A 20-year IT consulting veteran and author of “The Unix Guide to Defenestration” offers the second in a series of four articles that examine the state of computing as he imagines it will be five years from now. Last week’s article was on Microsoft’s vision for its future software. This week, he looks at where the Unix and open source alternative is likely to go and, next week, at the hardware both groups can expect to have available to run their software. The last of the series will look at the impact these changes are likely to have on the IT industry itself.
Future Computing, Part II: Unix vs. the world
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
This just has to be the biggest load of drivel ever written. Its nothing but a BSD ad.
The author seems to be a UNIX person dead set on his views. Can’t change that, but I really doubt IBM would go for BSD rather than Linux for the simple reason that Linux has much more momentum behind it, industrywide. But then again, everyone is entitled to their own opinions.
I seem to recall that the network computer was going to be all the rage starting in 1997, then again in 1999. The centralized computing model the author describes (under the minomer “distributed”) is simply not attractive when processing power is as cheap and small as it is now and getting cheaper and smaller.
I don’t think that Linux is going anywhere, I think he misses a big point when he talks about future technologies. Just because a new technology arrives does not mean it surplants an old one – even if it is a better technology. The real force that changes technology is mindshare not merit. Linux is going to probably surplant proprietary UNIX not the other way around. Development of Linux will continue at a blistering rate, perhaps the author is not aware of the many improvements in the 2.6 kernel. I don’t see why the 2.8 kernel would not have all the features that Solaris does today, and perhaps by the 3.0 release it may surpas Solaris.
I do agree with the author that Solaris will be a contender, I just don’t feel that it will eclipse Linux. Linux has too many zealous followers to be dismissed off-hand. Linux will continue to displace many older UNIX boxes, but probably not not brand new ones. AIX, HP-UX, Solaris, IRIX, will have market share for a long time.
By the time Linux has a Lion’s share of the UNIX world, DragonFly BSD will have been released and FreeBSD will also contend or even surpass Linux. However, the mindshare of UNIX will be Linux, without a doubt. Still, Solaris will not fall as fast to Linux as the other UNICes and Linux is going to hit a saturation point and its encroachment on UNIX and adoption will slow down dramatically, Solaris and BSD will definately be in the game. But, they definately will not be the UNIX market leaders, that will most certainly be a place held for Linux. When Linux finishes its consumption of its UNIX bretheren, this is when the fight between UNIX and Windows will become more interesting… The reall battle between operating systems is fought by the number and quality of applications available for them and how those applications run on their respective platforms.
Nice ideas but the truth is Microsoft is not going to sit by and let linux take over. I won’t say it will never happen but it won’t happen anytime soon. Microsoft always has a knack at eventualy coming out with better, faster, smarter products. (Microsoft haters save your breath). Once linux get to the point where it peaks in design and core changes are fewer between kernel revisions. Then software companies will be more likely to develope software for the linux platform. Linux needs some basic standards for software for it to be sucessful.
Hey,
thanks for posting the article. I enjoyed it even though I think the author got some pretty ridiculous biases.
Would be great if we could get more articles like this posted
It’s a thoughful analysis with good intensions, but I don’t agree on a couple of specific points.
We’ve already gone through thin client push before, it didn’t go too far. I agree with the other reviewer that PCs are getting cheaper. I would look at the new small-form-factor PCs. PC’s are “democratizing”, if you will, as opposed to the big centralized servers/mainframes with terminals. Users want to store local data, and they are hesitant to go back to the terminal world. Terminal-like thin clients might have a market, but I don’t think it’s going to take over the PCs.
I agree that Solaris probably gains from others’ demise and it has strong and improving high-end offerings, but it itself continues to be pressured by linux at the low end.
I agree that outside the U.S. there is a governmental and general user push for Linux, maybe some of that is to replace Windows. The SCO lawsuit might make some users think twice about deploying linux, but some users continue to deploy linux. I wonder if some distros, including those in Asia, might try to de-SCO-ize linux code as the code in alleged violation are revealed. But I agree that there is a chance that it might be unresolved for years, as they tend to be. There may be some “cloud” over Linux right now, but I wouldn’t count it out for the very long term for the above reason.
I agree that some linux users are probably thinking about defecting to BSD, but there are limitations. Apple is the only BSD that has suffiecient multimedia and user-friendliness to be on the desktop (for example, audio and video API). And for the server, commercial tech support for non-Apple BSD is probably not as widely available as linux. Device drivers are limited for linux, but even more limited for non-Apple BSD. And BSD doesn’t have hard real-time. And they have generally less push for modern fancy devices as linux. I use BSD, but these are areas that need improvement, IMHO.
I’m just curious, when you say “And BSD doesn’t have hard real-time” are you implying that Linux does? I’m not trying to pick a fight or anything as I agree with much of what you say, but if you are in fact implying that, then you’re wrong there. None of the free UNIX-like OSs that I am aware of yet support hard real-time operations.
“I use BSD, but these are areas that need improvement, IMHO.”
Same here, my thoughts exactly. They are all doing terrific work though. I am finding DragonFly to be particularly interesting of late.
>>Microsoft always has a knack at eventualy coming out with better, faster, smarter products.
My impressions of Microsoft are that they cater to people who want a shallow learning curve and less configurability through a GUI interface. I believe the new buzzphrase is “seamless computing”. They know their market well. The proposition they they come out with “better, faster, smarter” is diifcult to independently quantify although Microsoft’s marketing does a great job at this.
BSD is a better UNIX than Linux. The only real think Linux has over BSD is hype, the Linux hype will continue to die as year after year it continues to fail to deliver and people realize it is a technically inferior operating system. There will always be an open source alternative per-se, but Linux is in desperate need of a redesign of several portions that won’t be possible under the “Linux” name. The good news is that several portions of Linux can be ported over.
“…the Linux hype will continue to die as year after year it continues to fail to deliver and people realize it is a technically inferior operating system.”
Think you got that backwards. It certainly seems that Linux appears is spreading fast to people, governments and corporations, and as it does, attracts more and more people contributing code to it. I don’t believe it’s going to be losing momentum anytime soon.
I don’t really see BSD being a major force because of licence issues. BSD is largely taking stuff out of BSD and putting it in something else (sometimes the entire box and dice).
I think IBM and other’s are collaborating on linux to produce a stable platform based on all the “best of” and standard technology that can be produced. The GPL means that they can trust each other to contribute because in the end none of the partners can benefit unfairly from proprietary source. Using BSD means that they can’t protect code from other commercial closed source players, since anyone can take BSD and make a commercial modified closed source version.
BSD is less suitable for collaboration from commercial companies, but more suitable for using wholesale as a basis for products. However then we’d be back to UNIX with all its licensing woes.
Which is probably the genius of the GPL; Joe Bloggs with $10 can be in a partnership with undustry heavyweights on an even footing as both benefit from their respective contributions without locking the other out of the market.
> I don’t really see BSD being a major force because of
> licence issues.
Eh? I would say the same about Linux…
> BSD is largely taking stuff out of BSD
Oh? I wasn’t aware that it was using its own code! Good grief!
> BSD is less suitable for collaboration from commercial
> companies,
And Linux is?
> Which is probably the genius of the GPL;
You mean RMS’s vision of the world?
Posts like this make me not want to read OSnews anymore…
I think that linux desktops will be very popular in near future outside USA. The original land of Unix, USA, will remain M$ addicted.
The price factor and anti-USA sentiments will push linux first to government desktops and then to business, schools and domestic users.
Nobody ever gets predictions right and this guy is going to uphold that tradition. Also, I think he is sweet on MS.
Maybe I wasn’t clear enough.
BSD based projects by commercial companies will most likely be closed source, proprietary licences. The article seemed to indicate that this was an advantage. However this is no different to IBM+Sequent+SCO and that went nowhere and didn’t lead to a standard cutting edge OS in any way. Apple has done it and the OS is wonderful and they are sharing some of their code, but not all.
The problem with BSD is corporations are relucant share their code with BSD licence. With GPL at least their “partners” won’t just steal their code for closed source systems. (well in theory)
With linux all competitors supposedly they benefit from everyones contributions equally.
Cooperation with BSD based projects would be inherently closed source between partners on an equal footing.
and yes I’m cynical too. I should have put “genius” in quotes. Everyone has been saying for years that OS’s will be commodotized. Money will be made with services if you believe that then you’ll beleive that big IT companies want a state of the art, cheap OS to leverage products from.
Maybe I shoulda posted on slashdot, dumb comments by the barrelful there. I should also avoid the word standard and linux anywhere as that is a really dumb idea.
I don’t believe Linux will be a “losing momentum”. I just believe “Linux” faces many technical and political barriers that cannot be overcome. It is my opinion that Linux as a desktop operating system is already slow and unmanageable with its functionality still lagging several years behind. I also see no sort or long-term solutions to some of the long-standing problems in sight.
I said the same thing years ago and since then nothing has changed. Years ago everyone said I was wrong, these days even Red Hat itself says they have a very long way to go before they consider Linux desktop ready. I don’t believe I have ever read a good article covering where Linux falls short and why. Some day I will write one.
“Linux is in desperate need of a redesign of several portions that won’t be possible under the “Linux” name”
Great, can you name those areas? Linux has been entirely re-written more then once in its lifetime, why would that make it any less “Linux?” I dunno if you’ve seen the latest benchmarks but Linux 2.6 appears to be owning the BSDs in every area mentionable, all of Linux major historical performance areas are vanished. I am not about to claim that BSD is dead and/or dying, but the Linux hype has only been increasing and so has its technical capabilities. FreeBSD 5 still isn’t finished yet, what technical lead? Linux 2.4 has better CPU scalability then FreeBSD currently does, let alone the impending release of 2.6 (remember, SGI recently sold a 256 CPU Linux box).
I’m not trolling against BSD, but be sane. NetBSD’s SMP project is still at Linux 2.2 performance levels and OpenBSD’s has only begun, FreeBSD is still trying to finish release 5 (although making great progress lately). Linux has tbe mindshare and the raw manpower behind it, with lots nad lots of very bright people tinkering on it – as does BSD, but the numbers clearly lean towards Linux.
Why do so many journalists assume the SCO suit has any merit? The author dismisses Linux saying that people would be scared away from it for “legal and techincal” reasons. That has yet to happen, why would the future be any different? The author also says that moving to BSD would merit gains in performance, right – how? None of the BSDs have SMP scalability as mature as Linux 2.4, let alone 2.6! I suspect he’s employed by Sun. 🙂 Interesting read, but there were times I wanted to scream at the monitor.
I hope that by 2008 the OS finally has become more or less irrelevant. After all, users don’t really use an OS, they use applications. An OS is basically just a program that allows the user to start other programs.
With the availability of excellent cross platform toolkits like Qt and WxWindows and the expected growth of non-MS platforms, you might expect developers to start looking at these toolkits.
After all, the most obvious choice to targeting multiple platforms is to use cross-platform tools. Therefore, I expect that cross platform toolkits become more and more mainstraim, which should make the OS irrelevant.
That’s also one of the reasons I think the OSS world might want to try and embrace (&extend) the Windows platform instead of fighting it. I think it is much more important that users start using OSS Applications then that they start using an open OS.
I am much more bothered by users sending me attachments in proprietary file formats then I am bothered by neighbours catching the next outlook virus, for example.
What’s more, it’s the applications that tie users to the MS platform more then anything else. If we ever want a chance to de-trone MS on the desktop, we have to attrack more developers INCLUDING those that write proprietary software.
Therefore, we have to come up with an attractive alternative to the MS development tools.
That’s why I am working on a port of Debian’s apt system to win32 ( http://lists.debian.org/debian-win32/ ). I think the combination of cross-platform tools, a good build system and a good cross-patform distribution system will attrack developers and with every cross platform developer gained, MS loses a tiny bit of power and control making the Windows platform just a tiny bit less relevant.
Tell that to a Windows user who was infected with Code-Red/Blaster/Nimda/Nachi virus. I know, run the appropriate updated virus-ware, it is after all the users fault the operating system has massive and repeated security holes… The operating system has to allocate resources, in clustered systems, work with multiple processesors, multiple memory pools, run in distributed systems, collective memory spaces, etc, etc, etc… of course the OS will become irrelevant, makes perfect sense to me…
BSD will never win out because of the license. Yes I said the license. No matter where BSD goes any corporation and even Linux can take that code and put it into their OS. It will be cannibalized by everyone and anyone who wants the code. Corporations will “steal” the code and market it better than anything free can. That is the difference between Linux and BSD. Corporations alter Linux but they have to give those alterations back. This alone will put Linux on top. People all over the world can contribute to Linux but rich and powerful corporations are contributing as well, and those contributions are for everyone to share, not just paying customers. Besides, the 2.6 kernel is mostly on par with BSD or even better than it in some areas. BSD just can’t compete with the support Linux has behind it. Most people and companies aren’t even swayed by the SCO nonsense.
I’m just curious, when you say “And BSD doesn’t have hard real-time” are you implying that Linux does?
Sorry, clarification: I meant that linux has hard real-time _variants_ like MontaVista, not that linux itself is hard real-time.