A CA World panel of Linux luminaries that included Linus Torvalds predicted the demise of Unix over time, saying it will be replaced by Linux. Again, our recent editorial on the subject.
Is Unix on Borrowed Time?
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
and i’ll sell a gazillion records, make some films and win a bunch of oscars… design a 256-bit RISC machine … and the list goes on…
This just in! Shocking news! A group of entirely nonbias Linux zealots predict success for their life devotion.
Linux will only replace UNIX when it does a better job
and can do the same or less performance wise.
That means we are still left with a qaulity product
The problem is that Linux is an immature and lowly engineered operating system that has few of the architectural sophistication and innovation of other UNIX varients. It embodies the past (old world monolithic approach) rather than the future (federations of objects, distributed O/S). Linux is also very popular with the low-brow market (the hobbists, the hackers and community). It worries me that pandering to Linux could cause slow down of innovation or work on more important issues. When Linux started to veer into the embedded OS market (competiting against the likes of VRTX, QNX, etc), I was in fact quite worried. These embedded OS are highly engineered for what they do, yet putting Linux into this environment is just as bad as putting Windows onto a mobile phone — the OS wasn’t designed for this type of environment. We use Solaris for our mission critical application servers, and you look at the sophistication in the performance engineering of Solaris (e.g. the ability to fine tune and monitor the OS) compared to Linux, and Linux is a mere teenager. It really disturbs me whtn “Linux is the future” is raised: anyone with any theoretical knowledge will tell you that Linux is an old-world operating system. All hats off to Linus for stimulatting an open-source movement, and an open-source operating system, but be careful about distinguishing the fact that as an innovation from the technical innovation. Some objective critique of my comments would be welcome, but I expect that the maturity of the large part of the Linux world will buy me sniping comments and various backhanded slaps than a real discussion. At the end of the day, Linux has many of the big problems that Windows has: the marketing spin, the attempts to fit it into everywhere and anywhere (places that it is not suited), etc.
Unix isn’t gonnna die any soon. But even if it does, linux will not take its place because it is a Unix-clone (pretty bad copy too!). Inferno and Plan 9 is the future and generally distributed systems that can be used like embedded oses too! I’m sick of reading this b*** s**** over and over again.
I don’t think that ‘old-world’ is a problem for Linux. As an architecture it does just fine. With the advent of NUMA support etc it can scale on larger and larger boxes. It has the active support of some big guns e.g. IBM. How many paid engineers making the next version of Solaris or AIX or whatever? Contrast that to Linux.
*asp too – shows how well Microsoft is for handling real-life tasks.
Byt what a fscking pain – how about eweek get with the program, and put up a site that can handle the pressure.
*BSD is dying…erm I mean… lol
UNIX won’t die. There will always be demand for it. Linux can’t/won’t support *all* architectures unless there’s high enough demand or someone enough insane to port it heh…
UNICOS on Cray will survive a long time, as will Solaris. I’m not sure about the others.
I predict the demise of every currently existing operation system over time.
Whenever I peruse the analysis of Linux versus the other operating systems, I begin to understand why Linus Torvalds refrains from participating in these corporate, political and social “Linux versus the others” feuds.
What many of my fellow operating system enthusiasts fail to understand is that open source software do not compete. They merely evolve to the needs of their users. At worst, they compete against themselves. HP-UX, SOLARIS, Windows, Apple, IBM may can all throw in the white flag tommorrow, depending on what their balance sheet looks like or whether another great depression occurs.
The life of Linux, however, depends on how many souls it has touched and how many developers are enthusiastic about developing for it. Not on the economy, not on wall street analysis, not on profits and definately not on the outcome of a balance sheet at the years ending.
Linux, theoretically, can live forever as long as people exist as well as their rights to freedom. The others (commercial operating systems), technically speaking, only exist as long as their shareholders are pleased and the register keeps ringing.
Writing codes for fun, for perfection, for exhibition, for research, for the benefit of others, for openness, for critic by others, is quite different from writing codes for the corporate mission statement one is alleged to or an economic contract.
The more these simple and conspicuous facts begin to sink in, the more you’ll begin to understand Linux is in no competition with others, but itself. As long as every subsequent release of Linux is better than the previous, I think Linux will gain more grounds. It is only natural.
I use Linux for fun and for productive endeavours. Now, I could more or less do that with other operating systems (commercial ones in particular), but Linux is open, it is free, it avails me with choices unprecedented in the history of any other operating system in the world, and it has proven naysayers wrong over and over again.
I care not for whether Linux becomes popular or not. I don’t think Linux is overly hyped. I only wish hardware manufacturers and vendors, and proprietary software developers and corporations, acknowledge Linux and other open software standards in their design and implementation of their respective products, services and projects. Peharps, this is the only selfish reason why I enjoy and support the supposed “Linux hype”.
Regards,
Mystilleef
More linux zealots…. has anyone even seen the UNIX tree (www.levenez.com/unix) and these are NOT all of the unices that exist. Are you telling me that all these will perish and linux will be the last one standing?
It’s like saying Macs will overtake windows in the next 10 years hahahahahahaha (and I am a mac-user mind you)
Espacially when you see how CA’s product behave under Linux, I really really prefer working with the commercial vendors When I play with CA’s product (which is my main job btw).
—
http://homepage.mac.com/softkid (updated 15/07/2003)
There’s are three reasons why Unix will die (not in the everybody-turns-off-their-systems way, more like VMS):
1. Unix stopped innovating many years ago and went into maintainance mode, while Linux caught up and is about to add innovations (reiser4, /sys…)
2. There is, of course, a cost advantage of Linux and x86. Linux/x86 may not be an option for everybody, but it reduces the revenues of the Unix vendors significantly
3. Unix is proprietary and depends on the vendors’ faith. When the revenues go down, the vendors will have less money to put into Unix. When they put less money into it, they will have an even harder time to compete and more people will switch to Linux. It is a death spiral, and Unix users are in the middle of it. There’s some vendor lock-in. Unix user should be prepared that vendors will raise prices (to compensate the fewer sales and because the remaining customers need Unix’s stability more than those who are going to switch).
ho very nicely put Mystilleef
What about performance alone? I’ve used linux and FreeBSD, and I noticed that there was a huge performance difference between the two, being that BSD was much faster. Don’t all of these commercial distibutions of UNIX run much faster than linux?
Ask not what your OS can do for you but what you can do for your OS.
No doubt the Unix young guns were saying the same about VMS back in the day 😉
As for Unicos i wouldn’t mind having a look at that. Pity i can’t afford cray hardware lol
Apparently Unix is able to handle heavier processing loads, however Linux development is making Linux more scaleable and powerful. The 2.6 kernel is a leap in that direction. Linux conforms to a POSIX standard interface so you can call it a UNIX compliant platform, and it runs a lot of the same or slightly different software, but it’s a modern platform and there is nothing preventing it’s progress. Through the OSDL kernel development on Linux will be supported with more resources and at the same time this will not have any effect on bringing the control of Linux under one organization or dictator. There are only good things in the future of Linux, however there will probably be many scams and barriers thrown up in the USA because Linux is out of their control and it’s treatens their business who controlled the market in the past.
Personally I think that the USA should support Linux with a service model like IBM rather than try to impose a dictator like Microsoft. People don’t want to be controlled to the point where they have very limited freedom. It’s a doomed proposition. So the industry needs to adapt, and Unix is being replaced by Linux, because it’s better business, more money, etc.
Linux, FreeBSD, OpenBSD, NetBSD, OSX, … they are all cousins and have a lot in common. They share not only a lot of libraries and tools (what has much more code than in the kernel itself) but also a lot of concepts and, the most important point, as far as they keep the source core open, all them can benefit from the progress and innovation of the each other.
So, I can’t find any reason to stop, except perhaps money, any of these projects to keep going on. If people that work in Linux kernel find some breakthrough, they will be followed by *BSD and vice-versa.
Also, in the very improbably hypothesis that one project stop, the innovations and ideas they gave to us will still exist inside the others.
I see these projects as competitors in an evolutionary process and don’t mind which name they have as far as they keep being open source.
People don’t want to be controlled to the point where they have very limited freedom.
Apparently a majority of us Americans do. That’s what we call freedom. The freedom to choose Microsoft or Apple. Isn’t that nice?
Now whenever someone mentions freedom I have to ask what kind of freedom they really mean. Who would have thought there are so many different types out there.
What about performance alone? I’ve used linux and FreeBSD, and I noticed that there was a huge performance difference between the two, being that BSD was much faster. Don’t all of these commercial distibutions of UNIX run much faster than linux?
The issue of perfomance is as vague as it is subjective. What performance means in the desktop sphere is quite different from what it is interpreted as in the server realm. Besides the definition of performance varies from user to user.
While desktop users errorneously misconscrew performance for responsiveness, the network and system administrator worry about system load, uptime, efficient resource allocation, reliability and stability of running processes and deamons.
With regards to desktop responsiveness, it vary from distro to distro. Linux distributions like Redhat, SUSE, Mandrake, ${substitute commercial distro here} are generally configured to run a ridiculous amount of deamons out of the box, the price for letting Mr. Knowitall to the job you are supposed to do to begin with. The unnecessary amount of deamons and/or services clugs the systems resources, consequently making your desktop seem less responsive.
In addition to the above, because Mr Knowitall doesn’t know your system’s specification, a lot of user, software and hardware settings are preconfigured using the safest ( euphemism for slowest) and most generic configurations and optimisations.
Thank goodness Linux offers alternatives. There a number of distributions which allow you to customize, configure and optimize your system to your hearts content, thereby resulting in system responsive. Linux from the Scratch (LFS) Slackware, Crux, some variants of Debian that use modern archictecture optimizations (i686 not i386) and the most reputed amongst them, especially with regards to optimization, Gentoo, amongst a host of others offer desktop responsiveness equal and in many cases better than any operating system in the market.
Whether FreeBSD is more responsive than Linux and vice-versa is absolutely subjective as a lot of variables factor in. With regards to performance, however, I would argue that both of them are playing on the same level with little advantages over the other. As of today, I don’t think any of them are milestones ahead of each other.
Technically speaking, the commercial unices can outperform Linux, but surprisingly enough, it is bearly noticeable the average user. In fact, most of the commercial unices make horrible desktop computers. I’m talking about Solaris, HP-UX, AIX and co.
Regards,
Mystilleef
I agree with these guys. Unix will “eventually” be replaced by Linux. Note the word eventually. Theyre not saying next month or next year. It will a long time sure, but it will happen. This is why the likes of IBM and HP are ramping up Linux development. Another point from the article that is completely true is this line:
“He too said Unix will be replaced by Linux over time as it simply does not make economic sense for many businesses to be developing on an old proprietary operating system anymore.”
When you consider what the 2.6 kernel and future rapid development is doing for Linux …..and the fact that Unix is stagnet and costs money, this is spot on correct. Right now Unix makes a lot of sense in a lot of places. In the next 5 years though it will start to fade and in 10 years it wont be the OS of choice.
“No doubt the Unix young guns were saying the same about VMS back in the day 😉 “
We’ll kill guiVMS soon!
True….and tell me how relevant VMS is today…..lets be realistic.
all these hardware companies are making Linux more Highend…so one day in the near future, there will be no reason not to use Linux over your own product.
Was when Linus states that the exiting things that are happening are well above the level of the kernel. Some people have quipped that the OS ends with the minimum dependancies that are required to run Bash (or whatever shell you want) (read [email protected] for some real fanatics). Everything above that is just ‘applications’. In that light, most of the things that make OSX the favorite of Mac fanatics are the high-level applications. Same thing goes for the people that use Windows, and the GNOME v. KDEers.
When Eugina posts some new editorial or link about systems like NeXT and BeOS, the items that are most commonly remarked on are much higher up the food chain than the level that the kernel hackers are working on. I don’t think that there is ANY REASON WHATSOEVER that an environment such as those can be built on top of a Linux core. Don’t like X? Than build something else on top of the POSIX core.
Abstraction. This is what you need to run a user-freindly environment. OSX isn’t cool to Mac guys for its darwin core as much as it is cool for its ‘iLife’ applications and GUI environment. Linux is cool to Linux lovers because the level of abstraction is very minimal and appeals to the tweakers out there. At the same time, that very exposure to the axles and gears is what tend to put off your average Windowsian or OSX lover.
Linus understands his place in the OS universe.
I agree to a point……
I agree on the fact that yes, Jon Q user isnt all that interested in the kernel. However, I completely disagree when you say: “I don’t think that there is ANY REASON WHATSOEVER that an environment such as those can be built on top of a Linux core.” Thats just completely false. Anyone can put whatever they want on the Linux core….they just have to put their minds to it and do it. Take a look at Enlightenment E17 for linux rendered in OpenGL like OSX as an example.
“True….and tell me how relevant VMS is today…..lets be realistic”
Well don’t tell the 1 million users of VMS that’s it’s not relevant.
haha…. 1 million….maybe, but Im inclined to scoff at that. There is really no way to prove that many people are using it. But thats really besides the point…..take a look at the millions of Unix + millions of Windows + millions of Linux + millions of Netware and then tell me how relevant VMS 1 million user base is.
I think it’s probably more likely that *BSD will replace Unix. The BSD license makes more business sense than GPL does.
I think the million figure refers to the number of machines running VMS and not the total end users. Also VMS is a multi-billion business for HP.
> Don’t all of these commercial distibutions of UNIX run much faster than linux?
Well the general advice of picking the right tool for the right job applies to this case as well. Commercial Unices do a much better job on an enterprise level. When it comes to serving massive amounts of data (TeraBytes) while sustaining predictable performance and stability, Linux has still much work to do. Linux and the *BSDs are great platforms for much smaller requirements (i.e. web / file / print serving). A commercial Unix may be overkill, depending on your volume and needed functionality.
Generally speaking, yes, commercial distributions run much faster than Linux. The scalability of the RISC architecture is still yet to be challenged by Linux based solutions. I’m thinking x86 and AMD platforms here. However, the penalty for superior performance comes at a cost.
1. Unix stopped innovating many years ago and went into maintainance mode, while Linux caught up and is about to add innovations (reiser4, /sys…)
I think you are not aware of advances in operating systems design. Even back at university (10 years ago for me) it was obvious that Linux was an old-world rather than a new-world operating system (e.g. at the time one of our classes was experimented with Ameoba as a distributed OS). The Linux innovations you talk about are really just refinements and amendments to unix (little ‘u’) itself, these are not innovations in the broad landscape of operating systems.