Slashdot Mirror
Unix vs. Linux Career Prospects
TreeRat writes "Here is a CNN article covering the possible career prospects for IT professionals knowing Unix vs. Linux. It concludes (surprise surprise) that most Linux professionsals are using other forms of Unix at work as their primary OS, but that demand for Linux people should be much higher in the future. "
[... ]Combine that with the fact that it is free and runs on damn near any hardware and most everything on it is written in C, which can be ported to any other platform or OS, and that it can do most anything Solaris or Irix or AIX or HPUX or VMS can do for a slightly monumental price, there is literally no stopping it; it will eventually dominate.
I like Linux. But you try to run Linux on a serious 4 CPU Sun Enterprise 4000 server; Solaris and Irix have much better thread implementations than Linux. I, along with three others in our systems group, are responsible for hundreds of Linux boxen at work. We also run a bunch 'o Sun {sparc,x86} SunOS 4/Solaris 2, AIX 4 on RS/6000, IRIX 4/5/6 on everything from Indigo to Challenge-S hardware, and SCO on x86 hosts; plus we have NT all over the place on people's desks -- many prefer Linux over NT on their desks but require a Windows application to perform their job. That's life.
I don't want to give up this heterogeneous environment of differing hardware and operating systems. We spent most of the seventies and eighties figuring out that homogeneous protocols was the solution to heterogeneous platform incompatibilities; standardizing on only x86 throws into the toilet nearly 30 years of computer science research. Just as standardizing on only Windows or Linux would lose us yet another degree of freedom when trying to solve various problems with a computer. Let's face it, A world of homogeneous Linux and only Linux would get stale... When that happens I think it will be time to re-invent those old LMI/Symbolics LISP/OS systems on modern hardware, begin a re-implementation of PRIME/OS or VMS, or maybe even take a job selling shoes.
Linux is not the only free and open source operating system which ports to a wide variety of hardware; NetBSD also shares this bragging right in about the same proportions of differing architectures supported. And that is Free software's most important contribution, not just that Linux and NetBSD are free for the taking (sans the greater licensing freedom any commercial organization would enjoy with NetBSD), but that finally we have the choice of a portable and common cross architecture OS environment and development API; this is almost the realization of true CPU independence in software. That success is a computer science milestone, which while irrelevant to those who have standardized on only x86 as a platform, is historicaly significant relative to much of the early thinking about handling architecture diversity in the early days of computer science.
But this is not entirely a theoretical excursion. Embedded systems manufacturers and possibly even new workstation manufacturers (just look at the ARM based Netwinder) could exploit performance gains from new architectures as these OS's grow in popularity. Since they're designed to be ported, and since the APIs are common, application software should port easily. WOW! I get choice, and my software "just works," and someone makes a buck. Sounds like good 'ol capitalism exploiting the potential of a new market previously unavailable in the closed source proprietary era.
Amazing how long it seems to take the mainstream press and corporate pundits to realize concepts that our community pretty much takes for granted. ...)
For one, knowledge, especially computer-related, can never hurt you, but only help you. So what if you don't use it directly at work? It's the concepts, and the exercise of your mind, that's really important.
And since industry generally follows the bucks in any situation, I think linux specifically and Open Source in general will continue to become more viable contenders in the corporate IT realm.
(But of course, you already knew this
illum oportet crescere me autem minui
I dimly remember from my first logic course (before another one in symbolic logic) that "A false premise justifies any conclusion." What this means is that given a conditional statement which is true, if you then negate the premise, any conclusion could be true. For example, let us say that the statement, "If it rains tomorrow, I will go to the movies," is true. And if it doesn't rain tomorrow? I might or might not go. Either conclusion could be true or false. A false premise justifies any conclusion.
What has this to do with the article? The false premise is that Linux is a completely different OS than UNIX. This fundamental mistake undermines their whole article. If Linux is actually in most respects a form of UNIX, which I believe it is, then we can conclude with the article that Linux skills alone won't cut it, because they say that UNIX skills alone won't cut it. Or we can say with the article that Linux is a UNIX and UNIX isn't going away, therefore Linux skills are valuable. Or... well, you get the idea.
For reasons known perhaps only to the author, this article opposes Linux to UNIX in terms of skills and knowledge. This is a false dichotomy. Many of the pro-UNIX people quoted learned their way around Linux, and I should think that a good grasp of the fundamentals of any UNIX (or UNIX-like) OS is readily transferable to another UNIX OS. I do not go into a panic switching between Linux and IRIX, for example.
If one believes that Linux is UNIX for all practical purposes, this article is actually quite encouraging, because it tells us that UNIX skills are in demand, that Linux is a great way to learn networking, configuration, and sysadmin stuff, and that businesses are considering Linux as a low-cost, reliable UNIX OS.
What disturbs me about this article is the author's separation of Unix and Linux. By doing so, it is implied that Unix is this monolithic OS and that Linux is different. The last time I checked, IRIX, HP-UX, AIX, Solaris, etc... each had their own distinctive quirks. It is not as bad as the differences between BSD4.3 and SYSV ten years ago, but there are differences.
Fortunately, some of the quotes pointed out that Linux can serve as a very good means of educating yourself about networking, security, etc... This should have been the main focus of the article. The fundamentals of Unix/Linux OS's are the same. Once you have mastered these fundamentals, moving from one flavor of Unix to another is relatively easy. I realize that becoming an omnipotent master of a particular flavor of Unix does take time and experience. However, having proficiency in another variation of Unix makes this transition quicker and more efficient. But only if you RTFM!:)
Some clueless HR managers probably don't realize the fundamental similarity between the various unices. "I'm sorry, we are looking for a person with experience with HP-UX, but you are only an expert on Linux." I take the attitude that if you could master one form of Unix, you could probably master another one.
I believe that there is general consensus that Unix is Unix is Unix. The fundamentals are the same, and knowledge of one variation can be easily transfered to another flavor of Unix.
As another illustration of this, consider the case of SunOs vs. Solaris. SunOS4.X was primarily based on BSD4.3 while Solaris is based on SYS V.4. When companies made the switch from SunOS to Solaris, did they fire all their sysadmins and replace them with ppl more experienced with SYSV? Or did those old Sun sysadmin make the transition to this "new" OS?
On a similar note, what happened when SGI switched from NeWS to X/Motif? Or what happens everytime there is a major update to the "same" OS, like HP-UX9 to HP-UX10?
The thing is tho, once you've learned a few Unixes, you can pretty much wing it with any of them. I learned Unix initially on IRIX 5.1, and Solaris 2.? - I decided I liked it, got Linux and put it on my PC, and learned some admin stuff. (The hard way. I only fried two hard drives, a sound card and a cheap mono monitor before I learned to stop doing that... :)
:)
The first actual admin work I did was on an IRIX 5.3 system (volunteer work is great, you gain experience without needing to already have it. If you're a Comp.Sci. or Engineering student and you've got a chance to volunteer for stuff in your field, GRAB IT - it's the difference between getting a cool job later, and being told 'Sorry, we're looking for someone with more experience.') Then my first paying admin work was on systems with IRIX and Solaris both - while there's a few major differences, most things you learn apply anywhere. Now of course, I'm working with Linux and IRIX, and I replaced the Solaris box here with another Linux install. (Sorry, but the latest solaris is too slllloooooowwwww on a Sparc1 system.
Once in a while, I even get to poke at my friend's *BSD box. (He installs the latest whatever is available. New Linux kernel? Change system over. New FreeBSD? Change... New NetBSD? Change... Eventually he'll find something that supports all of his freaky obscure hardware... And even then, most things are the same, it's just the minor differences to get used to. Just like talking to a bunch of people (in english) with really thick, and different accents, once you get past the fact that the guy over there emphasizes a different sylable than the guy over here, they're all still speaking the same language. Except for my old data structures prof. We're not sure what language he speaks.
Does anyone know how transferable Linux knowledge is to UNIX? Of course, it depends on the UNIX that one's talking about, I know. I assume that Linux knowledge will transfer to a sys V based system better than to a BSD based system, since most Linux dists are based on sys V standards (correct me if I'm wrong).
/usr/bin/perl, but IRIX puts it in /usr/sbin/perl, so scripts can break if you don't change the header to compensate.)
Well, it also depends on which distribution of Linux you're using - for example, if you want to learn about boot time configuration, Red Hat uses an arangement very similar to what IRIX uses, but Slackware (last time I checked was I while ago, so it may have changed) uses a completely different one. (I'm talking about the layout of the rc files here.) In a lot of things though, it will only be a little different. Shell commands are almost always the same. Some systems will want different parameters for a few commands - though that is mostly because of different versions of utilities. (For example, in current Linux distro's, 'ps' dosen't need a '-' in front of it's parameters, but IRIX yells at you if you don't include it. ps is an external utility, but I'm too lazy to replace it on the IRIX systems right now, so I just remember the '-'.) Some other things will be totally different, like file systems. Formatting drives in IRIX 4 was scary, it took me an hour just to interpret and calculate the parameters, where in Linux it's just a case of typing one command and remembering to tell it which drive to wipe. If you've got all the same versions of external utilities tho (like NFS, Perl, Tcl/Tk, (t)csh, etc...) then you should have only a few differences to worry about, since most of the environment will be the same. (Oh, and make sure you know where those utilities are. On Linux perl defaults to
Confusion over the word unix is commonplace. In this instance, the article splits unix systems into two categories: Unix and Linux, as if all unix systems besides Linux are somehow authentic or legitimate, while Linux is not.
What makes a unix system a unix system? Most of us already know the answer to this but probably have no desire to spell it out. If you know what a unix system is, inside and out, you know perfectly well that Linux is just one implementation of unix, as are Solaris, AIX, HPUX, etc.
The code base does not have to stem from the original tree to be considered a legitimate unix system, nor does it have to be proprietary. As Neal Stephenson points out, "Gnu is an acronym for Gnu's Not Unix, but this is a joke in more ways than one, because GNU most certainly IS Unix. Because of trademark concerns ("Unix" is trademarked by AT&T) they simply could not claim that it was Unix, and so, just to be extra safe, they claimed that it wasn't."
Despite the pedantic point that the trademark is now owned by someone else, he hit the nail on the head.
If I'm not mistaken, the trademark has been owned, at least, by the following: Bell Labs, AT&T, Novell, X/Open, and The Open Group. However, the ownership of the original code has passed from Bell Labs to AT&T to Novell to SCO. Again, if I'm not mistaken, the trademark and the original code base are now owned by two different entities. It is in my opinion ludicrous for anyone to assert their rights to the word unix anymore - though I would not encourage anyone to try naming their product XYZ Unix or some variation along the lines of UnixWare, I think that everyone has every right to describe their systems appropriately, including people who run Linux; and Linux is a unix system.
If you look at The Open Group's discussion of their trademarks, you'll see just how silly the whole idea of claiming ownership to the word unix really is; these measures are too late. Language has owned the word for some time now. Hardly a day goes by that I don't hear, in common speech at work, at least one of their examples violated on how they don't want you to use the word unix. Specifically, I constantly hear people refer to "the unix box," "a unix box," or "one of the unix boxes," when in fact they are talking about different types of machines and different implementations of unix.
Do not let this word game encourage people to perceive Linux as somehow less deserving of the name than its commercial counterparts. A unix, by any other name, is still a unix.
Steven Pritchard, a contract HP-UX system administrator at a Peoria, Ill., manufacturer, sees a lot of interest in Linux across the organization -- especially since major vendors began to support it -- but no actual applications.
Yet.
"The demand for just Linux skills isn't all that high yet," Pritchard says. "Most of the Linux people I know don't have Linux-specific jobs."
Yet.
Those who ignore the future are doomed to regret it. Linux is the future, there's no denying it (flame all you want, but it's the truth and you all know it). The article mentions it being "barely 8 years old;" how old is Microsoft now? Twice that? Compare how far Microsoft's "advances" have been in that amount of time to how far Linux has come in half that time. Now think how far it's going to go in the future with more and more supporters and programmers hacking at it with each passing day compared to how far Windows 2000 (tm)(r)(s)(c) supporters and programmers will be able to take it in the same amount of time. Clearly, since Linux has already surpassed it in speed, reliability, stability and the ability to be almost instantaneously patched against things like bugs and security flaws (how long does it take MS to come out with a service pack for bugs they knew were there before they even released it?), it will always be ahead of MS in those areas. Combine that with the fact that it is free and runs on damn near any hardware and most everything on it is written in C, which can be ported to any other platform or OS, and that it can do most anything Solaris or Irix or AIX or HPUX or VMS can do for a slightly monumental price, there is literally no stopping it; it will eventually dominate.
Now, taking all that into consideration, does it make sense for any company or sysadmin/programmer to just ignore it now? Perhaps it doesn't have all the 3rd-party applications that Windows has... but it will. Perhaps support for a few peripheral hardware devices and interfaces doesn't exist or needs to improve... but it will. And some bright, sunny, shiny, happy day in the very near future, when Linux... or even just *nix... rules all, those of us who have believed in it all along will be so far ahead of those who have ignored it all along that they won't have a chance of learning all they need to know to <buzzword> compete </buzzword> with us. (By "us" I mean companies and individuals...)
Prognostication is the key to future success. A lot of us already know the future and are well prepared for it; a lot of people don't have the vision or imagination to see it, or are so comfortable with the status quo that change frightens them. They will be left behind. Oh, eventually they'll catch up; but for a good year or two, we'll be the hottest commodities on earth.
I for one can't wait...
"All truth passes through three stages: first, it is ridiculed; next it is violently attacked; finally, it is held to be self-evident."
linux is a UNIX variant. In fact all things known as UNIX are really UNIX variants. In my experience 95% of what you do on one system applies directly to all of them. This applies between Linux distributions as well. The shells are the same on most UNIXs, all use X windows, ftp, telnet or ssh, similar commands, and similar directory structures.
If you can use Linux you can use Solaris 7, AIX, FreeBSD, OpenBSD, NetBSD, SunOS, etc, etc.
Anyone who cannot cope with mathematics is not fully human.