“The Open Source community’s answer is to ignore Microsoft’s incredible technological lead – because it is proprietary and not a standard – and instead focusses on their own cool thing and self-gratifying cool features, because they think they can do better (which is where the ignorance combined with ego comes in).” Join the discussion at Advogato.
The Consequences of Open Source Arrogance
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Childish my ass…..Microsoft = M$ PERIOD. 1) its a lot quicker to type than spelling out the whole word and 2) Any company that is a convicted monopolist, has billions upon billions of cash, and still thinks $900+ for an OS and office suite is a fair price, and is constantly finding new ways to charge business’ and consumers more money automatically gets the S replaced with a $
On a side note….with regards to what OSS software is as good or better than Windows counterpart….we cant forget MPlayer. MPlayer just destroys any other media player on any platform. Why is it the best? Cause it can play EVERYTHING…you name it, it can play it. Its also has a very small footprint and is very fast.
1) its a lot quicker to type than spelling out the whole word and
I hope you’re joking. Your argument would be a valid justifaction for calling Microsoft MS, not M$. M$ is still the childish/immature/idiotic/whatever way to refer to the company. Abbreviations have nothing to do with it. Seems like your understanding of logic is quite flawed…
I also don’t feel like getting into the technical details, and you don’t seem too interested in the truth anyway, but it is not illegal to have a monopoly, and as such MS (<–) has not been convicted of such. Maybe you should go back and re-read your Slashdot article.
“One more time, when you pay 0$, you get the amount of quality your money have purchased.”
Well more than half the SW I ever bought is from companies that are long gone. The rest might still be here in one form or another, but rapid changes can leave me feeling pretty abandoned sooner or later no matter how much I paid.
> No, that is not correct. But what IS correct is that all
> the BSD innovation occured when BSD was still a research
> project at UCB. So even this isn’t a very good example
> of OSS innovation.
Do not agree, just to note: SMPng, KSE, snapshot support for filesystem (not transactions), GEOM (which is abstraction layer for accessing to storage devices: i.e.: you have some very wierd device which can store data and anouther, different wierd device which can store data. If you have drivers for both (I mean hardware drivers), you can merge them via GEOM using appropriate “plugin” to make RAID0 array, for example, or make RAID array from NFS-served storage and external firewire device and USB flash drive, it doesn’t matter what is physical layer).
Here you can find detailed description:
http://people.freebsd.org/~phk/Geom/
Just wondering
M$ is and always will be M$. Thats the way it is, if you want to whine about it fine.
“I also don’t feel like getting into the technical details, and you don’t seem too interested in the truth anyway, but it is not illegal to have a monopoly, and as such MS (<–) has not been convicted of such”
Who says Im not interested in the truth? Im certainly not interested in it from you since you obviously havent followed the case very well. Federal Court ruled that Microsoft……excuse me…M$..ILLEGALLY maintained their monopoly. Since it was established that they BROKE the law with NO APPEAL, the only thing at issue with the recent appelette court was the remedies. And even those are being appealed by Virginia and Massachusetts. Shall we go on….lets talk about why the other states dropped out of the action. Oh ya M$ bought them off. This was widely publicized by several major affiliates recently. I believe the deal was like 7 to 9 million per state to drop out of the appeal of CKs decision.
Regardless, M$ WAS convicted of illegally maintaining their monopoly.
“Have you actually used the version of Netscape that was given to the Mozilla team and then used the current version of Mozilla? I would say quite a bit of innovation has occured.”
I’ve used it (mostly on Windows. But on FreeBSD some). I wasn’t all that impressed with it, and the Java support is buggy. I did file a bug report in Bugzilla, but I don’t think anything has been done about it yet. (Not a big deal. The Java bugs are more of an annoyance than anything else. Not serious.)
“See my comment on cloned software in my last post. All software is cloned nowdays; even Microsoft’s. It doesn’t have any bearing on the discussion so please get over it. It is a non-issue.”
It’s an issue when all you are doing is copying software and not making any improvements on it. There is still inovation in software. For example, Sun’s dynamically patchable kernel is pretty inovative.
“I think EMACS sucks and I never use it, however, considering the bloated mess that is VisualStudio, which even includes a kitchen sink, I don’t think bloat can be considered as a disqualifying factor.”
Well, granted this is opinion here, but I think EMACS sucks. And it goes completely against the UNIX philosophy. As far as VideoStudeo, yes, it sucks too. Give me VIM for my programming editor and I am happy.
“Perl doesn’t encourage bad programming, but encourages people to use it by being flexible. If you’re a C programmer you can code perl C-Style. If you prefer a different language, perl can acommodate you as well.”
Perl does encourage bad programming. Even experienced perl programmers will admit that they often have a hard time figuring out what their own code does when they have to revist it a year later. Perl allows too many shortcuts and stuff that are little used and give you no clue what they do just by looking at the code. With Perl it is far to easy to say “I don’t need to comment this because it’s obvious what it does.” But a year later, it isn’t so obvious anymore.
Compare that with say… Python… A language where you can say “I don’t need to comment this because it’s obvious what it does.” and 9 times out of 10, you are right. It is obvious what it does and it is easy to figure it out again when you revisit it a year later.
“Funny, it is exactly the same with all Microsoft products as well. So I plead again, can we please get off of this “clone” focus and discuss something that is relevant?”
I think it is very relevant because OSS software, and particularily the GPL, discourages inovation in software because people can’t capitalize on their ideas.
“Try checking out openmosix.sourceforge.net sometime.”
Sorry, MOSIX was developed by Hebrew University as an academic research project in supercomputing. It was another gift horse given to the OSS community.
“Do not agree, just to note: SMPng”
Lets remember that FreeBSD didn’t get decent SMP support until the BSD/OS code was merged in, which of course, was not developed in an OSS environment.
“I have no saw one recent (meaning less than a year old) benchmark that says Apache is slower than Zeus.”
Apache 2 fixed this problem to a certain extent, but not completely.
Apache 1.x was slow by nature because its single threaded programmming caused it to constantly fork. Forking a new process is time consuming.
Apache was hampered even more when running on Linux because Linux has a relatively lame process scheduler. Improvements were made in kernel 2.4, but it is still relatively lame.
“I beg to differ. The only reason why Solaris is better is because of the maturity and the SMP code.”
The Solaris kernel device driver mechanism is also superior. As is the fact that it can be dynamically patched on the fly. There is no Linux kernel where you can do that. And the SMP support is a pretty big thing on server platforms. Also, Solaris has a better process scheduler.
“I tried to learn PHP and ASP once, while PHP can accept ASP scripts, pure PHP is very different than ASP.”
Remember that you don’t have to write your ASP scripts in Visual Basic. That is just the default language. But you can also write them in Java, or any other language which you have support for. I believe IIS has J-Script suppor by default. You just have to specify in the ASP header that you are using J-Script instead of VB.
“Seriously, what these guys have done hacking in their spare time is astonishing.”
That doesn’t change the fact that it is light years behind MS Office. And to the average person, it isn’t very useful. For one, it can’t work with Word documents. For anyone who has to colaborate, that is essential. Like it or not, Word is an industry standard and any other word processor that wants to seriously compete needs to have Word support.
StarOffice’s word support, on the other hand, is nearly perfect. It even supports revision tracking.
“I think it is very relevant because OSS software, and particularily the GPL, discourages inovation in software because people can’t capitalize on their ideas.”
This is an excellent point from a developers point of view. Although Im very happy with the software OSS provides for me, Im really looking forward to the day .NET and Mono are the basis of most windows and linux apps. When that happens, it will be much much easier for ppl and companies to capitalize on their innovations by building great software that is easy to run on/port to any platform. Developers will, for the first time be able to have a nearly identical codebase for their project that will run on windows, linux or unix.
Although its way early in the game, here is an early example of whats in store for developers with (.NET, Mono, C#):
http://www.desktoplinux.com/news/NS2017249197.html
> > Do not agree, just to note: SMPng
> Lets remember that FreeBSD didn’t get decent SMP support
> until the BSD/OS code was merged in, which of course, was
> not developed in an OSS environment.
SMPng is rewritten implementation of SMP obtained from BSD/OS 5.0.
Please refer to:
http://www.freebsd.org/smp/
Do not agree, just to note: SMPng, KSE, snapshot support for filesystem (not transactions), GEOM (which is abstraction layer for accessing to storage devices
None of those are innovative in the sense that they are something new and original.
SMPng: fine grained locking, woo. Just like virtually every other modern operating system.
KSE: woo, kernel threads. About time.
Sanpshot support: woo, something every other LVM toting OS supports.
GEOM: woo, a proper block layer. /me stares at his Linux based HDD-NFS RAID1 setup.
I’m sure you’re stoked that FreeBSD is crawling out of the precambian, but other operating systems proprietary and otherwise have already been there and done that.
>> “Ignorance with ego”
Thru the years I learned some things about human nature. One is that Murphy´s laws apply: the expert ones are those who look less savvy. A corollary is that calling others ignorant reveals ignorance of the accuser.
About ego — well, frankly, my dear… Do you expect us Linux users to get docile and pay Microsoft without questioning? Or maybe the idea is that hipocrisy game of “tolerated, if not promoted, piracy” of MS products?
Nein, danke. Even when free software solutions are worse than their MS equivalents, copy proprietary products annoy my soul — so Microsoft will be “no game” until the day they offer free products for Linux (like Internet Explorer).
Even them they´ll have a hard time, as many programs (Mozilla, Galeon, Opera, Konqueror, among others) already kick IE´s lame ass in lightness, speed, security, features, and overall looks… 😉
Microsoft has lead in the same sense gasoline has lead.
Free or open source software is unleaded.
And, besides, using a MS product will make me look idiotic. Even more!
Obvious flame bait, but I am going to bite it anyway…
“I’m sure you’re stoked that FreeBSD is crawling out of the precambian, but other operating systems proprietary and otherwise have already been there and done that.”
Oh really? Is that why FreeBSD benchmarks up to 30% faster than Linux at serving up HTTP requests? Is that why FreeBSD’s process scheduler blows Linux’s scheduler away? Is that why FreeBSD manages memory better than Linux. And perhaps most telling of all, is that why Warner Brothers chose FreeBSD to render special effects for The Matrix, giving the reason that the Linux software they were using performed better under FreeBSD’s Linux ABI layer than it did under native Linux? Come on…. When FreeBSD can run Linux software better than Linux itself can run it, there is a problem here.
“Oh really? Is that why FreeBSD benchmarks up to 30% faster than Linux at serving up HTTP requests? Is that why FreeBSD’s process scheduler blows Linux’s scheduler away? Is that why FreeBSD manages memory better than Linux. And perhaps most telling of all, is that why Warner Brothers chose FreeBSD to render special effects for The Matrix, giving the reason that the Linux software they were using performed better under FreeBSD’s Linux ABI layer than it did under native Linux? Come on…. When FreeBSD can run Linux software better than Linux itself can run it, there is a problem here”
Nearly all of the above issues will be a much different story next year when the 2.6 kernel comes out…rememeber Linux is still the new kid on the block. Despite that, it is rapidly replacing Unix in the market. As far as the Matrix rendering, yawn. Linux has been and will be used to render several movies. Off the top of my head I can think of Titanic and the new StarTrek: Nemesis movie that just came out. Im sure there are several others too.
Dont get me wrong, Ive got nothing against FreeBSD, but Linux is the future.
Is that why FreeBSD benchmarks up to 30% faster than Linux at serving up HTTP requests?
It does? I’m having trouble imaganing under what load, given that it’s bio and buffer subsystems are fairly poor and it completely lacks a multithreaded network stack.
Is that why FreeBSD’s process scheduler blows Linux’s scheduler away?
I really don’t think it does. Maybe you can provide us with some relevant benchmarks? FreeBSD’s processor scheduler is very simplistic, below the rather dumb stock Linux 2.4 one in many ways. Most certainly a couple of orders of magnitude below Ingos’ rather nice O(1) scheduler which most distros are shipping with now. FreeBSD 4 even lacks kernel threads for goodness sake.
Is that why FreeBSD manages memory better than Linux.
FreeBSD’s memory management is rather good in low end situations, I’ll give you that. Anything out of the ordinary and it’s a different story… remember that Linux 2.4 isn’t an unstable, newly introduced kernel anymore. It’s stable, tested and performant.
Furthermore, the VM in 2.6, like so much in 2.6, is simply top of the class.
And perhaps most telling of all, is that why Warner Brothers chose FreeBSD to render special effects for The Matrix, giving the reason that the Linux software they were using performed better under FreeBSD’s Linux ABI layer than it did under native Linux?
Sure, in 1999! Let me repeat: 1999! Since then we’ve seen Linux clusters adopted by practically every major movie house and Linux clusters appearing in the T500 list.
“Nearly all of the above issues will be a much different story next year when the 2.6 kernel comes out…”
This is called vaporware. I don’t care what kernel 2.6 WILL do. What can Linux do for me now?
“It does? I’m having trouble imaganing under what load, given that it’s bio and buffer subsystems are fairly poor and it completely lacks a multithreaded network stack.”
Under extremely high loads. That’s where Linux becomes unstable and where its lame process scheduler starts to bog down.
“FreeBSD’s memory management is rather good in low end situations, I’ll give you that. Anything out of the ordinary and it’s a different story… remember that Linux 2.4 isn’t an unstable, newly introduced kernel anymore. It’s stable, tested and performant.”
I’ve seen FreeBSD servers handle loads that cause 2.4 Linux to crash and burn. Linux is still unstable with very high loads.
Under extremely high loads. That’s where Linux becomes unstable and where its lame process scheduler starts to bog down.
and
I’ve seen FreeBSD servers handle loads that cause 2.4 Linux to crash and burn. Linux is still unstable with very high loads.
Nice. I’ve seen FreeBSD fall over in a catastrophic deadlock when a process simply malloced too much memory. I’ve attempted to port sophisticated multithreaded apps to FreeBSD only to be thwarted by the incredibly piss poor thread support.
For every anecdote you have, I can provide a counter. I had, obviously mistakenly, assumed that you had rigorous benchmarks to unerpin your claims.
“Nice. I’ve seen FreeBSD fall over in a catastrophic deadlock when a process simply malloced too much memory. I’ve attempted to port sophisticated multithreaded apps to FreeBSD only to be thwarted by the incredibly piss poor thread support.”
Granted FreeBSD’s thread support sucked until 5.0.
But I wonder if you know how to properly tune FreeBSD? Admitidly, FreeBSD sucks out of the box. Even SysAdmin didn’t know how to tune FreeBSD, which is why they ran such a horrible article. SysAdmin didn’t even know enough to enable soft updates on the file system.
As far as things being unsubstantiated, I didn’t see you back up your claim that FreeBSD is just now coming out of the cambrian era.
Granted FreeBSD’s thread support sucked until 5.0.
The hypocrisy is strong in this one!
This is called vaporware. I don’t care what kernel 2.6 WILL do. What can Linux do for me now?
So, we can’t talk about the features in 2.6, but you can talk about the features in 5.0
@_@
I didn’t see you back up your claim that FreeBSD is just now coming out of the cambrian era.
I think I did; all of those features are already present in other operating systems, including Linux, and have been for ages. Only a complete cretin would claim that FreeBSD 4.x – the current production version – isn’t lacking in all manner of modern features.
Fine. Accepts heavier loads than Linux? Great!
While I would like to see some numbers — *when the need for such loads come up* — I have no hidden agenda preventing me to recommend FreeBSD to customers.
A few moments ago I searched for “pchealth” in Google and IE executed an illegal operation (I´m at some relatives´ house now). This is totally unexplainable in straightforward terms. It´s like that butterfly moving its wings and causing a storm thousands of kilometers away.
This is a typical example of “technological lead”.
I’ve used it (mostly on Windows. But on FreeBSD some). I wasn’t all that impressed with it…
Just curious. What’s not to be impressed with. It is more secure than IE, It supports tabbed browsing. It has a decent spam filter. And so on and so on.
It’s an issue when all you are doing is copying software and not making any improvements on it.
Such as what. For example, I think ogg vorbis in a huge improvement over MP3. I think that OpenSSH in a huge improvement over its closed source counterpart. In my opinion Zope and OpenACS are an improvement over any of the content management systems I’ve used under Windows. And the list can go on and on.
Perl does encourage bad programming.
Perl is an inanimate and intangible thing. It cannot sway developers one way or another regarding programming skills. I think this statement is as funny as Sun claiming Java encouraged good programming. Well guess what. There is a hell of a lot more to programming than syntax, and Java doesn’t help a lousy programmer become a good program regardless of what Sun says. The same is true in the opposite direction with Perl.
If you can’t tell what your software does a year later, then your a lazy programmer who doesn’t comment code.
Perl allows too many shortcuts and stuff that are little used and give you no clue what they do just by looking at the code.
That’s what comments are for. If you don’t use them that’s your fault.
Compare that with say… Python…
Python is a very cool language.
I think it is very relevant because OSS software, and particularily the GPL, discourages inovation in software because people can’t capitalize on their ideas.
Ogg Vorbis, Zope…but I repeat myself…again.
So, we can’t talk about the features in 2.6, but you can talk about the features in 5.0
FreeBSD 5.0 does exist for now
5.0-RC2 released. And 5.0-RELEASE will be ~January 10th, 2003
Surprised? http://www.freebsd.org/releng/ will help you
“This is called vaporware. I don’t care what kernel 2.6 WILL do. What can Linux do for me now?”
“FreeBSD 5.0 does exist for now
5.0-RC2 released. And 5.0-RELEASE will be ~January 10th, 2003″
Vaporware my ass, its no more vaporware than BSD 5. Kernel 2.6 is available right now too (under the 2.5.xx development tag) just as BSD 5 RC2 is. http://www.kernel.org To call somthing vaporware that has been in development for well over a year, is in feature lock, readily available, and close to an official release doesnt make any sense. Regardless…..2.4.xx kernels r0><0rs
Its pretty sad that you dont care what any future enhancements may bring, but if you must be that way then fine. As far as “What can linux do for me now”….whatever you want. I was simply making a point that 2.6 will do all that the 2.4 kernels can do only much much faster with a lot more features and improvements.
The fact is that discussions (nearly) useless becausa TechLead or not, M$ has the money… stop.
Maybe OpenSource is copying in many ways from M$… The main reason is only to fill the Gap between universal standards existing through many OSes and M$ own standards trying to destroy OS intercomunication (and competion). Extremely obvious. M$ is a company and it will try all can to eliminate competition with all the weapons it its possess (many!).
On the other side M$ with all the minds hired in the last 20 years has copied (implemented? integrated) more from other platforms/OSes than all other together copied from them. From multitasking/threading (existing on game machines like Amiga 10 years before) to Active directories (BeOS?), and so on, and still didn’t catch up in some features like datatyping, localisation, simple and unified interapplication scripting for ALL applications on the platform, and, and and….
Just curious. What’s not to be impressed with. It is more secure than IE, It supports tabbed browsing. It has a decent spam filter. And so on and so on.
IE has the most technological advanced SPAM server integrated in it 🙂
“So, we can’t talk about the features in 2.6, but you can talk about the features in 5.0”
5.0 is due out in less than a month and the RC2 is already available. When is kernel 2.6 due out? Not for a hell of a long time yet.
“I think I did; all of those features are already present in other operating systems, including Linux, and have been for ages. Only a complete cretin would claim that FreeBSD 4.x – the current production version – isn’t lacking in all manner of modern features.”
As if Linux versions prior to 2.4 weren’t? I mean come on. Linux didn’t even have a decent filesystem. ext2fs was crap. And it was dangerous on servers with high transaction load.
“Just curious. What’s not to be impressed with. It is more secure than IE, It supports tabbed browsing. It has a decent spam filter. And so on and so on.”
It’s Java support is buggy, it doesn’t render complex pages correctly, and its still fairly crash prone on Windows.
“Perl is an inanimate and intangible thing. It cannot sway developers one way or another regarding programming skills.”
Even Eric Raymond, a well known Perl advocate, admitted after trying Python that Perl encourages bad programming, and that someomes he couldn’t decypher his own Perl code when he revisted it a year later.
And god forbid you would have to maintain an application written in Perl by another programmer.
“There is a hell of a lot more to programming than syntax, and Java doesn’t help a lousy programmer become a good program regardless of what Sun says.”
Of course there is. But syntax determines to a great extent how maintainable your code is, especially by others.
“I’ve seen FreeBSD fall over in a catastrophic deadlock when a process simply malloced too much memory. I’ve attempted to port sophisticated multithreaded apps to FreeBSD only to be thwarted by the incredibly piss poor thread support.”
BTW, I’ve seen Linux lock up tight because of a buggy paging system. I’ve seen it do this under even virtually no load.
As if Linux versions prior to 2.4 weren’t? I mean come on. Linux didn’t even have a decent filesystem. ext2fs was crap. And it was dangerous on servers with high transaction load.
2.4 has been in production for two years now. Comparing a current production release of one OS to a previous generation production release of another OS just smacks of desperation.
FreeBSD is only now catching up to the features Linux has had for two years or more. Furthermore, 5 is not production ready, and is unlikely to be for a while yet (yes, I’ve got four machines running it in the lab). Read the FreeBSD 5.0 EAG.
BTW, I’ve seen Linux lock up tight because of a buggy paging system. I’ve seen it do this under even virtually no load.
Nobody cares about your anecdotes.
Why are all of you arguing? The very short article that spawned this thread made one point, a point that could be debated.
We know that Microsoft is like a shark, never sleeping or resting, always on the hunt. That is not meant as a bash – when you’re number one, you have to do that.
If there is arrogance in the OSS movement, it is the type that must be there. If you are programming or working for OSS in some other capacity, you have to believe that it at least has the potential to be the best. If you did not believe that, then there would be no point in being involved – it would be an act of futility. When participating in something like OSS, one has to have a psychological impetus to keep going, to keep moving ahead.
If a person looks at Microsoft and its Roman like armies or programmers and business tactics, and you want to change this situation, you have to get yourself into a certain state of mind – you have to be a believer. If that seems like arrogance, then so be it. It is better than giving up and surrendering. Diversity in computing is essential and must be there. For OSS to keep moving ahead, those who work in it and support it have to have some ‘tude.
“FreeBSD is only now catching up to the features Linux has had for two years or more. Furthermore, 5 is not production ready, and is unlikely to be for a while yet (yes, I’ve got four machines running it in the lab). Read the FreeBSD 5.0 EAG.”
Shows how much you know about FreeBSD. RC-2 is available right now, and the official release will be available on January 17th.
And the reverse is also true. Linux is just now catching up to the stability and performance of FreeBSD for server use. That game can be played both ways.
“BTW, I’ve seen Linux lock up tight because of a buggy paging system. I’ve seen it do this under even virtually no load.
Nobody cares about your anecdotes.”
Nobody care’s about yours either. So your malloc example on FreeBSD is no better than my example about how Linux can’t even manage a swap file correctly.
Shows how much you know about FreeBSD. RC-2 is available right now, and the official release will be available on January 17th.
Read the FreeBSD 5.0 EAG
It basically says, “5.0 is not going to be as stable or even performant as 4.x. For situations where this matters, we recommend you stick with 4.x until 5.2 or so”.
Show how much you know about FreeBSD. Zealotry without the facts – gotta love it.
Linux is just now catching up to the stability and performance of FreeBSD for server use. That game can be played both ways.
Says you. Anecdotal rubbish once again…
Nobody care’s about yours either. So your malloc example on FreeBSD is no better than my example about how Linux can’t even manage a swap file correctly.
Wow, give the gimp a cookie.
This is exactly why I said:
For every anecdote you have, I can provide a counter. I had, obviously mistakenly, assumed that you had rigorous benchmarks to unerpin your claims.
Fancy that! I wasn’t trying to pass my anecdotes off as fact. You were.
“””Well, granted this is opinion here, but I think EMACS sucks. And it goes completely against the UNIX philosophy.”””
How is that? It’s composed of a bunch of little reusable lisp utilities that get autoloaded when needed.
The article mentions the critical technologies that Microsoft has and that are greatly innovative. There was a company once that produced an OS (BeOS) that STILL beets Windows XP performance when it comes to handling media. If you don’t believe me, download the personal edition and try it (www.bebits.com). It was the first personal computing OS to have a 64-bit journaled FS live incredible live queries, an integrated translation kit, and so on. MS does not match many of its features, and the API is OUTSTANDINGLY easy to use and incredibly powerful. Granted, it does have problems (networking…). My point is that this company (40 employees, at the largest), created revolutionary software. After having programmed under the Win32 API, having used MS technologies for the last 8 years, and having watched their technical progress, I have to say that it is competition like the open-source community (and little companies that deliver superb products) that is the reason for any of MS’s technological advances. If it were not for this, I highly doubt Windows would be what it is today. So, be thankful there is an “arrogant” open-source community that a) doesn’t agree with the greedy and unjust business practices that MS uses and b) that has enough of a brain in them to know that most MS technologies so far have been way below what they could and should have been.
“It basically says, “5.0 is not going to be as stable or even performant as 4.x. For situations where this matters, we recommend you stick with 4.x until 5.2 or so””
They always say that. So what is your point? They never recommened that people upgrade right away with a major version jump.
“Says you. Anecdotal rubbish once again…”
Is that why virtually all of the longest uptime servers are a BSD variant? And not a single Linux server made the list?
And you didn’t provide any evidence for your claims either. You are the one that first made the claim that FreeBSD is just now getting out the stoneage, and you chose a couple of examples and failed to even mention that there are many areas where FreeBSD is way ahead of Linux.
And as far as mulithreading on FreeBSD, I love your logic… Let’s blame the OS for the fact that you don’t know how to write a multithreaded app for FreeBSD. Makes a lot of sense.
I run many multithreaded apps on FreeBSD with no problems at all. I have written multithreaded apps for FreeBSD that run just fine.
“How is that? It’s composed of a bunch of little reusable lisp utilities that get autoloaded when needed.”
Autoloaded when needed… heh… Is that why the Emacs binary uses so much memory?
Besides, why would I want an editor/file manager/calendar/email client/web browser/whatever the hell else Emacs can do? To me, Emacs is a a good example of “jack of all trades, master of none.” It does a lot of things, but it doesn’t do any of them very well.
The basis for writing the article will become clear later next year.
In the mean-time, please consider this: that when ordinary business users need to use a computer, they automatically think (apart from “unreliable”) “Microsoft”.
Therefore, to break the domination of Microsoft, the answer is to BECOME Microsoft. After you have a fully microsoft compliant suite of tools and software, you can start to enhance and improve, and any attempt to “do better” will simply widen the gap between Open Source and Microsoft, not bridge it.
The longer Open Source leaves DCE/RPC and Microsoft DCE/RPC-based applications to Microsoft’s domination, the longer Microsoft has to use them as a strategic base to FURTHER widen the gap.
and the wider the gap gets, the less chance that the ordinary business user will be able to really choose between microsoft and open source systems without taking a significant risk, and you know exactly which way ordinary business users will swing when presented with risk and unknown.
> “The two [OS/2 and NT] share some history and some minor > bits of code, but the two diverged *significantly* well > before the first IBM release of OS/2 (v2.0) in 1992…”
>
> Their development has certainly diverged in many
> respects, but they share, for example, a filesystem
> (called “NTFS” for New Technology File System by MS
> and “HPFS” for High Performance File System by IBM).
>
> I’ll leave it to you whether you consider a common
> filesystem to be a “minor bit of code.”
While NTFS and HPFS share some general implementation concepts, the two are not at all compatible in physical structure, and calling them different implementations of a “common filesystem” is a bit of a stretch.
Besides, Linux can read/write HPFS partitions, and yet it and OS/2 are totally dissimilar in almost all other respects.
In addition, OS/2 2.x/3.x/4.x and WinNT have *completely* different kernels, the GUI API’s are different (Win32 vs. PM), OS/2 has the SOM-based WorkPlace Shell as well as REXX integration for both GUI and shell while Windows NT has a silly brain-dead shell, OS/2 has WinOS2 and MVDM subsystems which are totally unlike anything on NT or its derivatives, and OS/2 uses a completely different device driver model.
You might as well claim that Linux, Windows XP, OS/2, and BeOS are the same because they all support the FAT16 filesystem…
I agree that the DCE/RPC issue is very important. I’ve just kicked out Windows 2000 server in favour of Linux and Samba and it’s working great for me (with Windows 2000 clients). It’s been liberating in many respects, and there are certainly things I can do which I couldn’t before. But it does have its limitations and I can imagine the issues if mine was a larger enterprise to manage. For example, I’d like to see good tools to make LDAP much easier to manage and implement as an basis for authentication and management for some of the great available open-source software (qmail, samba etc), and in a way where it can be an AD server too, including system policy support for 2000/XP. The only tools I know of that are close to fulfilling that are Samba-TNG (which I haven’t yet tried) and Gosa which is promising, but at present, imperfect.
I know it’s really easy to say it, and a darn site harder for an open-source development team to deliver it – but that would be a hell of boost for *nix to eat into the Windows server market share, and it would be a good standpoint for further growth.
And as for *nix on the desktop, I’d really like to see a gtk2 port of a leaner OpenOffice and this idea of having a groupware app hinging on the XML/RPC backend of phpGroupWare sounds very interesting too (imagine a plugin for Evolution). I’d also like to see the gtk2 toolkit get a bit more *responsive* (Qt and Win32 GDI are still ahead there IMHO), although I’m definitely in the Gnome camp and I think RedHat are right to push it.