1" OS.

One of the strings on that contract was that all new hardware had to ship with SunOS 4 (the SVr4 based code) starting at some particular date. The Sun386i predated the SPARC1 you mentioned, and it was one of the last boxes that could ship with 3.x (the "old" BSD codebase). Sun wasn't allowed to continue trying to scale the BSD code -- the ATT contract said so. Those Sun4 machines couldn't (legally) be sold with the SunOS 3.x kernels.

As for performance ... odd, the Sun486i was faster than the Sun 4/110, not that it ever shipped. Quite a lot of Sun386i hardware upgrades had to get held back to continue making the early SPARC hardware look good by comparison. (In fact, I don't think any Sun386i product upgrades really appeared.) Same thing applied to I/O and memory busses, not just CPUs.

You seem to think that being gone for seven years doesn't count as "long gone" ... heh. His influence on current designs is rather indirect. Minor apologies for not spelling Andy's name right, one's used to just saying "Andy".

Re:"Open Systems for Open Minds" ...

[Anonymous Coward]

Sigh...

False. Sun didn't have the money to invest in SPARC engineering ("productizing the RISC machine", if you will) until it got a boatload of $$$ from ATT as part of its work to merge SunOS features into SVR4.

You sound quite authoritative on this, something that to be verified would require being awfully tightly into the company's balance sheets and strategies. Can you explain your sources?

Given that AT&T owned a serious chunk of Sun at that time, was this the only way they could drive this deal? Seems unlikely.

If ATT hadn't wanted to kill of the main BSD variant, Sun would likely still have been an "open systems" company. And if you don't recall that, you weren't at the company when that was all happening, or in the industry watching the dynamics of SVr4 and the "Oppose Sun Forever/1" OS.

Happens that I was.

One of the strings on that contract was that all new hardware had to ship with SunOS 4 (the SVr4 based code) starting at some particular date.

That would be "SunOS 5," but I get the idea. Once again, your source on this?

The Sun386i predated the SPARC1 you mentioned, and it was one of the last boxes that could ship with 3.x (the "old" BSD codebase). Sun wasn't allowed to continue trying to scale the BSD code -- the ATT contract said so. Those Sun4 machines couldn't (legally) be sold with the SunOS 3.x kernels.

Sun4 machines could run SunOS 4.x, which was BSD-based. The first machines that truly could not run SunOS 4.x were the Ultra 1 systems, which started shipping in late 1995. Once again, source on the AT&T contract having this restriction, which seems to go against future events, would help here.

As has been mentioned before, lots of machines that could run BSD-based SunOS came out after the 386i: all the SPARCstations, for example, and the entire Sun-4 product line, as well as the last Sun-3 Moto-based systems, which came out after the 386i.

As for performance ... odd, the Sun486i was faster than the Sun 4/110, not that it ever shipped.

How odd was that, really? The Sun-4/110 was the slowest SPARC machine ever released, 7 MIPS with no integrated FP; I got to have one on my desk for a while, lucky me. The 486i by the time it was released would have been a contemporary of the next generation of faster SPARC machines, the SPARCstation product line.

So what this all boils down to a conspiracy theory that AT&T demanded the death of BSD, which somehow tied into a plot to kill everything not SPARC. If the latter were true, please explain the Sun-4/80 and Sun-4/370, both of which feature Moto 68040 and were competitive with SPARC performance in the repective niche, and were on Sun's price list into the '90s.

I think it's more likely that the 486i was killed because demand for the 386i had tapered off; it wasn't well-integrated into Sun's product plans, and was more a victim of East Coast/West Coast internecine battles than anything you've proposed so far (it's a good day when I can work "internecine" into a conversation :-). But the easier to swallow explanation was Scott's famous "all the wood behind one arrow" edict, which made good sense for a $1B company to follow, and served Sun and its customers well for over a decade. This is the one part of your story that makes sense, that Sun would want to go with one chip architecture rather than the way things were being done at that time, which was to have three separate chip architectures. It also made sense that that architecture be SPARC, for many, many reasons, none of them having to do with AT&T that I am aware of. If you can substantiate this, I'd be interested in hearing about it.

Re:"Open Systems for Open Minds" ...

[Big Jojo]

I could use a cluebat here. All Sun4 machines post-dated the Sun386i, and I never saw one run SunOS 3.x ... it was all based on the SVr4 code. (Though not the earliest versions) Don't be confused by the way the SunOS 4.x code merged the BSD-isms in with the SVr4-isms ... it was SVr4 at core, with a BSD veneer. Later versions of Solaris evolved even more.

No Sun4 included a Motorola chip; that'd have made it into a Sun3. You don't know what you're talking about.

I never denied the politics or the "one arrow" strategy impact; just pointed out that the way it was implemented had a magic "... and the answer is SPARC" clause, which is what nuked the original "Open" systems thrust of Sun. RISC had a marketing cachet that wasn't wholly deserved, although at that time it was more deserved than today.

Sources ... clearly you won't believe people that were there at the time. So failing that you'd want a copy of the relevant contracts and presentations. Sorry, no.

Oh

[KoolDude]

Sun is probably doomed.

Umm... but... I thought Doom was going to be MS exclusive

*logs off and runs away*

Re:News Flash: Linux still not ready for the deskt

[drooling-dog]

I think this sums up the consensus of the article -- Linux is coming, but not to the desktop.

Well, after years of dual-booting with Windows, Linux is the now the only thing going on my desktop, and I've gotta tell ya, I'm doing just fine (better than ever, IMHO). Maybe I'm dreaming? Or maybe the Microsoft fuddites don't know what they're talking about.

Microsoft Prices and Competition

[GregBildson]

Thankfully, Linux has prevented Microsoft from dominating the low end server market like they wanted. From the low end, Microsoft was then going to push into the high end. Can you imagine how much Microsoft would be making off this area if Linux had not stopped them?

On the desktop, we just bought 5 rather powerful developer PCs for $600 each. Of that price, $150 was for Windows 2000 (not XP thank you). 25% of the price is a rather large part of the cost. The decision between Linux and 2000 was pretty close on these boxes and getting closer all the time. Pretty soon, developer workstations could well be all Linux with OpenOffice and the like. I think the competitive threat to Microsoft will soon restrict their desktop and office pricing.

Re:Microsoft Prices and Competition

[RoLi]

Exactly.

With hardware becoming cheaper and Windows becoming more expensive, it will be soon 30%, then 40%,...

A Lesser Form of Unix

[MBCook]

"But for many applications, Solaris is overkill, and Linux, a less capable flavour of Unix, is good enough."

I don't know much about Solaris, so I'd like to ask you guys out there. What makes Linux less capable? What does Solaris do that Linux can't do (at least well enough)? Just wondering.

Re:A Lesser Form of Unix

[pherthyl]

I would think SMP support would be much superior on Solaris. Since it routinely runs on 32+ processor systems..

Re:A Lesser Form of Unix

[jc42]

While it's not nearly the whole story, an interesting benchmark is the top 500 supercomputers list. Sun and linux are both on the list. Sun seems to be the main supplier of supercomputers these days. But they first appear on the list as number 156. There are only two linux systems, but they are numbers 5 and 45.

I guess that doesn't tell you all that much about the top end, other than that linux and Sun are both quite capable of supporting raw, top-end gigaflops power. The actual computing is done by processes, of course, so this benchmark really just tells you that neither OS has any showstoppers. Neither interferes with the ability of a process to crunch bits and bytes.

Now on to the other benchmarks ...

Re:A Lesser Form of Unix

[repetty]

Go to Sun's web site; explore their product offerings and read the specs.

I did. I concluded that Linux and Intel hardware are simple toys compared to Sun stuff. They are not to be confused with each other in features, reliability, or price.

Linux is an alternative to Windows, not Sun.

Richard

Re:A Lesser Form of Unix

[sql*kitten]

I don't know much about Solaris, so I'd like to ask you guys out there. What makes Linux less capable? What does Solaris do that Linux can't do (at least well enough)? Just wondering.

Well, scalability is one major thing. Linux struggles with more than 4 CPUs and more than 4G of memory; Solaris handles hundreds of CPUs and a terabyte of RAM. Linux lacks the manageability of system resources offered by Solaris 9, which allows system resources to be prioritized for different tasks, with a guaranteed minimum available. And even if Linux could do this, it doesn't run on hardware than can be dynamically partitioned, unless it runs as a guest on z/OS, and in that case it's z/OS doing the work. Tight integration with the underlying hardware is another advantage for Sun; they know precisely every component in every system that Solaris runs on, because they designed and built it, so there are never compatibility issues. Solaris' high-performance, high-reliability filesystems are proven, not just betas (yes XFS is also proven, but in IRIX not Linux). Speaking of filesystems, Solaris has ACLs, whereas Linux just has the relatively crude user-group model. Linux doesn't have remote shared memory or IP multipathing (IIRC).

Re:A Lesser Form of Unix

[the gnat]

Linux is an alternative to Windows, not Sun.

Not if you need a simple workstation or webserver. I can see why people would want a Sun box for high-end stuff - we just bought a couple of SGI Origins for file serving, but would have been just as happy with Suns - but I can see few reasons to buy a simple 1U Sun server, other than binary compatibility with larger systems.

In addition, as the article points out, applications that once ran on giant multi-cpu boxes now run on clusters. Clusters can be a fucking nightmare, but the point is that if you need a lot of computer you don't necessarily have to buy it all in one box unless your applications specifically require many processors with shared memory. Google's Linux farm is a good example.

Re:A Lesser Form of Unix

[raptor21]

It depends on the hardware you are talking about. Solaris is an enterprise class OS. What that means is it is more suited to run on an enterprise class systems and run an enterprise on.

Apart from supporting oodles of memory(tera bytes) and CPUs (106). An enterprise class system needs to provide Enterprise class serviceability. Serviceabilty is probably the most understated of all the things talked about on slashdot.

How does linux point out to the service engineer which memory module of the hundreds on an huge server is faulty so he/she can replace it? What about debugging. Say the kernel panics for whatever reason how do you debug it? Having the source is no help if you can't get a stacktrace and debug a live system at a debug prompt. On Solaris you send the core dump to Sun or the FE/Sun Engineer can boot into the OS run mdb/adb on the coredump and debug at the customers site. Or boot with kadb and debug a live system. How do you do that in linux.

What about architectural things like changing the VM system mid kernel release? Say I have a panic on a customers site caused by a VM bug for which the fix lies in an kernel version 2 releases later. Now linus in all his infinte wisdom decided to accept a full VM rewrite in between then. Now The customer is screwed. We can't fix this because we can't move the customer to the latest release as it would need a full regression test to be done. We are not talking about a PC here, we are talking about hardware that costs hundreds of thousands to millions probably running a billion dollar business where uptime is important.

So until linux can solve these problems. I think Sun need not worry, eventually they should but not now.

Re:A Lesser Form of Unix

[jlanthripp]

If you read the post you replied to more carefully, you might notice that it mentioned Solaris - not Linux - as the OS that routinely runs on 32+ processors. I can name one vendor off the top of my head that sells 32+ processor boxes that run Solaris as a single system image - Sun Microsystems

Re:A Lesser Form of Unix

[ianezz]

Linux struggles with more than 4 CPUs and more than 4G of memory;

Please, put it on comparable hardware. Or let's talk about the very same problems popping up on Solaris for x86.

And even if Linux could do this, it doesn't run on hardware than can be dynamically partitioned, unless it runs as a guest on z/OS, and in that case it's z/OS doing the work.

Like somebody is going to care if it's doing it all by itself, just for the sake of it?

Tight integration with the underlying hardware is another advantage for Sun; they know precisely every component in every system that Solaris runs on, because they designed and built it, so there are never compatibility issues

That's the dumbest thing you could say: it's like saying that a good reason to choose a certain car is because there are no suppliers of spare parts for it that sells directly to you (but instead you have necessarily to go through the car "manifacturer"). It's not like you are going to replace the engine with a third-party one, but what about the windshield-wipers or such? Please don't tell that since you are spending X grands this doesn't matter, otherwise nobody would ever consider using something like Linux in the industry.

Solaris' high-performance, high-reliability filesystems are proven, not just betas (yes XFS is also proven, but in IRIX not Linux).

Yes, but where high-performance is a concern (video), people are moving to Linux, not Solaris. SGI has been wise enough to port XFS, because it knows Irix has no long-term future within the video industry.

Solaris has ACLs, whereas Linux just has the relatively crude user-group model.

And thanks to the No Such Agency, SELinux provides mandatory access control policies on more than just files. And they actually use it.

Re:A Lesser Form of Unix

[jlanthripp]

Good thing you're not a Patriot missile operator, eh? *grin*

Re:A Lesser Form of Unix

[jak163]

How does IBM run Linux on their mainframes if Linux can't support more than 4 processors? Did they modify it? Is this what the recent dispute with SCO was about?

Re:A Lesser Form of Unix

[sql*kitten]

How does IBM run Linux on their mainframes if Linux can't support more than 4 processors? Did they modify it? Is this what the recent dispute with SCO was about?

Linux is run as a guest of z/OS - z/OS sees all the processors and creates a number of single-processor virtual machines on which Linux kernels run. This is both good and bad; good because individual services can be completely isolated from one another, leading to ease of management for security and QoS (Solaris 9 Resource Manager can also do this), but bad because you are still locked into the proprietary vendor because you need z/OS, and now you have two OSs to look after.

Re:A Lesser Form of Unix

[fferreres]

It comes at 22, and at other places too. The Dell cluster uses a mix of servers, and it's not really a supercomputer (or I got confused at their page?) but a cluster of computer. It's nice to note that the IBM system at stop #5 is dated in 2000 while the rest of the systems are all 2002. So it's still at #5 after two years...nice uh?

Sun doesn't support Linux?

[Biolo]

Ok, Sun is not 100% behind Linux (yet), but that's because Linux isn't ready for the high end (yet). By high end I'm talking about F15
and F12K
servers. It is pretty close to having the capabilities to run on the Sun midframe stuff, for example I'm sure it would run fine on a 3800,
maybe even the 4800, but you start to reach its current limit with a fully stuffed 6800 system.

Now, step back for a minute and think why Suns UltraSparc and Solaris solution is so strong. Simple, at the risk of repeating the marketing guys the lure is that you can give your development and deployment guys a bunch of cheap Sunblade 150s or some cheap UltraSparc blades and whatever they come up with can be moved straight onto anything up to and including an F15K without recompiling. Put yourself in the place of a big corporation. Your putting together a new system, you have no idea just how big a load it will eventually have to take (say in 5 years). Today, sure you could run it on high end Linux box, but what happens if 6 months in the system needs a bigger box? If you chose Sun in the first place you simply buy the bigger box and move over. No porting, no redevelopment, and you know there is always a bigger, faster system you could move to. It buys you severe scalability that Linux isn't placed today to provide.

Now, about not supporting Linux, what about the LX50, the Sun Open Desktop that is coming soon, the Lintel blades (Coming Soon(TM)) the fact that the entire Sun One stack (web, directory, identity, etc, etc, etc) is either available now for Linux or coming soon, not to mention Star/OpenOffice.

So what is the perceived issue? I think people don't see Sun offering Linux on the UltraSparc range and thing they don't get it. Sun does get it, but look at their selling point for the last 10 years, total scalability. Linux doesn't provide this yet so they can't buy into it. What they are doing is making Solaris as compatible with Linux as possible, whilst at the same time helping Linux by providing software (openoffice, SunOne and much more) and I believe some kernel code too.

Believe me, when Linux is ready for the F15K class systems Sun will be ready for Linux to be there.

Disclaimer - I work for Sun, but nothing I have said here is not already public information.

Re:Sun doesn't support Linux?

[christophersaul]

Nice post.

Now cue all the posts informing us that all you need is a few single cpu Linux boxes and all the world's applications suddently scale ad infinitum.

Sun, not dead, just a change of direction.

[Zebra_X]

Sun really needs to rethink their strategies.

Solaris for Intel? None of the sparc binaries run on it, it's not any faster than linux. Linux 1, Solaris 0.

Solaris for web applications... absoultely not. Tried and true OS for sure. Though web sphere, atg, web logic and most other large scale app servers have linux ports. java's relability will never exceed the uptime of an OS i.e. the JVM or app server will crash before the OS does. That then precludes having a bullet proof OS. Web applications need redunancy, both from a geographic perspective and application. Doing so requires a duplicate hardware investment. Not such a good deal with sun. Linux 2, Solaris 0.

Maintainance... will be cheaper with linux rather than solaris. The reason being that Linux and it's friendly varients are all freely avaialable. To learn and use linux is not a big deal, solaris on the other hand need solaris hardware to run. As a result of easier access to hardware and software (linux) labor costs go down because the skill sets require to administer and maintain linux and linux apps are more freely available. The same is not true of solaris. Linux 3, Solaris 0.

Solaris, and it's hardware IS good for massive multi-proc applications. Data Warehouse with Multi-tera bytes of data? Linux and Intel are not suited to such tasks. Large transactional databases that require nearly 100% uptime and reliability, i.e. the database is nearly as reliable as the Operating system. Solaris is the OS for that application. Linux 3 Solaris 1

Sun is no longer suited to playing in the high(er) growth markets of dedicated servers, web applications, IT support devices (dns, dhcp, network management) and such. Their role is increasingly being boxed into ultra highend applications where a large number of processors, ultra high reliability and what sun has stood for still means something. Where the applications are almost as reliable as the OS, and that the OS and hardware is required to be up nearly 100% of the time and never unexpectedly. The difficulty they face is that that the role just described is not in particularly high demand. As IT budgets continue to shrink - decision makers are going to continue to look to linux to solve their problems.

Linux is cheap - costs less to maintiain - and the hardware can be repurposed. Sun just can't argue with that. Sun needs a change of direction.

Re:Sun, not dead, just a change of direction.

[GregBildson]

Totally agree. Sun's boxes are now only applicable for the super-highend and for the mission critical high uptime price insensitive. Sun will continue to make good money here. Their hardware business needs to target mainframe and supercomputer class users. These are the only areas where high profit per unit will arise.

It has always been possible for Sun to build a good software business. That remains a possibility but the odds are shifting against them (other than with StarOffice). Imagine if they had ever built a compelling suite of Java tools or applications on top of their APIs. To some extent, they have tried. They probably should have bought BEA Weblogic back during the bubble.

Sun will definitely still be around. They still have possibilities but I think they are aiming too high perhaps right now. Their dedicated customers still have value even as they decide to buy more Linux boxes.

Re:Sun, not dead, just a change of direction.

[aeoo]

Solaris, and it's hardware IS good for massive multi-proc applications.

Linux is steadily adnvancing in this area. It does not stand still. So at some point it will be as good as Solaris in that regard. Look at 2.6 kernel. Try to imagine where 2.8 or 3.0 will go. Now tell me how should Sun change direction?

Re:Sun, not dead, just a change of direction.

[Zebra_X]

The restriction that I was refering to was primarily hardware. Intel hardware, the most common and definately most compelling hardware platform to deploy linux on does not scale particularly well. If a machine was needed that required 32 processors IBM or Sun would likely be the only two choices. It will be a long time before linux will be able to play in this arena. There will always be a need for these machines.

Re:Sun, not dead, just a change of direction.

[Zebra_X]

Sun has demonstrated a particular ineptitude in software development. For example Java has become a rather messy collection of API's. Some might call it "robust" however Sun constantly defines new API's touting them as "innovations". However there is no leadership in execution. Why is it that developers are forced to rewrite sun written code to enchnance performance and make their applications reliable? The truth is that Sun is very fond of "reference" implentations. I.E. they are half assed, untested and unreliable. To get really good perfomance other have to start from the ground up. Blackdown, IBM's 1.3 JVM comes to mind, Caucho's XSL Parser and there are others. Sun's leadership with repspect to software is really none at all. Setting up a portal and creating "standards" through API's isn't really adequite.

Re:Sun, not dead, just a change of direction.

[Zebra_X]

I've done as you asked:

Sun v120 (There is no v150) 1U
SINGLE PROCESSOR CAPABLE @ UltraSPARC IIi 650 Mhz
1 Gig RAM
2x 36 gig SCSI 10K Drives
2x10/100 On Board-Nic
1x1 Gigabit Ethernet ( added to compete with dell svr.)
UltraSCSI II (No Hardware R
aid)
cd-rom
No redundant power supply
1 yr. wrty.
Price as configured: $5,757

DELL 1650 1U
DUAL Processor Capable w/ ONE processor @ P4 -1.24Mhz
1GB RAM
2 x 10K 36 GB Drives
Hardware RAID - UltraSCSI II On-Board
2x 1 Gig E-net On-Board
cd-rom
redundant supply available
3 YR. Warranty
Price as configured: $3,640

basically for every two sun servers - you could get 3 dell servers. they cost less to upgrade too.

under most circumstances i'd recommend that drives and supplies be redundant - it doesn't hurt and can only help. Sun charged 1K for the extra drive @ purchase time. Most notably sun lacked the one gigabit on-board ethernet.

A server's sole purpose is to push bytes and process. slower nic's = slower performance. The lack of mutli proc on sun's part isn't surprising - you have to pay for expandability. Also of note is the lack of redundancy i.e. Hardware RAID support and redundant power supplies.

The v120 is their "entry level server" that is clearly lacking in expandability, and reliability. As stated before - sun hardware at the web server level cannot compete. just as sun does a shoddy job with their software - they do the same with their hardware - and they won't even warranty their hardware beyond a year. Repairs aren't cheap.

As for maintainance - starting adim salary for linux is lower than that of sun. also, because linux is freely available - the likleyhood that a candidate has spent more time working with linux boxes in and out of a work environment is higher.

there was a study posted on /. a few months ago about the cost of ownership for windows, linux and solaris. solaris wasn't even in the running. on going costs were vastly higher than the other two platforms.

Re:Sun, not dead, just a change of direction.

[Zebra_X]

I've left the rest of the computing world out of the discussion becuase I think that the markets that we are talking about, those are web servers, network utility machines i.e. DNS, DHCP, Firewall and to some extent File Servers don't have any other real competitors but M$.

AIX, IRIX are the highend OS's that Sun needs to go compete against - Sun has the hardware line and the OS to do so. They continue to waste their money with lowend servers that will never reach a price point that is competitve against AMD and Intel.

With the advent of 64-Bit chips from AMD and Intel on the horizon, clock cycle for clock cycle AMD and Intel hardware will have the lowend single proc and dual proc sun boxes out matched. as far as remote management goes see the price comparison post... a full lights out remote management card from dell is $399. That's all that is needed to be competitve. Beyond simply clock cycles the Intel architecture is about to undergo another evolution. the new bus architecutures that will support switched bandwidths greater than 2GB/sec will lead to massive disk and network throughput with the help of SATA and 10 Gigabit ENet.

the interchangeability of hardware is important from an IT perspective. having parts laying around, or being able to repurpose a boxe's task is important most netoworks are still mixed usally containing some form of windows - the linux servers that get decomissioned can still be used for other tasks that require windows. the same is not true for sun machines.

In the end I'm not saying that linux categorically eliminates solaris from the world, that is not the case at all. I'm trying to make the point that Sun is playing in a world in which they do not have the ability to compete, one that linux and intel have licked. There are other markets that will have a need for some time for the high end n-way boxes that sun produces but not the areas in which linux is being adopted. Linux is not the one stop shop answer to all the licensing woes of the personal computing and IT world; but sun's "brand" no longer carries the same singular sense of reliability and depend that it once did - becuase there are alternatives. Change is slow in the IT world - with the dot com boom gone *boom* - saving money is cool, and budgets are slim. these all point to Sun >rapidly< rethinking their current direction.

In two to three years when the existing dot com then "legacy" systems come up for upgrade - intel and amd hardware will be very attractive from a price/perfomance standpoint AND they will have a *nix esq OS that is available for deploying all of their Linux ported applications. The choice, will be clear.

Re:Sun, not dead, just a change of direction.

[Zebra_X]

I'll give you that Gigabit Ethernet isn't totally necessary. It was one of the weakest of the arguments. thought the chipset provided by dell allows for load balancing and failover in hardware.

hardware raid IS important, very important, these machines are typically @ the front lines of handling sometimes unpredictable loads. the reason FOR hardware raid is primarily to augment the processor in pushing bytes to the disk. sure linux and M$ both do software RAID to and it's no big deal. but for the price that dell offers it - there is no reason to offer it at well. in my mind reliability is supported by redundancy. ALSO you have to agree that a 1 year warranty is inexcuseable. Of the past sun machines that i've used two of the three broke. one needed a mother board replaced in the first few months after buying it and the other's ethernet interface died after being shipped somewhere (bizzare). My feeling is generally that a machines warranty should be 2/3rds of it's useful lifespan.

i'll find you benchmarks...

just becuase thousands of businesses use java (i do) doesn't mean that it's good. but that was a bit of an exageration.

My point with the 3 for every two was primarily that you get 3 servers that are more expandable for the price of two that are not... i'm not sure that they are all that competitve primarily because the v120 isn't really playing in the same league as the dell, basic things like second processor and the third drive bay and the redundant PS.

again, i think sun has a place - but it's not in the emerging linux market. and to my point last post we will find that companies are going to move away from sun hardware in the future. give it another 3 to 4 years. only time will tell - chalk one up to philosophy.

Re:Sun, not dead, just a change of direction.

[christophersaul]

You'll need to go back and do the config against the newly released v210, then factor in the cost of the software that comes for free with the v210. Should turn the tables nicely on the Dell box.

Re:Sun, not dead, just a change of direction.

[Zebra_X]

competitive, yes. although have a look at the warranty... to get a 3 yr wrty it's an extra 3K... what gives?

Re:Sun, not dead, just a change of direction.

[Zebra_X]

computers aren't really cars as far as warranty's go. it's hard to justify a 2 year difference in warranty period. a similar server to the dell from ibm has a 3yr. warranty also. Sun parts cost more too.

To be honest I beleived this for a while too, until I read an article about the 2.6 kernel. Premptive multitasking? Welcome to 1995. Separate queues for each block I/O device, isn't this common sense?

is this really needed for the applications that we are talking about? solaris IS more advanced than linux yes I haven't ever refuted that, but linux is more than adequite and costs less to own than solaris.

Re:Sun, not dead, just a change of direction.

[Zebra_X]

I can save money both on hardware costs and labor costs. I can also build better and more reliable, less expensive web applications using PHP instead of Java.

Perhaps in your world wasting time and money is OK. But not in mine.

Sun's Doom is like Apple's Doom...

[Bug-Y2K]

Content profitability in a comfortable niche market.

Too long have people labored under the delusion that one company MUST dominate the "computer market"... only because one company (Microsoft) seemingly *has* dominated. Just because one does today, doesn't mean that is the natural order of the market place. If anything it is unsustainable, as Microsoft is beginning to find out.

Sun makes some excellent high end gear, and in that market niche they are by far the largest player. They aren't even competeing with M$ in that space... and there is plenty of money to be made there. Sure CEO Jeff won't get to cross-check Bill in the teeth as often as he'd really like, but hey... that is NOT what Sun is in business to do.

Shake the current "one must dominate" worldview out of your heads /.'ers. It won't work. Microsoft's whole strategy, both internally and externally, is "For us to win, they have to lose." You WILL lose if you play that game with Microsoft because they play it better than anybody... but if you play a different game... Steve Jobs' game... where "we need to make something of quality that some percentage of the market wants and not worry about Microsoft" then you will do fine. There are billions of dollars to be had and significant percentages of market to be owned. Sure, you won't have dominance, but you don't really *need* it.

Re:Sun's Doom is like Apple's Doom...

[displaced80]

Parent makes a lot of sense. It's funny that geeks should expend so much effort arguing about markets, popularity and economics. I'm in it for the tech. I can't understand any so-called nerd writing off platform (x) for reasons of market share, dominance, etc. Everyone who thinks of themselves as a geek owes it to themselves to get out and *try* every option available: tinker with every OS, link against every library! Disassemble every case! Hell, you might find something you like more than what you've got now.

Re:Sun's Doom is like Apple's Doom...

[NineNine]

I agree 100%. Sun's problem is that they're trying to be everything to everybody. Im-fucking-possible. Even MS knows that they're not on the top end boxes. They keep trying, but they don't make a lot of headway. MS's business is the desktop and low-mid range servers. Sun's business is very high end servers. Apple's business is very high end desktops. No business on the planet caters to everybody. Not a single one. Sun should quit worrying about MS on the desktop and focus on what they're good at.

Re:Sun's Doom is like Apple's Doom...

[Fzz]

I agree. Not to mention that most software written for Linux can easily be ported to run on Solaris, and vice-versa. What this means is that if Sun finds a niche where they deliver something the customer wants and they can be profitable doing it, they won't die because of lack of available software.

Linux is Sun's friend; if all the low-end servers moved to Windows, Sun would have a much harder time finding such a stable profitable niche.

Re: Right on!

[knightinshiningarmor]

You are absolutely correct!

Intel, probably due to the fact they are off on their own world, seems to think SUN is doomed. We all know, however, that it is in fact planets which are doomed. Here's an example to help you out, termos.

And, don't forget:
1. Hold to knowledge that SUN is not doomed
2. ????????????
3. Profit!

Re:News Flash: Linux still not ready for the deskt

[obsidian head]

OT: Your current sig:

"Going to war without France is like going deer hunting without your accordion." - Norman Schwarzkopf

is contradicted by a well-known urban legends site.

The Economist didn't get it

[Alomex]

Traditionally, the Economist has been one of the earliest media outlets to get technology. The first printed press reference to the Internet outside the tech press is from the Economist, ditto for Linux, but this time around they are way off the mark.

The reason Linux is so popular is not that is free. BeOS is also free. Linux is successful due to convergence of many different factors:

(1) Free
(2) Open source
(3) Unix compatible
(4) X-windows (X11) compatible
(5) designed for x86 (yes it runs on many other chips, still Linux is an x86 project from the get go)
(6) Multiple vendor supported
(7) Plenty of third party support

Moreover each of these things feed of each other. That is why Linux is so popular.

Re:The Economist didn't get it

[Oliver Defacszio]

Oh, this makes no sense. Of your impressive list, only 1 and possibly 5 mean anything to anyone who doesn't hang out at Slashdot. Furthermore, how popular is linux anyway when less than 1% of worldwide non-Slashdot users have any interest in it at all? Additionally, things like X11 "support" are subjective positives, as there are a large number of people who think X is a dramatic weakness in comparison to Win2000 and OS X. Oh, and it's a pretty wild stretch to call basement hackers "third party support," not that it's ever stopped pundits before.

All this wishful thinking, while fun to watch, is little more than that.

Re:The Economist didn't get it

[Alomex]

Of your impressive list, only 1 and possibly 5 mean anything to anyone who doesn't hang out at Slashdot.

It's funny to see how much you don't get. In practice Linux is only used by sophisticated users (who else can figure it out?).

pretty wild stretch to call basement hackers "third party support,"

Have you ever heard of Red Hat?

Re:The Economist didn't get it

[Oliver Defacszio]

Someone doesn't get it, that's for sure. In your original post, you listed reasons why linux "is so popular." You didn't say anything about how it's only popular to fringe groups like this one, which make up a comically small piece of the overall pie.

So, indirectly, you have supported my point. The fact that it holds any value only to "sophisticated" (read: smug) users, when coupled with the insignificant percentage of the world that falls into such a category, means that its not popular. I will avoid a re-hashed commentary on how the points you listed polarize the entire "popularity" even further.

By your logic, if I am incredibly handsome to only blind women (probably true), should I then write a comment stating "why I am so handsome"? People would rightfully laugh, and so I laugh.

Re:The Economist didn't get it

[Alomex]

The fact that it holds any value only to "sophisticated" (read: smug) users,

No. This is where you are wrong.

A sophisticated user is the IT department of a large corporation.

Linuz is starting to be popular among those, hence the article from the Economist. They certainly don't care about smug slashdotters.

Re:The Economist didn't get it

[Oliver Defacszio]

Like many others, I have worked in IT departments of relatively large companies and, while it seems that linux is somewhat popular in the trenches, I assure you that nerds are not the ones who cut infrastructure development cheques. Those who do are not particularly impressed by the points you originally made, with the possible exceptions I have noted.

The suits are well aware that "third-party" vendors such as Red Hat are hanging by a thread and that a vast majority of linux development happens in someone's basement. I have personally seen the tide turn in board meetings simply because Sun, Microsoft and others look rock-solid economically in comparison. Despite Microsoft's EOLing of several common products, it still looks like a better deal to Joe CEO than the image of some geek being fed up and dropping support at a moment's notice. Sure, others can take up the task, but that requires time and leaves the entire thing in purgatory until it happens. On the other hand, until Microsoft says, "that's it, no more support on product X," it will be supported come hell or high water. Big difference.

Re:The Economist didn't get it

[Alomex]

Look, you keep coming back to slahsdot geeks. For the umpteenth time, this is not the people I have in mind.

You are the only one in this conversation equating sophisticated users with slashdotters.

I'm thinking about CIOs and CTOs who understand the technology. They are not religious about Open Source. They like because they can recompile the kernel if it turns out the TCP/IP stack does not support enough simultaneous connections. They care about X11 compatibility because they need to port existing applications...

Re:The Economist didn't get it

[Oliver Defacszio]

Well, I didn't think I had related anything to the Slashbots in that last comment, but you are free to interpret as you wish.

What I am saying is that I've never met anyone with corporate signing authority who cares about the points you listed (with, again, the exceptions). I have, however, met a few who are turned off by specific aspects of the linux "revolution" that are often the parts that techies like most. For example, distributed peer support. Even CTOs or CEOs who may ordinarily consider alternatives get darty-eyed at the idea of a midnight crash with something like Red Hat and news groups as the only lines of support.

Microsoft and Sun, despite the costs, are guaranteed to be there in a year. Red Hat isn't and I'll bet dollars to doughnuts that this means a lot more in most offices than an ability to recompile the kernel.

I am sure that there are companies specializing in linux support, but I can't name even one. I haven't looked, mind you, but the fact that nothing comes to mind means that they can't be doing a very good job of promoting their presence.

Ha! It didn't come from the article

[A nonymous Coward]

They said, in a fine example of shades of gray, that rather than make M$ a loser, Linux will make them less of a winner.

The fact that a slashdot poster can make such a bold (and stupid) statement is not surprising at all.

Not sure how the author reaches his conclusion...

[bc90021]

The author states "Only in Mr Mundie's nightmare scenario would Linux and other open-source software wipe Microsoft from the face of the earth. Mr Ellison's prediction might then come true, but with a drawback: his own firm, Oracle, would be wiped out too." I'm not quite sure how this could be true, since Oracle still sells database software that runs on Linux. It was a really good article right there until the end...

Nice article in "The Observer"

[ctid]

John Naughton, author of the excellent "Brief History of the Future" wrote an interesting article about the costs of Windows versus the cost of Linux on the desktop. He's definitely not an unbiased observer (no pun intended), but his article has real numbers in it.

Because the article submitter is full of BS

[0x0d0a]

economist.. i dont think MS is really a loser.

Yet when one reads the story instead of just the story post, one notices that the story post is a load of crap. The Economist said nothing of the sort -- as a matter of fact, it said mostly that Sun was in for it, and that Microsoft isn't in for that much trouble (since its dominance is in the workstation market).

From the article:
The most likely outcome is that customers will face a choice between Linux, which is cheap and cheerful, and Windows, which offers more bells and whistles, is tightly integrated with other Microsoft products and is easier for unskilled staff to use, but costs more.

Okay, I'll certainly grant integration and easier to use for unskilled staff, but where does the Economist get more "bells and whistles" from? When I think "bells and whistles", I generally don't think of Windows...

Re:Because the article submitter is full of BS

[TKinias]

scripsit 0x0d0a:

Okay, I'll certainly grant integration and easier to use for unskilled staff, but where does the Economist get more "bells and whistles" from? When I think "bells and whistles", I generally don't think of Windows...

That's interesting. I think you may understand the term ``bells and whistles'' differently than I, and apparently the author, do. Would it make more sense if he had written, in lieu of ``bells and whistles,'' ``cute and briefly entertaining but eventually tedious and ultimately useless misfeatures''?

Clippy is ``bells and whistles.''

Re:Because the article submitter is full of BS

[0x0d0a]

"Bells and whistles", the way I've always heard the phrase used, refers nonessential but nice features. Having the ability to run on a machine with defective RAM would fall into this category...

"Bells and whistles" don't have to be misfeatures...

Re:Because the article submitter is full of BS

[Corporate Troll]

Having the ability to run on a machine with defective RAM would fall into this category...

Why in hell would you want to do that? If it's defective, it should be replaced.

Anyways, for me "Bells and whistles" is the same as the parent poster thinks: eyecandy, integrated webbrowsers, cutesy helpers, etc...

Re:Because the article submitter is full of BS

[KiahZero]

OK, but if the machine has defective RAM and knows it, wouldn't it be nice to be told so by the machine, rather than having to trouble-shoot? As much as I love to try and solve mysteries, sometimes it's nice for the computer to give me a smack on the back of the head and let me know what the problem is.

Re:Because the article submitter is full of BS

[Corporate Troll]

Agreed! That would be really nifty. Is there any system that actually does that?
Anyways, the three main causes of flakyness in computer are:
a) Bad cooling
b) Flaky RAM
c) Weak power supply

I personally had a) and b), which makes your computer completely unpredictable. c) was reported to me on several instances by other geeks. Now, if a computer does weird things, it's this little list that I check.

Re:Because the article submitter is full of BS

[KiahZero]

Completely offtopic to the rest of the article, but oh well... Is there a way to test the power supply to see if it's giving enough power? My computer tends to spontaneously reboot, even though the CPU stays at 60 C most of the time, and I'm reasonably sure the RAM is still good.

Re:Because the article submitter is full of BS

[Corporate Troll]

A way? Probably by measuring the voltage fluctuation and stuff like that. As I said: I never experienced it myself, I always buy "overkill" power supplies.
I think the best way to check is just to try another powersupply and see if the problem goes away. Of course, that's a problem if you don't have a powersupply lying around (I have a 320W and a 500W powersupply just lying in my spare parts box...Both of them never used... long story.)

I don't care for Linux

[erf]

I've found the random, cobbled together nature of the Linux kernel to be irritating. I don't find it to be well-engineered as a whole, and development seems extremely haphazard. How many USB interfaces were there??

The BSD's are also free, run the same software, have rational development models, and while there isn't a profusion of distributions, the ones that exist are extremely easy to customize and very simple to maintain. Linux does have mindshare though...perhaps it's another example of the less-well-made product beating out the superior ones by virtue of marketing.

Re:Not sure how the author reaches his conclusion.

[ClosedSource]

I think any writer who takes predictions from Larry Ellison seriously has zero credibility. What happened to the network computers we were all supposed to be using the last 7 years?

Sun

[DrWhizBang]

Sun is probably doomed.

Oh no! I was told we had a few billion years left! I... what's that? Oh, that Sun.

Sorry for the panic.

Re:Sun

[kcb93x]

Yes, a lot of good things are written in Java, but you don't write an OS in Java. You don't write games that will actually sell in Java. And you really don't write something you to run fast in Java.

So maybe Java is good for introductory training before moving onto C++? That's what it seems like to me. But, IANACS (I Am Not A College Student)

Yet.

The Slant

[evocate]

Why is Mundie listed as "Chief Technical Officer" and Ballmer as "Chief Executive [Officer]", but Ellison is listed as "Boss"?

How is Sun the main loser if Linux hasn't replaced many system in the the "highest echelon"? Are telecom billing systems and airline reservation systems running on Windows? I doubt it. E-mail, web servers, file and print sharing. Applications for which Solaris is overkill. Hmm, that sounds like Windows territory to me.

Ownership cost is mentioned, and again Micosoft gets the spin due to lack of full description.

Mundie's collapse predictions are left completely unchallenged. Why does Ellison need a grain of salt, but Mundie does not?

Button-pushers like "cancer" and "nightmare" in the closing paragraph definitely set a tone for the reader's afterthoughts.

Trying to discredit Ellison's prediction by assuming the demise of Oracle on the basis of Mundie's questionable prediction is just wrong on many levels. There is more than one business model - I'll leave it at that.

Overall, this is less obvious than something that comes out of Microsoft-funded "independent analyses", but more it's more insidious too. Did Microsoft influence this author? Are Economist executives invested in Microsoft? Why the divide-and-conquer routine against Sun? Why start with such an optimistic view of Linux only to end on such a sour note?

Re:News Flash: Linux still not ready for the deskt

[the gnat]

Yeah, but the fact that you know how to dual-boot means you're already more technically astute than 99% of the people out there. I see your point, and feel the exact same way, but people here seem to forget that Linux is a good choice - even, perhaps, the ideal choice - for the desktop for many people here because of its power and customizability, which the n00bs don't care about. I've been using Linux exclusively for several years and I now find Windows downright difficult and obnoxious to use.

My problem is with stupid people who say "Linux is not a desktop OS" and assume that I can't possibly be using it that way, and thus expect me to fix their Windows boxes, send me Word docs, and don't understand why I don't love Windows.

Re:Not sure how the author reaches his conclusion.

[TKinias]

scripsit bc90021:

The author states "Only in Mr Mundie's nightmare scenario would Linux and other open-source software wipe Microsoft from the face of the earth. Mr Ellison's prediction might then come true, but with a drawback: his own firm, Oracle, would be wiped out too." I'm not quite sure how this could be true, since Oracle still sells database software that runs on Linux. It was a really good article right there until the end...

It actually does make sense, even if it's not laid out as clearly as it might be. Mundie says Linux will kill all ``commercial'' software. Ellison says Linux will kill MS. Economist says the only way Linux will kill MS is if, as Mundie argued, it kills all ``commercial'' software. In that case, Oracle is dead, too. Therefore, according to Economist, Ellison will be right only if Oracle dies too.

Clear? Like mud?

MS Development Tools ie Visual Studio

[jtotheh]

We use MS Visual Studio at work......it's a very nice development environment - really nice debugging and help features. There are Java tools that are comparable but I think they are actually more expensive (when you get the necessary top-end version) - MS ASP.NET as a development platform is attractive for medium-sized projects. I don't think there are any Free alternatives that are as nice.

Re:MS Development Tools ie Visual Studio

[the eric conspiracy]

Eclipse + Tomcat + JBoss is a pretty decent free combination that offers a lot of capability.

Re:Not sure how the author reaches his conclusion.

[skillet-thief]

To say it perhaps a bit more clearly -- like mud -- it seems the Economist is just saying that the rise of Linux and FOSS in general could lead customers to replace Oracle DBs with MySQL, PostgreSQL, etc.

It isn't entirely inconceivable, even if there is still a pretty big gap between Pg, Mysql and Oracle.

Correction Mr. Mundie

[bluGill]

Sorry Mr. Mundie, but Apache is NOT a clone of comercial software as you tried to claim. Apache is a fork (clone) of the Origional NCSA web server, which was NOT commercial. (open source, but I'm not sure exactly what license was used so it might not meet the exact legal definition of Open Source). IIS, and the other comercial servers are clones of an open source webserver.

Of course this is all an accidemic exercise, but don't try to claim some high ground where Apachee has it.

More Lies!

[Mohamm3d Al-Sahaf]

Sun is not dead, but they are burning in their cublicles.

Linux will fall too. The only OS to survive will be Windows (tm) due to it total world domination, Iraqi-style.

bah!

[rutledjw]

You speak as though Sun has a solid business model...

From what we've seen working with them, that's not the case. For reasons beyond my meager intellect, my company has sold it's soul to the IBM hardware cult. However, Sun has some good software and services offerings we'd like to use. Sun keeps coming back trying to open the hardware door again.

While I'd personally like to let them do it (particularly with their blade product) the option simply isn't there. Their lack of focus on what we want them for (services and some software) is going to cost them legitimate services (primarily) and software (secondarily) business at some point in the future.

IMHO, Sun's real failing is their inability to address software and services as legitimate revenue sources on their own. Instead they are solely used to move hardware.

So basically you're right. You just said it in one sentence...

Re:News Flash: Linux still not ready for the deskt

[antiMStroll]

The Economist is qualified to make statements about the perceived financial benefits of different operating systems, they are not qualified to make predictions on software development. The only insight I gleened from your post is that the majority of Slashdot moderators run Windows.

MS Information Minister

[Anonymous Coward]

Linux is nowhere near Redmond! They are not even within 100 miles. My feelings are the same -- we will embrace and extend all infidels!

Ahem! I forgot the link:

[ctid]

Naughton's article in "The Observer"

Re:Unix apartheid

[NetMasta10bt]

UNIX is the way it is because it was well thought out and was designed the way it is. The learning curve is quite steep compared to the trivial days of DOS.

If you are not up to the challenge of downloading a set of RedHat/whathaveyou CD's then just buy a computer with it preinstalled. If you want that whole click next to install thing, check out a Walmart PC with Lindows preinstalled for $199.

Most home users I know, still have never installed Windows themselves.

Apache a CLONE??

[cmacb]

"According to Craig Mundie, Microsoft's chief technical officer, as the open-source movement grows, it will get better at producing free clones of commercial software."

I give up. What is Apache a clone of? I wish they had included an actual quote on this. Maybe the journalist was "interpreting".

Most of the article is on target though. The easy way to evaluate the strengths of the companies mentioned is to look at how diversified they are (or aren't).

IBM is no longer primarily a hardware company. They have a strong consulting division, they do fundamental research and grab pattents on REAL things (rather than new parsing algorithms like some companies I can think of), they have a very strong software development component, they farm out hardware manufacturing that is no longer profitable (disk drives) while hanging on to things that they do best and can make money on (chip fabrication).

Sun is primarily a hardware company. Their operating systems are (almost) exclusively sold to customers who use their hardware. Java and Star Office are far from being cash cows. Their weakness is that as Intel, AMD, etc, chips get cheaper there is less incentive to use Sun's higher priced hardware components. Supporting Linux helps them a bit, but it is the price of their hardware that puts them at a disadvantage.

Microsoft is a software company. They are trying real hard to become something else too, but like Sun they are having a heck of a time making anything else work. They don't really do fundamental research, but instead try and grab patents on programming concepts so that they can bully other companies in court when it suits them. They don't really make any hardware, but instead stamp their logo on a few things to make it seem that they do. They do select good subcontractors for mice and keyboards, I'll grant them that. Everything they do except Windows and Office lose money. Prospects for either of those (because they are already so successful) can only go down. They currently have a scatter shot approach to the "next big thing" which consists of trying everything at once and seeing if any of it takes off. Few companies have the money to do this. But they will bleed themselves dry rather quickly if they are not carefull. Something tells me they are not going to be carefull.

Apple is trying to diversify too. Since they are starting small the only way they have to go is up. It would be nice to see them further popularize the power-pc server. My personal experience with OS X is that they are rushing versions of it out the door too fast. I've decided to wait for XI (or whatever they call it) and switched to running Linux on my iBook. Those gel buttons are cute though.

Re:Apache a CLONE??

[Anonymous Coward]

"The easy way to evaluate the strengths of the companies mentioned is to look at how diversified they are (or aren't). "

Strength of company sure...

Strength of product, no...

General Motors is one of the most diversified companies in the automotive industry. But one could hardly call any of their vehicles superior.

Oh wait nevermind... the rest of your article is nonsense as well. Microsoft not doing fundamental research? Huh?

Re:Apache a CLONE??

[MyHair]

I give up. What is Apache a clone of?

IIRC, Apache started off as a clone of NCSA httpd. I can't find the reference (I thought it was in the config file or man page, but no), but I think they wanted identical functionality and config files at first.

Of course it has grown into so much more now.

Re:Apache a CLONE??

[hpa]

IIRC, Apache started off as a clone of NCSA httpd. I can't find the reference (I thought it was in the config file or man page, but no), but I think they wanted identical functionality and config files at first.

Of course it has grown into so much more now.

Actually, Apache is a derivative (i.e. a fork) of NCSA httpd, which was also Open Source. A "clone" implies it is a re-implementation. The name "Apache" derives from being "a patchy server", which betrays its heritage as a collection of patches to NCSA httpd.

Re:Apache a CLONE??

[MyHair]

Actually, Apache is a derivative (i.e. a fork) of NCSA httpd

Looks like you're right. I couldn't remember if it was a clone, fork or if it shared some code. Now we know.

The name "Apache" derives from being "a patchy server"

That's what I've always read, but when finding the above link I ran accross this which claims it was named with respect to the Apache Nation and that the "A PAtCHy server" reference is incorrect, although popular. I get the feeling they're trying to rewrite history to be politically correct. Given how open source project names come about I find the common explanation more likely, but what do I know?

I suppose next they'll tell us the Gimp was named in honor of disabled people or the character in Pulp Fiction.

Re:News Flash: Linux still not ready for the deskt

[00_NOP]

I think the real reason that this is not happening is the one hinted at the article in terms of TCO (or at least the perception of TCO). There are a lot of low garde, low paid Windows sysadmins out there and they are a powerful conservative force in the offices of the world.

However, if I was running a new company I'd have nothing to do with Windows, it's too expensive, too inflexible. I'd just hire a better sysop.

Where I work the admin is very resistant to Linux (it frightens him to be honest, because he knows how to plug in a network patch board, but he doesn't really know how to configure TCP/IP) but I was bale to do things for 'free' (ie just my time) in Linux/apache that he said would cost multo.

I think I know why.

[twitter]

You wonder how Linux could help Sun. The article thinks it knows why Linux could hurt Sun:

Many people who would once have bought expensive Sun boxes running Solaris are now running Linux on cheap, PC-like machines instead.

What they missed is that free software has brought many more people into the world of Unix. Those people don't mind making the jump to Solaris when they feel the need for better hardware. Sun will continue to reap the benfits of their excellent hardware even when people buy that hardware and then load it with free software. Sun is far from doomed because Sun has many clues.

Microsoft, on the other hand, has not got a chance thanks to free software. They have been unable to penetrate the server market with their inferior wares because Linux is easier to use than all the point and drool tools. Microsoft can expect hardware makers to revolt soon and they will lose the desktop. Most Linux distros are already "good enough" for the desktop. With enough a little more hardware support, there will be few people who want Microsoft's wimpy, DRM crippled stuff. There's no way this user could ever go back and there's nothing special about me. The only thing Microsoft has to offer is a bullying monopoly. "You have to learn it because its there," people have told me. Without that, poof, nothing left to offer but stuff that's losing money and depends on monoply rents.

Re:News Flash: Linux still not ready for the deskt

[RoLi]

Linux is coming, but not to the desktop.

Translation:

We said for years that Linux will never, never, never come to servers (and were wrong), so we now say that Linux will never, never, never come to desktops.

Why is it that journalists (and some posters around here) completely lack the ability to acknowledge developments in the future? Why is it they are about 5 years behind things? (by that time it would have been correct to say that "Linux is coming", today, in 2003, it's already a major, if not dominating player in many computing fields.)

More Linux propaganda for the brainwashed

[Maxamoto]

Linux will be gone (thank God) in less than 5 years. Microsoft and the TCPA will make sure of that. 1. Microsoft is an extremely competitive company, bent on domination. The Linux community is still fighting over which is better; Gnome or KDE. 2. A correctly configured Microsoft OS is inherently more secure than any distro of Unix or Linux. 3. Microsoft OSs will work with every piece of x86 hardware out of box. Linux still requires endless hoop-jumping. 4. Microsoft is in bed with all hardware vendors. Linux is not. 5. Microsoft adheres to industry standards. Linux does not. 6. Microsoft provides tools to administer Linux and Unix clients. The only too I've seen Linux offer to administer Microsoft clients is rdesktop. 7. Microsoft dominates 96% of desktop OSs. Linux certainly does not. 8. TCPA will eliminate Linux from the desktop, making it impossible for coders to write server applications for Linux servers. 9. Microsoft is politically savvy. The Linux community is not. 10. Microsoft Operating Systems are simply better in every aspect that Linux Operating Systems. They've had 2 decades to perfect the desktop environments. Linux has not, and will not, thanks to the TCPA. I've seen the future. Make Mine Microsoft.

NEWS FLASH

[Erik Hollensbe]

NEWS FLASH:

LINUX IS COOLER THAN WINDOWS

I really wish someone would keep a tally of these articles. Jesus, 3 years ago this was starting to get old, now it's just stupid.

Did someone pay for that story or something? I'm sure Rob & Crew have heard this enough times, too.

Sun is probably doomed?

[Anonymous Coward]

What kind of conclusion is that? You made absolutely no support to that claim. Sun is a HARDWARE and software company. The could very change their business model to fully supporting Linux on their custom hardware. Businesses would jump at the chance to run Linux on proven server hardware and recieve support for hardware and software from one vendor (no more it's the hardware/no it's the software support bounce arounds). I think the rest of this makes sense but the Sun is probably doomed statement is just plain dumb and tacked on at the end.

Open source competition stifles innovation?

[mariox19]

Okay, here I quote the most deliberate piece of misinformation and propaganda:

"According to Craig Mundie, Microsoft's chief technical officer, as the open-source movement grows, it will get better at producing free clones of commercial software. [...] Such products reduce the incentive for commercial firms to innovate...."

In short, competition from quality products stifles innovation -- according to Microsoft. Commercial firms have no incentive to actually come up with products having more features, more stability, better documentation and ease of use than open source products because no one is willing to pay for such benefits?

Well, as they must be fond of saying over at Microsoft, "If you believe that, I've got an operating system to sell you."

I'm curious.

[Kaki Nix Sain]

My background: been using Linux for a good while now (current=gentoo), so I'm not a newbie from the config-it, use-it, script-it angle. However, I've never had anything but short, infrequent inclinations to dig down and learn about the mechanics of how the fundimentals of the OS work. So from that angle, I'm total newbie.

Now to the question. You say "...that's because Linux isn't ready for the high end (yet). By high end I'm talking about F15 and F12K servers." Which gives me a jolt of curiosity. What are some parts of the mechanics in Linux that prevent its move to high end systems? Are there any fundamental issues that will prevent Linux from 'growing up' and eventually running on those systems? Must some parts of the Linux mechanics be scrapped/overhauled? Or is it more a matter of time and will to add/extend pieces that will make the move possible?

Re:I'm curious.

[Biolo]

The issue is the granularity of the locks in the kernel, as well as some other scalability issues like how well it can actually manage 106 CPUs in a single OS instance, handling thousands of devices (disks, etc) and stability in doing so.

Perception is a major part of it, IT managers don't have case studies of major orgs using Linux on high-end hardware so don't trust it there, whereas Solaris is commonplace in that space.

Linux is definitely getting there, finer grained spinlocks help, the number of devices it can handle concurrently is increasing, I think in about 3 years it will be ready for the F15K's of this world.

Background - I work for Sun, but am a Linux advocate. I'd truly love to have Linux everywhere, I keep getting bitten on the arse by differences between Solaris and Linux (as well as HP-UX, AIX, etc) (anyone tried to use killall under AIX?!!). This said I also think that UltraSparc is a massively better CPU base than the x86, especially at the high end. No one has yet been able to put together a 6800 class or above system that gets taken seriously,which is why Intel is playing with Itanium to try and break into that space.

Re:Unix apartheid

[NetMasta10bt]

But the average user never gets to the learning curve part because they simply cannot figure out how to get it on their machines

Nono, you don't understand it's not broken, it is the way it is because it is extremely powerful. Thats like saying you've got a 1000 horsepower dragster, but YOU wanna drive it, so you want to have it only run on one cylinder.

The key is to have proper front end to all the intricacies that lie within... ie: MacOSX .

No I'm not your average end user, but I think the default desktop on RedHat is pretty well layed out. I even have my Mother running it at home.

Microsoft is not going to go anytime soon...

[sciwhiz007]

As much as some of us would like it to happen, Microsoft is not going to lose it's desktop share very soon. It possesses several advantages over its competitors, whether we accept it or not.

The most obvious reasons why Microsoft dominates are because it locks down consumers, it's UI (although simple, and in my opinion, primitive) is appealing to the average Joe, and because it uses a very simple point-and-click interface.

The average user only needs his/her PC to write documents, play a few games, check e-mail and surf the internet. He/she does not care about the OS wars, and doesn't want anything at all to do with the command line. He/she wants an interface that's familiar to him/her. If the interface is too powerful or configurable, (or too different from the standard Windows UI) he/she gets worried. He/she does not want to write config files, and does not want to reinstall the OS that came with his PC. He/she also wants to use what he/she uses in his workplace. He/she does not want to lose any formatting, or have to install new products, or run an emulator. These are some of the reasons why Microsoft dominates.

Lastly, Microsoft has a huge edge, i.e. money. It uses this money to aggressively advertise and get the message out to the common man. Most users have not HEARD of UNIX/Linux. If Linux wishes to dominate the market, it needs some serious advertising. Also, more hardware manufacturers must incorporate Linux into their PCs, and distro companies must obviously convince the main manufacturers. (This is already happening, and I am glad).

Although Linux certainly satisfies my needs, I don't think it's SIMPLE and EASY enough for the average Joe.

Re:News Flash: Linux still not ready for the deskt

[The Cydonian]

First of all, great new term, fuddites. Really, love the way folks here invent new words; I know for a fact that my lexicon is changing.

Second, about Linux-only desktops. I've been dual-booting myself for the last two years, and quite honestly, prefer Linux to Windows. But the fact remains that virtually all workshops/conferences these days demand presentations be done in ppt format only. (You can get away with pdf for documents). Yes, I have been looking at non-MS alternatives; I have OpenOffice on both Windows and Linux, I have tried OperaShow, Flash, Director and others as alternatives, but no, I still can't get away from ppt's hegemony. Fact of life, I guess, no use being religious about it.

Re:Unix apartheid

[NetMasta10bt]

It will defeat Windows if for nothing other than that it's free, as is it's office applications. That's really what it all comes down to.

Companies all over the world are starting to switch right now, and in turn so will there employees...

You'll soon see tv commercials from that guy hocking his video learning CD's, showing just how easy it is to switch.

Re:Sun is doomed

[Not The Real Me]

Next time...read the article before posting.

Small and medium sized companies do not need 104 processor systems running 8 OSes. In a Fortune 500 company, a department level server does not need 104 processors. 5 years ago, Sun sold a dual Sparc 336 system to the department I worked for. Price tag? Over $100,000.00

Sun's days of selling department level servers and servers to small and medium sized businesses are numbered.

Sun will still compete in the enterprise level but they will become more of a niche player.

Winer image vs Looser Image

[fjpereira]

I don't think MS will go out of business in the near times, but as any other empire, one day it will vanish in the air.
However, the first step before MS starts loosing marketshae is loosing mindshare and it's articles
like this that make people start thinking that MS isn't a safe bet.
Right now MS is the market leader and most people just follow the leader: many people making purchase decisions just buy MS products because it's the market leader and it has an invencible reputation, even if their products are inferior to the competition.
As a mather of fact, many people just buy MS products and don't even look for alternative solutions.
However, this kind of articles are the first wave of change. Change the way as people perceive MS and IT market. It changes the invencible perception about MS as more and more people start beleiving that the day when MS will desapear is geting each time more closer...

Basic economics..

[segfault_0]

Microsoft and Sun will both be big losers in my opinion, for reasons of simple economics. Commoditization of software has been going on for a long time. This is why you see Microsoft constantly screaming from the rooftops about their new features, which 90% of people dont care about. Its the only thing that makes them different from your Linux's and OpenOffice's.

There will come a point when it really just wont functionally matter which os or productivity suite you use. At that point in time, the cheapest alternative always wins. So unless Microsoft plans on paying us to use their software - time and open source software will inevitably have their way with them.

Re:Basic economics..

[Anonymous Coward]

Hey Batman, riddle me this: If commoditization has been going on for a long time then why didn't Sun die 10 or 15 years ago?

Have to agree with you.

[eniu!uine]

I switched from Windows to Linux a while back and there really isn't much I did with Windows I can't do with Linux. I can still browse the web, play media files, do word processing, and even play a few games(those that have Linux ports like Terminus). I'm not exactly technically inept, but I'm no sysadmin and I've been able to get it working. The install was just as simple as Windows was. For those that think Linux isn't ready for the desktop my question is: what are you using it for? Although the lack of third party software is a little annoying, that's hardly a problem with Linux itself. Most people that I know who run Windows never paid for it and the two who bought XP that I know can't use it now because they made too many hardware upgrades. If I had paid for Windows+Office+Photoshop etc. I would have been looking at easily the cost of another computer. This way I get the same stuff with some added benefits(like NEVER getting a virus) and I get to be honest about it. I used to reinstall Windows every few months, but I haven't reinstalled Linux yet. Even when it comes to the interface I think KDE beats the hell out of Windows. I don't know how I ever lived without multiple desktops. Yeah, there are things you can't do with Linux(things you won't be able to do until Linux hits mainstream desktop use), but I can see cases where Linux on desktops would definately be a more cost effective approach. Where I worked in telephone sales we had hundreds of computers all running the same application. In this situation why pay hundreds for NT? Why have Windows for library computers? The search application is probably web based anyway. People need to be pragmatic about this stuff, not just do what everyone else is doing.

Re:Have to agree with you.

[randallman]

I certainly agree with the library bit. I think open source OS linux/bsd/etc is perfect for deployment in these single use applications. Easy to configure and admin several terminals and STABLE, STABLE, STABLE. You left that part out. I've never crashed a linux box. The worst I've done is freeze X. Then I ssh from another machine and restart X. Never been forced to reboot a linux machine.

Re:Unix apartheid

[Oliver Defacszio]

It will defeat Windows if for nothing other than that it's free, as is it's office applications

Nice try. History has shown that people are happy to pay for something when there are free alternatives when the paid product is even marginally better or the free product requires more energy. Why else are there shopping mall tax return kiosks?

Linux is already free, but Windows 2000 and XP are still better products for a desktop PC. Windows: 98%, Linux: ~1%. Thus, you're wrong.

Not to cheap Sun stuff

[HuguesT]

The cheap Sun workstations are just PCs with Sparcs instead of Pentium, running Solaris. IDE disks, PCI bus, USB, you name it.

No difference, not even so much in price. You can run Linux on these boxes too if you want.

Re:News Flash: Linux still not ready for the deskt

[SN74S181]

I don't remember anybody saying Linux would never, never, never come to servers.

It's been used extensively in that role for over half a decade now in many organizations.

I think there's substance to the people who say there are Linux problems with the desktop, and it's hand-waving to say 'they said the same thing about servers...' It just isn't so.

Re:No Confusion: Sun is Doomed

[Moses Lawn]

Are you the "BSD is dying" guy?

Um, what's this Microsofa thing?

[WillASeattle]

Is it some kind of shell you run on top of the kernel? I mean, I recognize Linux and Unix, but I can't find anything called Windows - unless you mean X-Windows? Is it any good? Where can I get a download of it - any ideas?

Re:Unix apartheid

[fredrik70]

Hey, Linux wan't created for the noble cause of displacing MS Windows. It was created because Linus et. al. wanted a free Unix-like OS plus the fun of implementing it at the same time. *BSD pretty much got the same history, a bunch of geeks wanting to do the OS of their choice and have some fun along the way.
If people like and use, yeah great. And please make it easirer and more friendly to use.
But I'm not sure Linux can (or should) be squeezed into a please-all-be-all OS just because of the political goal of kicking MS' butt.

Here we go again...

[TheOrquithVagrant]

Here we go again with more confusion of OS and Hardware capability and other factual errors.

Linux handles more than 4 Gigs just fine, on 64-bit architectures. The x86 architecture on the other hand needs terrible cludges (PAE) to handle more than 4 Gigs. Solaris does even worse than Linux at this on the x86 architecture. That said, Solaris's VM in general is probably more roboust than the current mainline Linux VM.

As for Linux's SMP scaling, it scales fine up to 8 CPU's on Intel hardware. SGI claims that on their Altix architecture, it scales well up to 64 CPU's. Given that they are probably the only ones who have even tested Linux on a SSI computer of that scale, they are the only ones who have actual facts to say anything about its scaling. Would be nice to see some figures, though, for sure.

Solaris's "high performance, high reliability filesystems"? Oh you mean Veritas VXFS? That's an add-on product, not part of Solaris. A fairly expensive add-on product, at that. Veritas sells VXFS for Linux too, in case you wonder, but you rarely see it on Linux systems because the need is simply not as great.
Solaris's native filesystem UFS is a _joke_ for performance, and the journalling that is now available for it is a bolted-on kludge. Hence, any serious Solaris system is likely to have VXFS installed. The exception is when the machines are running applications that use raw partitions for their data, which is very common to see (Oracle, Tivoli Storage Manager, etc).

The XFS port for Linux is not a beta, and has not been for a long time. It is not yet part of the main kernel tree, but that doesn't make it in any way a "beta" or "untested". You would be surprised if you knew how many commercially sold NAS solutions run Linux with XFS under the hood. XFS is still an add-on product, but one that is in some ways more capable than VXFS, and on top of that, free.

Oh yea. And XFS supports ACLs on Linux and works well together with samba, which is part of the reason why its common in those NAS boxes.

Re:News Flash: Linux still not ready for the deskt

[Bob Uhl]

Well, after years of dual-booting with Windows, Linux is the now the only thing going on my desktop, and I've gotta tell ya, I'm doing just fine (better than ever, IMHO).

Exactly. The market for desktop machines is fairly stagnant, from what I've been hearing. People are switching from Windows to Linux and from Windows to Mac OS X; no-one I know of is switching from Linux to Windows, although some are switching from Linux to Mac OS X. Slowly but surely, Windows marketshare is decreasing. It's going to take awhile, but a combination of ever-improving Unix desktops and ever-worsening Microsoft policies will drive users to the Unix world.

The real fight will be between OS X and Linux, in a dozen years or so. OS X will either lose or become free software. Apple will become a full-time hardware manufacturer, creating some of the best and prettiest `lifestyle appliances.' And free software will keep on truckin'.

Disk mirroring out of the box...

[jotaeleemeese]

... that does not suck.

CDE (don't laugh, it is predictable and stable).

No need for kernel rebuilds.

Support. First rate support (I have not experienced Red Hat nor IBM, from the first I have reservations, from the second I believe they will be as good as ever).

Yeah, sure.

[jotaeleemeese]

Like if not having control of the desktop has eliminated Sun, IBM, Apple and HP.

And surely they will follow blindly whatever MS's TCPA stuff mandates for Windows users.

And like if there will not be computer manufacturers willing to do business with people that will not like to be controlled in their computer requierements by third parties.

And surely all the old hardware running today is going to disappear from the face of the earth as soon as MS and its friends make TCPA.

And most importantly, Linux is just code, free code. As long as there are people willing to experiment with hardware (no matter how protected) it will be possible to do Linux stuff.

Re:News Flash: Linux still not ready for the deskt

[Archie Steel]

Three words: Codeweaver's Crossover Office. Rejoice.

Sig

[TKinias]

scripsit Corporate Troll:

A nerd without Karma is like a pretty girl without breasts.

Um, so a nerd without karma is like a pretty guy? Well, that's about half right ;)

Re:Sig

[Corporate Troll]

Uhm...Not really, since due to the "without karma" and "without breasts" both the pretty girl and the nerd are specialized forms of the collections "pretty girls" respectively "nerds".

Ergo, you cannot go back to the less specialized form. You cannot say "a nerd is like a pretty girl (with out without breast)", nor can you say "a nerd without karma is like a pretty girl".
Changing the gender, doesn't really help, because never did I state that a "petty girl" was equivalent with "pretty guy", nor did I state the gender of the nerd in question (might have been a girl in the first place!)

So essentially, your deduction based on my statement is incorrect. Funny...but incorrect.

And if you mean that "a pretty girls without breasts" == "a pretty guy", you should check the part below the belly button. For the record: I know some girls with very tiny breasts that I would consider "damned hot", I just got recently converted to the D-size formats. (Not that's not "FORMAT D:")

Re:News Flash: Linux still not ready for the deskt

[jak163]

Installed Redhat 7 on a partition. Can't connect to my ISP. PPP daemon dies. Tried two different ppp programs, two different ISPs, Redhat tech support, two postings to Usenet. No I haven't tried a different modem. But the point is I'm not my parents but I still can't get onto the Internet with Linux. I think it has to get a little easier for the desktop to succeed there.