The Steady Decline of Unix

stinkymountain writes "Unix, the core server operating system in enterprise networks for decades, now finds itself in a slow, inexorable decline, according to Network World. Jean Bozman, research vice president at IDC Enterprise Server Group, attributes the decline to platform migration issues; competition from Linux and Microsoft; more efficient hardware with more powerful processor cores; and the abundance of Unix-specific apps that can now also run on competitor's servers."

Comments

Uh huh

[MightyMartian]

So the bulk of Unix's decline comes from competing *nixes, in particularly Linux.

News at 11.

Re:Uh huh

[mtrachtenberg]

Nah, it's not Linux eating into UNIX's market share. It's stuff like the Debian, RedHat, Android and Ubuntu OS's.

Re:Uh huh

[who_stole_my_kidneys]

The decline is from the price point. My last place of employment had 1 HP UX server that costed upwards of 25K for software and specific HP hardware to run on. migrating to windows cost a fraction of that in OS licenses and hardware, even though it took 8 windows servers to do what the one UX server did, it was still cheaper.

Re:Uh huh

[PPH]

Likewise, there has been an alarming drop in the number of people who use Kleenex when they blow their nose.

Re:Uh huh

[Freshly Exhumed]

So Unix is dying *yet again*? I hate that trope.

Re: Uh huh

[samkass]

The biggest UNIX vendor in the world-- Apple, Inc.-- has had its UNIX laptops increase in market share in almost every quarter for the last 5 years. And although it's not certified UNIX like its desktop sibling, iOS is based on the same core... not sure what value differentiating this specific market segment offers. In the server, Linux seems to be doing just fine, and is close enough to UNIX for it not to matter.

Re:Uh huh

[major_handicap]

Especially if it's owned by Oracle...

Re:Uh huh

[gstoddart]

Combine that with the fact that Solaris is now in the hands of Oracle, who are squeezing out everybody who doesn't have a support contract and pissing off people who used to use it ... or that HPUX is still in the hands of HP (where technology goes to die) ... and what's even left?

AIX is still around, but I have no idea of how widespread. Beyond that, I'm hard pressed to think of another commercial version of UNIX I've encountered. (That doesn't mean they don't exist, but they were never in any shops I was in.)

That pretty much leaves Linux as the primary UNIX-like-thing for most people.

Re:Uh huh

[localman57]

Why would anyone in their right mind switch servers from HP-UX to Windows?

Why wouldn't one? Based solely on who_stole_my_kidney's anecdotal argument, and your rather content-free counter-argument, I'd be inclined to follow his advice.

Re:Uh huh

[localman57]

Yeah, but that's just splitting hairs. If you factor in all the *nex's , people are blowing about the same amount. It's just more diversified.

Re:Uh huh

[telekon]

"Reports of my decline have been greatly exaggerated."
    -- UNIX

Re:Uh huh

[Anonymous Coward]

Why do we even need to have this conversation? If it looks like a duck, swims like a duck, and quacks like a duck it isn't a duck unless it is branded a duck? This is so fucking stupid. Linux is a UNIX type of operating system, so UNIX isn't in decline.

Re:Uh huh

[fuzzyfuzzyfungus]

Apparently SCO UNIXware is still around, though I assume that it's more of an absurdist performance art piece with a couple of legacy customers than an actual operating system at this point.

Re:Uh huh

[Anonymous Coward]

>>>W
>>>>H
>>>>>O
>>>>>>O
>>>>>>>S
>>>>>>>>>>>>>>>>>>>>>>>>>>>H

Re:Uh huh

[mcrbids]

Oracle is a flavor of SQL. So is MS SQL Server and PostgreSQL. I'm guessing that you meant MS SQL Server?

Re:Uh huh

[mrclisdue]

"However, reports of our decline are spot on!"
        --Eunuchs

Re:Uh huh

[Anonymous Coward]

Because we here at Slashdot like to be accurate, especially when poking holes into other's arguments.

Linux is a kernel, nothing else. It's the distributions which are supposedly eating into UNIX's market share, but really just chewing away at Redmon's marketshare.

UNIX market share only appears to be dwindling because it takes less hardware to do the same jobs they were doing just a few years ago.

People are consolidating 10 to 20 servers onto single or two small/medium sized servers.

Total server counts go down, productivity goes through the roof - the numbers are just that numbers, without any details as to why they've shrunk.

Re:Uh huh

[Shatrat]

Because instead of replacing 1 HP-UX server with 8 Windows servers, it could have been replaced by 1 Linux server.

Re:Uh huh

[Anonymous Coward]

Switching from a serious platform to a limited Fisher Price(tm) My First Server(tm) whose only benefit is the ease of setting up some basic functions. Uh huh. Doesn't sound like a good idea.

You got multiple choices that are not only better, but also similar to HP-UX enough that the migration cost is a tiny fraction of that for Windows.

Re:Uh huh

[Anonymous Coward]

Oracle is the reason that my employer is switching from Solaris to Linux. We were one of Sun's biggest customers, too.

Re:Uh huh

[MightyMartian]

I'd be curious to know what exactly the HP-UX server did that could so easily be moved over to a cluster of Windows servers, myself. Apart from believing that one can buy eight servers with Windows Server licences and come out much less than $25k, I'm just trying to sort out what this server would have actually been running that one could simply go "Oh well, we're going to Windows now."

I've found damned few cases in my experience when wholesale moving from one platform ecosystem to another platform ecosystem was a viable activity in and of itself, unless it was part of a long term strategy of retooling and recoding. I've seen some organizations move from Unix to Linux, but generally with the notion that porting apps was relatively easy or had already been done. But to move from *nix to Windows is a big deal, unless you're running everything in Java EE, in which case why would you completely change your ecosystem with other *nix variants out there?

Re:Uh huh

[Bill, Shooter of Bul]

Because Unix is architecturally closer to Linux than Windows, would be the obvious answer. But seeing as there aren't many details given about the role of the servers*, it would be wise to reserve judgement in the particular case, even if it makes little sense as general advise.

*Often times under further questioning admins doing similar switches, the irrefutable answer becomes " exchange integration" or something of the like. Someone offering advise can't very well redesign the entire IT operations over coffee for a company he doesn't work for.

Re:Uh huh

[gstoddart]

Apparently SCO UNIXware is still around

Must be those people who paid the extortion fee.

I find it astounding that any organization wouldn't have long ago asked themselves WTF they're running it for -- and I can't imagine there have been any updates to it in a long time.

Or, it's alive and thriving -- I have no idea. It's been a dead horse for years to me. If someone told me they'd be putting in a new server with SCO Unixware on it -- well, I'm not sure of how long it would take me to stop laughing.

Re:Uh huh

[K. S. Kyosuke]

If HP-UX is expensive enough, Windows can actually be an attractive choice. Remember, just because you have a technically better solution, you can't just slap a $100k price sticker onto it and expect that people won't go for the cheaper option that can be boosted up in functionality with a bit of third party stuff.

Re:Uh huh

[MightyMartian]

Other than to honor market doublespeak, so far as I'm concerned, you can lump Unix in with Linux, OS-X, QNX and the other variants and -likes that make up the *nix ecosystem. It's a helluva lot easier to port an application from Unix to, say, Linux, than it is to port from Unix to Windows, unless you use a compatibility layer like Cygwin. Man, I wouldn't want to use Cygwin too much on a production server. The only time I ever did it was to get a decent radius server running on a Windows machine. It worked reasonably well, but I was very happy to move to a Linux server due to glitches.

Re:Uh huh

[ganjadude]

pfft, I dont care what they say 2014 WILL be the year of Unix on the Desktop!

Re:Uh huh

[reedk]

But GNU/Linux is POSIX-compliant, and that's what people mean when they lump Linux with UNIX or call it "another UNIX."

Re:Uh huh

[Penguinisto]

The decline is from the price point. My last place of employment had 1 HP UX server that costed upwards of 25K for software and specific HP hardware to run on. migrating to windows cost a fraction of that in OS licenses and hardware, even though it took 8 windows servers to do what the one UX server did, it was still cheaper.

Agreed, but for one thing: Amortization.

In a previous position, one of my clients had an ancient IBM 9370 mainframe going. Mind you, for the business size (about 200 employees) and what they used it for (a string of automotive dealerships), it was 1) overkill when purchased (they overestimated their expansion plans by couple of factors), and 2) hellishly expensive. I think they paid a solid 7 figures for it, but cannot remember exactly how much.

Thing is, its amortization schedule was roughly 2 decades at least if I were to guess. They bought it in 1984, I last saw it in 1999, and I bet it's still running today if their CFO has anything to say about it. Anything after the amortization date is pure gravy for them, methinks.

Another reason why I say it's likely still around:
- It does pretty much everything they need it to.
- Unlike most x86 servers (but not unlike most high-end UNIX boxen) could probably take a direct 1-megaton nuclear detonation and keep going without so much as a dropped routine.
- You could swap out everything but the power cable without shutting it down (I think only a microcode and certain IPL patches would force any real cold downtime). Yes, I'm including RAM and processors.

And to top that off, I see a lot of similar situations with other ancient boxes (HPUX, AIX, SunOS, you-name-it.)

Re:Uh huh

[mellon]

C:\YOU\LOSE> tar cf - . |ssh windoze2 "tar xvf -"
Broken Pipe
C:\YOU\LOSE>

Re:Uh huh

[maxwell demon]

If it weren't Unix, it weren't an Unix variant. But since not all Unices are identical, every single Unix is an Unix variant.

Re:Uh huh

[binarylarry]

Gnu's
Not
Unix
But Is
For all intents
And
Purposes

I for one welcome our GNUBFAP overlords!

Re: Uh huh

[Em Adespoton]

Yeah; I've been wondering what exactly they mean by UNIX here -- are we talking POSIX compliant OS (they almost all are these days), something based on BSD/AT&T code (BSD derivatives like OS X and FreeBSD, plus SVr3+ derivatives like HP:UX) are are we talking purely SVR 4+, and thereby mean SCO offerings when we say UNIX?

See http://www.levenez.com/unix/ for a nice list of UNIXes. Interestingly, Windows NT isn't there, even though it is POSIX compliant.

Re:Uh huh

[Chris Mattern]

Thing is, its amortization schedule was roughly 2 decades at least if I were to guess. They bought it in 1984, I last saw it in 1999, and I bet it's still running today if their CFO has anything to say about it. Anything after the amortization date is pure gravy for them, methinks.

If only it didn't require power and A/C, and if only it didn't require support. Power costs for one of these beasts is most likely all by itself more than it would cost to buy a modern replacement.

And support from IBM is astronomical, if it's even available. If you're not paying for support, well, that's another kind of cost. But you won't find out about it until the bill comes due.

Re:Uh huh

[mellon]

Most likely _all_ of the world's largest organizations run Linux. But only _some_ of them run Windows Server 2012. So I think the OP got it right.

Re:Uh huh

[CanHasDIY]

Because HPUX sucks more than Windows. It still doesn't mean Windows is any good.

I think they did a South Park episode about that...

Re:Uh huh

[PPH]

while some of the worlds largest organizations run Windows Server 2012.

Larger organizations can hide the TCO of an army of IT people better then small ones.

Re:Uh huh

[loufoque]

Isn't IBM pushing RHEL these days?

Re:Uh huh

[fahrbot-bot]

Linux is a UNIX type of operating system, so UNIX isn't in decline.

The article is mainly talking about the Unix versions like HP-UX, Solaris, etc... and the iron used to run them, focusing on installations that require many 9s of reliability, fault-tolerance / fail-over and up time. Their argument is that those systems are more mature, reliable and capable (and more expensive) than most Linux systems. Many installations are realizing that they don't need that all that and less "capable" Linux and/or x86 systems are just fine - for many things. Personally, I believe in using the right tool for the job, not necessarily the best and/or most expensive tool. The trick is defining the job correctly.

Re:Uh huh

[kotj.mf]

It's still surprisingly big in retail - it runs a crapton of local back-end stock keeping applications at the major retailer that I worked for awhile back, and I've since heard that they've virtualized it to deploy on their "next generation" in-store platform.

20 years ago, it was really the only game in town for Enterprise UNIX(tm) on Intel, and given how much it costs to design, buy, and deploy ANYTHING that's going in to 2000+ remote locations, it's going to stick around for quite a while more.

Re:Uh huh

[Anonymous Coward]

Windows NT was certified as POSIX compliant.

Windows NT!

Re:Uh huh

[Anonymous Coward]

Unix has three definitions.

1. Legacy Unix, based on the AT&T Unix source code. eg: Solaris and AIX.

2. Certified Unix(tm). eg: Solaris, AIX, and OS X (apparently not included in their decline of Unix numbers).

3. Unix-like operating systems. eg: Linux, *BSD. (*BSD is also somewhat legacy in that AT&T incorporated their source code).

Re:Uh huh

[lipanitech]

Does this factor in mac osx considering it's a unix based operating system.

Re:Uh huh

[maxwell demon]

Care to back that up in any meaningful fashion? Show me where there is a single Linux server that can do the same thing as a HP UX server does that cannot run Windows. Bring it on, fanboy.

Exhibit 1: A Linux server, brand new.
Exhibit 2: A HP-UX server, 25 years old, clearly unable to run Windows.

SCNR

Re:Uh huh

[mellon]

Added bonus: they can use it to heat their building in the winter. I once spent a couple of months sitting next to one of those babies in an un-airconditioned space in the summer. Despite being close to the water, where it was consistently cool outside, that machine kept it nice and toasty inside. I still twitch a little if you say dazzdee. You can keep amortizing a machine while it's powered off and in a warehouse, and save yourself a bunch of money emulating a 370 in software on some reasonably powerful Xeon server. Or just sell the thing for scrap and write it off as a loss.

Re:Uh huh

[Penguinisto]

I can give the counter-arguments against using Windows:

* You're guaranteed to suffer every month for maintenance (Patch Tuesday), and require multiple machines not just for capacity-matching, but for redundancy if you want anywhere near the same uptime. In spite of an MCSE/MCSA being cheaper, one competent UNIX admin can maintain 3x the machine count than an MCSE/MCSA can - unless you feel like springing for a lot of pricey add-ons/upsells to keep admin FTE headcount down (e.g. automation via SCOM,SCCM and etc). It doesn't take too much for that SA contract cost to match or exceed the HP one, especially if the Microsoft products have the word "Enterprise" in the product title/license.

* All that aside, I haven't even touched on increased space, power consumption, cooling/HVAC, and etc... the costs scale up almost exponentially in larger installations.

Re:Uh huh

[lgw]

I highly doubt IBM had anything to do with it. I'm sure it was a UCS system (Reynolds and Reynolds, today). They were grandmasters of hidden costs. But power is likely not one of them - they're surprisingly efficient for what they do.

Re:Uh huh

[Anonymous Coward]

server that costed upwards of 25K

"cost" works fine there. Please. I'm begging you.

Re:Uh huh

[Samantha Wright]

Except for, y'know, Ken Thompson's original distributions like V7. Which are about as unvaried as you can get.

Re:Uh huh

[SimonTheSoundMan]

If only it didn't require power and A/C, and if only it didn't require support. Power costs for one of these beasts is most likely all by itself more than it would cost to buy a modern replacement.

That's why we have OS X. It's powered by the users own sense of self importance.

It's Unix certified too.

Re: Uh huh

[lgw]

They're talking about the old-school, high-dollar Unix systems: Sun, HP-UX, AIX, and so on. All but Sun have largely vanished, and Sun is pretty much dead. It's all Linux now. And that's a huge change in mindset from when Slashdot was new (which isn't that long ago, really), when everyone was debating whether some "free" OS would ever be taken seriously in datacenters, beyond maybe stateless web servers. Yup, pretty much it was.

Re:Uh huh

[aix tom]

Well, a highly-modified closed-source "Unix A" kernel is also only reminiscent of another "Unix B" kernel.

Re:Uh huh

[0123456]

That was rather my point. I have a hard time seeing what you could be doing on an HP-UX machine that would more readily move to Windows than Linux.

My guess is some MBA said 'no-one ever got fired for buying Windows' and they had to kludge up whatever used to run on Unix into multiple Windows servers to keep his/her bonus up.

Re:Uh huh

[MightyMartian]

It's sufficiently compatible that sources can be compiled. It's as much a *nix as an oddball like, say, Coherent.

Re:Uh huh

[MightyMartian]

All I can say is "I'm glad I don't work there." I can only imagine the nightmares they must be going through moving from HP-UX to a Windows Server cluster.

Re:Uh huh

[oPless]

Yep, and it was pants, so they bought an equally half assed, but much more complete implementation of a posix subsystem called interix and rebranded it.

It was pretty good, but then again ... Windows NT

Re:Uh huh

[MightyMartian]

The POSIX subsystem in Windows is partially POSIX 1 compliant. In other words, it's pretty much useless for any kind of meaningful port. You have to turn to Cygwin if you want something approaching full POSIX 2 compatibility.

Re:Uh huh

[mlts]

Historically, this is what has killed most of the high-end UNIX architectures except for Oracle and IBM.

Even factoring in Microsoft's price hike, it can be cheaper to go with a stack of 1U boxes (or a blade/enclosure array) all running VMWare or Hyper-V than it is to continue paying the service contract costs on the older hardware.

Of course, this advantage also applies to Linux, *BSD, and Solaris x86.

In my experience, the trend seems to be from the older platforms (PA-RISC, POWER, SPARC) to a UNIX that runs on x86, rather than a complete platform change to Windows.

Re: Uh huh

[samkass]

MacOS X's core OS is open source. You can download the kernel and recompile it and swap yours in if you want to, and all the standard user space stuff is basically FreeBSD.

Also, it is a certified UNIX 03 operating system, so it is more "UNIX" than Linux, which is what I assume you're comparing it to.

Re: Uh huh

[turbidostato]

Easy: a supported version of ls
(which is to say, making the life easier for the sysadmins you already have in your payroll).

Re:Uh huh

[gravis777]

No kidding. I had to go back and read the headline twice after reading the blurb and the article before it made sense.

Well, no DUH that Unix is loosing ground to Linux. I honestly cannot remember the last time I saw a data center with true UNIX machines (oh wait, yeah I can, it was 14 years ago, unless you want to include OSX as a UNIX, in which case that was five years ago). This has been happening for like 15 years.

When I saw the headline, I was thinking "*nix is loosing ground to Windows?" (which also wouldn't have been a huge surprise).

No news here.

Re:Uh huh

[Gilmoure]

Gnubfap is my next band name.

Re:Uh huh

[Sylak]

If it looks like a duck, swims like a duck, and quacks like a duck it isn't a duck unless it is branded a duck?

It's not a DUCK®, it's a Waterfowl That Attenuates Quacking Noises. For copyright reasons, of course.

Re:Uh huh

[Entropius]

I dunno, that trope is dying. Netcraft confirms it.

Re:Uh huh

[Desler]

XNU is a Mach 2.5 kernel with the BSD 4.3 userland layered on top. And it is not "reminiscent of Unix" it is certified Unix.

Re:Uh huh

[Mdk754]

Oracle is a flavor of SQL. So is MS SQL Server and PostgreSQL. I'm guessing that you meant MS SQL Server?

Oracle Linux indeed exists, basically RHEL. Oracle does a whole pile more than just databases.

Re:Uh huh

[Millennium]

Why do we even need to have this conversation? If it looks like a duck, swims like a duck, and quacks like a duck it isn't a duck unless it is branded a duck?

Welcome to the world of trademark law.

Re:Uh huh

[jitterman]

I'm at a mid-sized hospital, and we have AIX for one specific application that is about 3 minor revs behind. Our current plan upon upgrade is to move to the Linux release (there is also a Windows release, and while we have plenty of Windows servers, our systems team wants this app on Linux). Like you, I'm not sure how many people use AIX, but we are leaving it for Linux.

Re:Uh huh

[Mdk754]

Misinterpreted. My bad...

Re:Uh huh

[Anonymous Coward]

Posting anonymously for reasons that will become very obvious in the next sentence.

The company I work for has almost 3,000 SCO boxes. Well, they used to be boxes. Now they are VMs. We have legacy software that just won't fucking die that runs on SCO, and doesn't need to be updated until a whole multi-year big-dick-budget clusterfuck of a project replaces it and many other things with a new massive bloated system powered by Oracle.

We might be one of SCO's biggest remaining licensees, but I can't wait until we turn that shit off for good.

Re:Uh huh

[Bill, Shooter of Bul]

Trust the spell checker? I don't trust the idiot putting things into the spell checker. Remind me to fire the guy that types my posts.

Re:Uh huh

[DutchUncle]

People forget that once upon a time computers were expensive,and you didn't just go get more of them at the local market. They were designed for hot-swap, stability, and more stability. Of course if one were starting today one would buy a cluster of 64-bit machines with RAM all over the place and run rings around a 9370, but stability has its place.

Re:Uh huh

[chris200x9]

man, too bad it's pronounce guh-new otherwise it'd be nub fap.

Re:Uh huh

[RabidReindeer]

Unix has three definitions.

1. Legacy Unix, based on the AT&T Unix source code. eg: Solaris and AIX.

2. Certified Unix(tm). eg: Solaris, AIX, and OS X (apparently not included in their decline of Unix numbers).

3. Unix-like operating systems. eg: Linux, *BSD. (*BSD is also somewhat legacy in that AT&T incorporated their source code).

I don't think you can classify BSD as "Unix-like". My understanding that it is considered full-on Unix. And the basis for Solaris, no less.

Re:Uh huh

[Darinbob]

Most of the source is open, it's extremely compatible at the API level. The only thing that makes it not-Unix is that it's BSD derived instead of AT&T V7 derived. Most "real" Unix systems were closed source.

Re:Uh huh

[s.petry]

By this statement "windows cost a fraction of that in OS licenses and hardware" means that you have no clue how proprietary Unix works. Buying the hardware gives you license to the Unix. Hardware maintenance means you pay for maintenance on your Unix. If you jump into your time machine, long ago you used to have to pay extra for compilers and custom graphics drivers. Never did you have to pay for the Unix (no, SCO does not count!)

When you move to PC based hardware, especially when moving to Windows, you lose up time and get instability. SunOS, HP-UX, AIX, and Irix were always very stringent on hardware. This increased price of course, but damn if I don't have Sun E3s that still run from the 1980s. That, and they simply work without people putting their hands on them at all. Compared to Windows and weekly reboots? No thanks!

Linux adds some of the stability to the bulk hardware, but the bulk hardware is simply not as good. Making bulk means you lack the amount of QA put into proprietary hardware. Issues with driver compatibility are not as easily seen when you can mix and match what ever is cheapest. Our Dell and HP is not "bad" per-say, but not like it is with IBM P Series, Sun, or HP Unix.

8 machines to replace 1 does not save money. This is typical bean counter bullshit (fuzzy math) and not reality. A decent server is at least $2.5 K means you spent $20,000 on just hardware. Add OS licenses, AV protection, TS for remote admin work, Software do to _anything at all with_, additional network support, load balancing or HA to compensate for down time, and admin time, and you have nearly doubled the $25,000 you claim was too much for Unix.

Re:Uh huh

[Roger W Moore]

Many installations are realizing that they don't need that all that and less "capable" Linux and/or x86 systems are just fine - for many things.

Are they? Or are they just realizing that a cluster of redundant, possibly virtualized, machines is just as reliable even if each single machine is not? Two linux boxes with 99% uptime each running the same service redundantly is equivalent to one machine with 99.99% uptime but I bet the linux boxes are cheaper.

Re:Uh huh

[Darinbob]

However, some modern Linux distributions are probably much more reliable and professional than the "real" big box Unix systems were in 1990.

Re:Uh huh

[s.petry]

AC, what you've just said is one of the most insanely idiotic things I have ever heard. At no point in your rambling, incoherent response were you even close to anything that could be considered a rational thought. Everyone in this room is now dumber for having listened to it. I award you no points, and may God have mercy on your soul.

I think you should read some history regarding the BSD and System V split instead of making up stories.

Solaris and AIX were based on BSD, HP-UX and IRIX were based on System V. All 4 of those were certified Unix, all 4 of those were "Legacy" Unix operating systems.

Your third item is closer to the truth than your first two items, but notice that what you state inside the parenthesis contradicts your first and second statements.

Re:Uh huh

[mlts]

I wonder if the embedded kernel that is found in ESXi would be considered a UNIX. If that is the case, it definitely would help stats, although I seriously doubt a public stat gatherer would be able to find a hypervisor port in any installation that is installed with any sense of security at all.

Re:Uh huh

[swb]

Datagate?

Re:Uh huh

[BitwizeGHC]

Microsoft: "For you the day your servers all went down simultaneously was the most important day of your life, but for us it was Tuesday."

Re:Uh huh

[Darinbob]

There is more than the actual hardware and operating system. Where is the software that actually does the work? That is the most critical component, and very often the most expensive part. Replacing with a free unix doesn't help if that unix won't run the software.

Re:Uh huh

[jitterman]

Right - integration engine (Cloverleaf).

Re:Uh huh

[morgauxo]

Exactly! If GNU wasn't ESSENTIALLY Unix then RMS wouldn't have bothered to name it that! I mean, come on, Unix is right there in the name. Why name it that as opposed to giving it a whole new name of it's own that speaks to what it IS rather than what it ISN'T unless he knew that Unix was so much a part of it's identity that he had to try to define it as not Unix in it's name?

Re:Uh huh

[Guy Harris]

If it looks like a duck, swims like a duck, and quacks like a duck it isn't a duck unless it is branded a duck?

It's not a DUCK®, it's a Waterfowl That Attenuates Quacking Noises. For copyright reasons, of course.

Actually, trademark reasons. Oh, and nothing can be a "DUCK®", it can only be a "DUCK® waterfowl". Can't use a trademark as a noun, after all....

(Unfortunately, commenting on this - which I figured I'd do at some point - leaves me unable to moderate your article +10^100 Funny.)

Re:Uh huh

[Arker]

It's been a few years but I still remember when they tore out my sweet, reliable Netware server and replaced with Windows NT. It took many months of having these guys from Washington state flying in twice a week and huddling and working, and the thing was down more often than not, but eventually they did get this working. Using two machines, each with significantly more power than the one they replaced. And it was still noticeably slower and less reliable. But I am sure someone upstairs in the corporate structure got a nice kickback on the deal.

Re:Uh huh

[Guy Harris]

Especially if it's owned by Oracle...

Well, technically speaking, it costs Oracle money to use the name Unix, but they presumably pass that cost on to the customer. However, I suspect that cost is a small fraction of the total cost of the system, even if all you're running on it is Solaris, free-as-in-beer applications, and stuff you developed in-house.

Re:Uh huh

[rthille]

It's so closed-source that they make it available on git-hub: https://github.com/opensource-apple/xnu

Re:Uh huh

[fahrbot-bot]

However, some modern Linux distributions are probably much more reliable and professional than the "real" big box Unix systems were in 1990.

True, but, last time I checked, this is 2013. :-) And, while your sentiment may be valid for Linux, it's probably not for the hardware on which many/most Unix systems run.

Re:Uh huh

[unixisc]

Er, no. SunOS was based on the old BSD, not the current 4.4 based *BSDs. Solaris, OTOH, was based on SVR4. Also, since BSDs haven't undergone #2 above, they, like Linux, fall under category #3

Re:Uh huh

[MightyMartian]

I can see backup servers as one area. Considering most backup solutions for Windows have *nix plugins/agents, that makes some sense, although there are certainly *nix backup solutions just as comprehensive as any you'll get for Windows.

Re:Uh huh

[evilviper]

We might be one of SCO's biggest remaining licensees, but I can't wait until we turn that shit off for good.

McDonalds? I seem to remember them basically being the ones keeping SCO afloat.

I HATED being a SCO admin in this millennia. In the past they weren't as much of a joke, but they didn't keep-up AT ALL. Their USB support was a nightmare that kept crashing servers. Their boot-disk support was a nightmare I never entirely rangled into working, because floppy disks became too small and their boot CD support was too damn messy, and merging the right drivers just didn't work out.

SCO made sense for one reason... They were the first Unix on x86. Once FreeBSD / Linux was decent, they should have been dumped, EVERYWHERE in a second. Instead, far too many companies clung to the crufty old platform, and just kept raising the pain threshold.

Fortunately, I walked away from the only company I've seen still using it extensively, and haven't had to touch it since.

Re: Uh huh

[Guy Harris]

Yeah; I've been wondering what exactly they mean by UNIX here -- are we talking POSIX compliant OS (they almost all are these days), something based on BSD/AT&T code (BSD derivatives like OS X and FreeBSD, plus SVr3+ derivatives like HP:UX) are are we talking purely SVR 4+, and thereby mean SCO offerings when we say UNIX?

Well, "POSIX-compliant" can be split into "POSIX-compliant and the POSIX APIs are the core system APIs and the APIs on which a lot of the other system APIs are built" (UN\*Xes, including Linux and OS X) and "POSIX-compliant but the POSIX APIs are somewhat of an add-on and the core system APIs you're expected to use for most programming are different" (Windows with the Subsystem for UNIX or whatever it's called, z/OS's UNIX System Services, and the like).

But what people often mean by "Unix" when it's being contrasted with "Linux" is "commercial UNIXes in server rooms", which is closest to "something based on BSD/AT&T code", except that OS X doesn't count (not much in the way of computer-room systems runs it).

(There's also "OSes that have passed the Single UNIX Specification validation suite", which is closest to "POSIX-compliant OSes", and identical to it if you're talking about "POSIX" as meaning "Single UNIX Specification" and "compliant" as meaning "passed the validation suite" and not just "we sure intend this to match the SUS, but we haven't tested it" - Linux is more-or-less in the latter camp, although I'm not sure that they haven't picked some relatively minor places in the SUS to ignore).

Re:Uh huh

[amorsen]

Because instead of replacing 1 HP-UX server with 8 Windows servers, it could have been replaced by 1 Linux server.

Only if the software runs on Linux. Quite a few important commercial software packages started out on HP-UX and Solaris (and AIX if you were "lucky") and was then ported to Windows once people started using x86-based servers and workstations. Porting to Linux would obviously have been a lot easier, but it didn't happen until later or not at all.

Running such software under Linux either meant running Linux on RISC hardware and using a compatibility layer or running the Windows version in Wine. Neither was particularly appealing.

Re:Uh huh

[fahrbot-bot]

Many installations are realizing that they don't need that all that and less "capable" Linux and/or x86 systems are just fine - for many things.

Are they? Or are they just realizing that a cluster of redundant, possibly virtualized, machines is just as reliable even if each single machine is not? Two linux boxes with 99% uptime each running the same service redundantly is equivalent to one machine with 99.99% uptime but I bet the linux boxes are cheaper.

That all really depends on the requirements and usage. The HP systems I've used are very capable with hardware and software support for redundant and/or fail-over NICs and SCSI etc... as well as a large back plane and LOTS of RAM and CPUs (I once used a Unix system at NASA with 1024 processors.) A cluster of real/virtual systems is not always equivalent. For example, we once had a CPU fail on a T600 and the system simply deconfigured it and rebooted - note: there are / were Unix systems like Tandem Non-Stop on which *any* component could be replaced on-the-fly. Most Linux systems are not as capable in this respect - perhaps we are talking about different types of "reliability."

The upshot (and the point of the article) is that there are more choices and people are taking a harder look at what's actually required. In many cases, smaller, less capable/expensive Linux (or BSD) systems are adequate, but sometimes you really do need something more. It's not a dig against Linux, just that there are different tools for different jobs.

Re:Uh huh

[Guy Harris]

Does this factor in mac osx considering it's a unix based operating system.

No. To quote the article:

Unix, the core server operating system in enterprise networks for decades, now finds itself in a slow, inexorable decline. IDC predicts that Unix server revenue will slide from $10.2 billion in 2012 to $8.7 billion in 2017, and Gartner sees Unix market share slipping from 16% in 2012 to 9% in 2017.

I rather suspect OS X's share in the Unix server market is pretty small, making its share of the server market even smaller. Unix desktop market, pretty large, but that's a different matter.

Re:Uh huh

[rthille]

IIRC, SunOS was based on BSD, Solaris switched to System V (at the user-interface level anyway, not sure about the kernel and programming APIs)

Re:Uh huh

[evilviper]

It's just more diversified.

No, it's LESS diversified. Linux dominates. AIX is healthy. Solaris, and HP-UX are still alive but on life support. All the other formerly significant players are GONE.

Irix, Digital-Unix/Tru64, OpenVMS*, SCO OpenServer/Unixware/Xenix, SunOS, BSDi, Ultrix, QNX, NextStep, MINIX, etc.

They've all got so little market share as to be virtually nonexistent today.

HP-UX porting to Itanium put it on a roller-coaster ride to irrelevance. And Solaris won't be too far behind. AIX is kept afloat by IBM making damn good hardware, but it's been tough, and once the hardware falls behind, the OS will decline as well.

It's ALL going to Linux. I'm a fan of the BSDs myself, but Linux isn't a bad second-choice, and I look forward to having a lot more job openings ahead of me.

Re:Uh huh

[Rob Riggs]

Are they? Or are they just realizing that a cluster of redundant, possibly virtualized, machines is just as reliable even if each single machine is not? Two linux boxes with 99% uptime each running the same service redundantly is equivalent to one machine with 99.99% uptime but I bet the linux boxes are cheaper.

Exactly. Hardware and software architectures have changed a lot since 1973. Redundancy that used to be done in one piece of hardware -- "the server" or "the mainframe" -- is now handled by "the cluster". We still have expensive hardware when you look at the servers, network infrastructure, storage infrastructure, clustering and/or virtualization software and monitoring systems. But individually, we can take our pick of vendors for each of these components and that competition is what keeps the costs down.

Our vendors know that they cannot screw us (as, for example, Sun/Oracle does my previous employer) because they will very quickly find themselves with one less customer. There is healthy competition in the marketplace. And we work to avoid vendor lock-in.

We can also identify bottlenecks and selectively upgrade the pieces as needed. The cluster is organic in that regards. Our software runs on the same cluster it did years ago -- but all of the components have been upgraded numerous times, just like the cells in our bodies.

Re:Uh huh

[mooingyak]

7: Fewer IT admins needed (no "windows" guy/"UNIX" guy.)

Depends on the shop. Here we'd have to hire a windows guy, so it would be more admins.

Re:Uh huh

[unixisc]

Uh, no. Solaris was based on System V, a break from SunOS, which was based on BSD. I dunno about HP/UX or AIX, but DEC's Ultrix, and later, even their OSF/1 was based on BSD. In fact, you had battle lines in the 90s b/w Sun/USL on one side - UNIX International, and IBM/DEC/HP on the OSF side. That was before the 2 camps ultimately closed ranks.

Re:Uh huh

[evilviper]

Another reason why I say it's likely still around:
- It does pretty much everything they need it to.
- Unlike most x86 servers (but not unlike most high-end UNIX boxen) could probably take a direct 1-megaton nuclear detonation and keep going without so much as a dropped routine.
- You could swap out everything but the power cable without shutting it down (I think only a microcode and certain IPL patches would force any real cold downtime). Yes, I'm including RAM and processors.

"everything" is a subjective matter, changing from year to year. No doubt it was meeting expectations when purchased, but I'm sure they'd love to have MORE features that the box doesn't provide, and they just don't dare touch it due to hardware restrictions and age.

You can get super-heavy-duty x86 servers, but almost nobody wants to pay for them, so they're hard to find.

x86 servers can do the hot-swap thing, too, if you're so inclined. SMP processors in a redundant configuration. Memory module mirroring, multiple controllers, etc. It's a small reconfiguration to do it, but folks don't do it because they want cheaper x86 boxes with reasonable reliability, not more expensive and very reliable. And you can even swap out the power cables, one at a time, without down time. Reliability is typically handled with clustering instead of single-box redundancy.

Re:Uh huh

[ultrasawblade]

QNX isn't a UNIX derivative at all. I think it has some POSIX compatibility stuff thrown in but it's not a unix.

Re:Uh huh

[Penguinisto]

I'm sure it was a UCS system (Reynolds and Reynolds, today).

Bingo has been called (though it was called FDCS when I joined up with 'em). :)

And yes, UCS was the absolute king of fees, but I daresay that folks like Microsoft, Dell, HP, *ORACLE*, et al. have learned rather well from them.

Re:Uh huh

[Anonymous Coward]

Bring every system down once a month, whether they need it or not, so that when they do you have the structure in place. Much easier than doing it once or twice a year and nobody knows what they're supposed to do.

Re:Uh huh

[Just Some Guy]

And I strongly doubt that any of the remaining Unixes are more capable in any meaningful way. Each have their own cool features, to be sure (eg Sun's SAM), but Linux isn't exactly a tiny little featureless wisp of a thing. If you have a specific need that a proprietary Unix can serve better than Linux, that'd look pretty attractive. Most businesses don't have any especially difficult specials needs that someone else hasn't already solved better with Linux, though.

Re:Uh huh

[multisync]

Advise is a verb meaning "to give advice". Advice is the verb you couldn't find

Advice is a noun.

Re:Uh huh

[maccodemonkey]

It's "OS X" not "OS-X". It's also not a "variant". It is Unix.

a highly-modified closed-source BSD kernel is only reminiscent of unix.

For a closed source BSD kernel, it's remarkably easy to download from Apple's web site.
http://opensource.apple.com/source/xnu/xnu-2050.7.9/

Either that, or you'd be wrong.... But no. That's totally crazy!

Re:Uh huh

[lgw]

I liked those FDCS "mainframes" that were really PCs (well, PS/2s) with a 370 processor on a PCI card. Clever product, really.

Re:Uh huh

[Ungrounded Lightning]

Running such software [which had been ported from UNIX to Windows but not to Linux] under Linux either meant running Linux on RISC hardware and using a compatibility layer or running the Windows version in Wine. Neither was particularly appealing.

It was a server! What's wrong with running the apps under Wine? What were they running that Wine would have caused them problems?

Re:Uh huh

[serviscope_minor]

Except for, y'know, Ken Thompson's original distributions like V7. Which are about as unvaried as you can get.

That was ancient history.

The reason for the truly astonishing quantities of cruft in autoconf was to deal with the irritatingly huge variety between the major unicies at the time, IRIX, SunOS, AIX, HPUX, Linux, Solaris, some BSDs, OSF/1 and the various unix like horribless that cropped up in greater or lesser compatible versions on various other obscure systems.

Re:Uh huh

[bobstreo]

Unix has three definitions.

1. Legacy Unix, based on the AT&T Unix source code. eg: Solaris and AIX.

2. Certified Unix(tm). eg: Solaris, AIX, and OS X (apparently not included in their decline of Unix numbers).

3. Unix-like operating systems. eg: Linux, *BSD. (*BSD is also somewhat legacy in that AT&T incorporated their source code).

I don't think you can classify BSD as "Unix-like". My understanding that it is considered full-on Unix. And the basis for Solaris, no less.

4) Linux based phones (android)

Re:Uh huh

[VortexCortex]

Gnubfap is my next band name.

Sadly, I think the folks of 4chan have a strangle-hold on every incarnation of "noob fap" you can imagine...

Re:Uh huh

[VortexCortex]

POSIX != Unix. Gnu's Not Unix. It's POSIX though, and that's what counts.

Re:Uh huh

[Archangel Michael]

All of the redundancy you just mentioned, is now available in VMWARE, abstracted away so that the Hardware practically doesn't matter any longer (as long as it is x86/x64). A failed CPU on a VMWARE box would cause the machines to migrate over to a standby, and be up and running before anyone even knew. And I would get a notification and swap out the blade out of the chassis and be ready for the next "fail over".

As for the upgrading of hardware without ever downing a system, that is easy. VMWARE already handles that by migrading the virtualized box off the affected machine, you upgrade (add ram, CPU etc), reconfigure the VMWARE stack and migrate back. I can migrate the data stores with live data as well. I've replaced both Hardware(CPU, RAM) and Drives Systems(Equalogic to Nimble) while machines were running. This has nothing to do with "Linux" being able to handle it, since VMWARE does it without "Linux" (or Free BSD, Windows ...) ever knowing.

I won't install Server on bare metal ever again. It is more expensive to install on bare metal, but only if you value your time.

Re:Uh huh

[fahrbot-bot]

I pretty much agree, but there are either special or certain applications for which Unix - on larger / more expensive hardware - surpass many/most Linux and/or x86 solutions and that's what the article went into. Most of the capability rests with the hardware and the OS modifications to handle them. A LOT of people just don't need that level of reliability and are now turning to Linux (or other x86) solutions as they tend to be smaller and less expensive.

I've worked on just about everything from PC to Cray-2 systems using Linux and almost all versions of Unix (yes, I'm old) and they each have their applications. Many (most?) things can be done on smaller (x86) systems using Linux/BSD, but when you *really* need the big iron, usually nothing else will do. Sure, some of that could be done with a large cluster of smaller systems, but sometimes 100+ small systems is more hassle and expensive and actually less reliable than 1 large one - especially when you factor in the footprint, cooling, electricity, SA costs, etc... I imagine that sometimes it breaks the other way.

I also imagine that many people posting here have never actually worked on a large, expensive computer system - Unix or not ... Seriously, some of the systems on which I've worked were bigger than my car - and don't get me started on EMC disk arrays (we had to reinforce the floor *and* lay steel plates just to wheel ours in).

Re:Uh huh

[lister king of smeg]

If it looks like a duck, swims like a duck, and quacks like a duck it isn't a duck unless it is branded a duck?

It's not a DUCK®, it's a Waterfowl That Attenuates Quacking Noises. For copyright reasons, of course.

I can only imagine what sorts of names Debian will come up for this.

arctic quacking goose would be my guess knowing their love of the ice themed branding ala ice dove, snow weasle...

Re:Uh huh

[lister king of smeg]

the difference is that free BSD is a kernel a user land a bunch of programs and distribution it has vertical integration or cathedral style development. Linux on the other hand is just kernel that many people build stuff around remix and make a billion different distros that all collaborate. It is at the core a perfect example bazzar style development everyone doing there own thing but making it all work together.

Re:Uh huh

[Forever Wondering]

BSD is AT&T [pre v7] Unix derived. http://en.wikipedia.org/wiki/Berkeley_Software_Distribution

I was doing Unix device driver work in startups since 1981 [v7, system III, system V] and we always referred to them with "Unix" attached (e.g. "Unix System III" or "Unix v7" or "v7 Unix"). Likewise, BSD was "BSD Unix"

POSIX was born out of a desire to produce standards that allowed Unix-compatible systems to be created that were not AT&T derived [i.e. eliminate the AT&T licensing stranglehold]. AT&T could copyright Unix source/binary distros. But, it couldn't copyright the API's [Note to Oracle ;-)]

The non-Unix derived systems that truly aspired to look, act, feel like Unix [and be free] were: Minix, Linux, Mach. No doubt there are others that didn't quite make it to completion. Obviously, Linux took the crown [I loaded my first distro in 1993]. Though not an OS, the other piece of the puzzle is GNU/FSF providing "clean room" reimplementations of the standard POSIX (nee Unix) utilities.

I stopped using the term "Unix" a long time ago. Now, I usually say Linux, *BSD, Solaris, or "POSIX-compliant OS", etc.

Re:Uh huh

[unixisc]

I'd imagine that the main driving factor would be going from a proprietary, HP-only architecture - be it PA-RISC or Itanium - to the x86, which is something plenty of companies make - Dell, IBM, et al. PA-RISC is dead, and on the Itanium i.e. HP's Integrity servers, the only options, aside from HP/UX, are Debian Linux and FreeBSD. So if one wanted to reduce one's dependence on HP, given everything that's happened w/ the Alpha and OpenVMS, it makes sense to migrate away from the Itanic. And if one is doing that, one could stay w/ Debian or FBSD, but one would definitely have to leave HP/UX.

I agree w/ you that it's an expensive process. But w/ companies abandoning platforms or making them more expensive to continue running, the writing is pretty much on the wall for customers tied up w/ such proprietary solutions. Just look at OpenVMS customers, who were pretty much raped when they had to buy Integrity servers to replace AlphaServers running their software. Since DEC is no longer around, their investment is no longer protected, so as HP customers, they'd do well to look at alternatives. The best options out there are Linux and xBSD, and they might as well go that route.

On the question of why Windows, chances are that once they make a decision to go to x86, the platform that supports the widest choice of software is Windows, and so as an application server, it's an automatic choice as a result.

Re:Uh huh

[jambox]

IBM still sells a lot of AIX for running Tivoli and stuff like that.

Re:Uh huh

[MightyMartian]

One of the (few remaining) movie rental stores where I live is still running it for the rental/POS software. He updated a few years ago from the Wyse terminals to PCs running terminal emulation software, but that's about it. On the one hand, wow that's archaic, but on the other hand, I guess as long as it works and the hardware keeps humming, there's not much reason to switch. Still, the day will come when the platters will seize and he'll be left trying to find tech support for it. I have no idea whether you can get an old install running in a VM or not, which would be the logical way to go.

Re:Uh huh

[Capt.DrumkenBum]

Damn it! Now I have to change my password.

Re:Uh huh

[unixisc]

But neither Oracle nor Sun used the name UNIX, either then, nor now. Back in the day when that trademark cost money, every vendor called it something other than UNIX - SunOS, Solaris, Iris, AIX, HP/UX, SCO OSE, Dynix, CLIX, et al. The only company I can recall calling it UNIX was Novell's UnixWare, but then Novell got USL from AT&T. So it would cost Oracle nothing. Oracle has Solaris aimed at legacy SPARC servers, while for x86 servers, it uses RHEL rebranded as OEL.

Re:Uh huh

[Tough Love]

QNX is a Unix-like operating system, just as Linux is. It is a Unix derivative or not depending on how you define "derivative". Certainly not a derivative in the copyright sense, and certainly a derivative in the API sense.

Re:Uh huh

[InfiniteLoopCounter]

Doh. I realized just after clicking submit that Linux is actually the kernel in GNU/Linux and that my post could look quite dumb if I didn't clarify.

Re:Uh huh

[s.petry]

Solaris originally added some System V enhancements to SunOS, but did not change it from BSD. As Solaris progressed more and more System V was implemented and added but it was never truly System V compliant.

The easiest way to tell by looking for BSD or System V is to review the init system. System V has an /etc/rc.config.d, /etc/init.d, and /etc/rc?.d directories where BSD uses /etc/rc.* boot scripts. Solaris took a hybrid approach and went to an /etc/init.d with /etc/rc?.d directories without the /etc/rc.config.d structure. Of course that was dismantled for the proprietary init replacment service in Solaris 10. The original Solaris implementation of the System V init had copies of boot scripts in /etc/rc?.d directories instead of links. Some of these were still around in Solaris 8, but cleaned up fully in Solaris 9

Re:Uh huh

[MightyMartian]

Heh. I remember Coherent. I had the 386 edition, which didn't support virtual memory, which heavily limited its usefulness. I had wanted to update from the old Tandy 6000 running Xenix (with two whopping 20mb MFM drives). At the end of the day, I abandoned the efforts, and the old Xenix box ran the show until the business closed. I inherited that old Xenix machine with the idea of running a BBS but by then one of the hard drives made a very high pitched squeeling sound (though it still booted), and I ended up giving it to a friend of mine.

I still had the Coherent floppies until about five or six years ago when I finally turfed them. It was a nice toy made utterly irrelevant by Slackware.

Re:Uh huh

[Tough Love]

The article is mainly talking about the Unix versions like HP-UX, Solaris, etc...

The article is mainly blowing smoke out its ass.

Re:Uh huh

[Darinbob]

There were the unix wars awhile back, and I knew a few people at AT&T related companies who would frown whenever BSD was mentioned. The definitions of what is true Unix has changed over the years.

Basically since V6/V7 things have branched often but subsequent merging was less common; sometimes only the ideas merged rather than the actual code. There may be a common ancestor but the amount of that common ancestor that is still present is extremely small.

Re:Uh huh

[MightyYar]

It's powered by the users own sense of self importance.

Did the BMW patents expire already?

Re:Uh huh

[riverat1]

And the HP UX server probably had one Sys Admin who only needed to work on it part time while the Windows servers probably require at least 2 of them. I'm Sys Admin for a Solaris system running our ERP system and I doubt I spend even 10% of my time administering it. We have 3 guys working on the Windows network and sometimes they ask me to help.

Re:Uh huh

[MightyYar]

My mom's real estate office still runs it. I was helping her restart Apache and I was floored she logged me in and it said SCO. Rooting around (pun intended), it seemed pretty primitive - it wasn't even set up to recycle the log files, and they were HUGE.

Re:Uh huh

[Myopic]

You would follow bad advice because someone's insightful rhetorical question isn't phrased in the form of a detailed argument? Damn. That's foolish. I hope you are not serious.

Re:Uh huh

[sk999]

One should probably classify BSD as "Legacy Unix" as well, since it is derived from 32V (which is AT&T Unix version 7 ported to a VAX). For several years, BSD (at Berkeley) and System V (at AT&T) were developed in parallel (part of the "Unix Wars"). Thus, Ultrix and SunOS were BSD based, while AIX and HP/UX were System V based. There was sort of a merger of System V and BSD with the release of SVR4, which begat Solaris, among others, and which soon led to Certified Unix(tm).

Re:Uh huh

[bertok]

Ok, I'll bite.

I'm a Windows admin, but I just went to a training course to learn about a high-end enterprise product that runs on top of Linux. I've dabbled with Linux-based stuff before (proxies, VMware, ESX, etc...), so it's not exactly new territory, but I figured it's 2013, it'll be interesting to get a glimpse into the current state of the "Linux Enterprise" world.

My experience was this:

-- You still need to patch, or install 140+ dependencies to install one application. Same difference.
-- You still need to reboot. A lot. More than I thought. I suspect that it is possible to avoid most of them though by judiciously restarting services, but the effort is much higher and the outage level is practically the same, so what's the benefit, really?
-- Things that really ought to be automatic, aren't. I spent a good 50% of the lab doing really fiddly things like cut & pasting iptables rules to open firewall ports. The installer really should have just done that for me.
-- Binding services together and just generally getting things to start up and talk required an awful lot of error prone manual labour. The lab guide was liberally sprinkled with warnings and "do not forget this or else" sections. Lots of "go to this unrelated seeming file, and flip this setting... because.. just do it or nothing will work."
-- I love the disclaimer in the training guide: "Linux configuration scripts do not tolerate typos, are case sensitive, and are not possible to validate before running the associated service." Fun stuff. I can't wait to diagnose random single-character problems in 10 kilobyte files when the only error is that one of a dozen services barfed when started.
-- Wow, the 70s called and wanted their limitations back: spaces in file names? You're risking random failures! Case-insensitive user names? Nope. Unicode text? Hah! IPv6? In theory, not in practice. GUI config wizards? Nope. Text-based config wizards? Not many of those either. Want to make a configuration change to a service without having to stop & start it? You're dreaming! An editor more user friendly than vi? Eat some cement and harden up princess!
-- I love the undecipherable command-line wizardry. I'm not an idiot, but how-the-fuck would I know what "-e" does on some random command? There is just no way without trawling through man pages using a command-line reader with no mouse support and keyboard shortcuts I don't know. Compare this to a sample PowerShell pipeline "Get-Process -Name 'n*' | sort -Descending PagedMemorySize". You'd have a hard time finding an IT engineer that can't figure out what that does.

I keep hearing about the supposed efficiency advantage of Linux, but I just don't see it. Given a Hypervisor, PowerShell, and Group Policy, Windows administration a piece of cake in comparison.

Re:Uh huh

[mellon]

Sure. But I know lots of cases where a large windows shop has Linux because they have to. The cases I'm aware of where a large Linux shop has a copy of Windows to run Exchange, that's the _only_ windows machine in the shop, and it does nothing else, and it isn't even an outward-facing SMTP server. If you have data to the contrary, by all means trot it out.

Re:Uh huh

[dbIII]

From the other end supercomputing clusters made of many networked servers also get described as a single system.

Re:Uh huh

[dbIII]

Don't worry, you are in the same ignorant boat as most who have insisted on gnu/linux or the earlier renaming attempt of LiGnuX, and things like android or ulinux which have no gnu code at all render the entire silly renaming irrelevant. The gnu OS is called hurd.

Re:Uh huh

[Guy Harris]

But neither Oracle nor Sun used the name UNIX, either then

Well, actually, Sun's OS originally announced itself in the boot message as "Sun UNIX 4.2BSD Version {version number}", or something such as that, until AT&T got cranky; "SunOS" first appeared in the boot message in, as I remember, 4.0 (at which point it was also more "4.3BSD" than "4.2BSD").

nor now.

Define "used"/"uses". They don't use it in the OS's brand name, but they sure use it on, for example, the Solaris 11 overview page ("Brings the reliability, security and scalability of the #1 UNIX OS to the enterprise cloud"). Apple doesn't use "UNIX" in the name of their OS, either, but they used it in various advertising materials, e.g. "sends all other Unix boxes to /dev/null", and The Open Group told them they had to certify (Mac) OS X in order to use the trademark.

Back in the day when that trademark cost money

It still costs money:

Annual fees apply, which are referenced by the Trademark License Agreement:

• License fee for the TMLA to remain in place
• Product registration fees
• Program fees (royalties)

Separate fees apply for the test suites.

every vendor called it something other than UNIX - SunOS, Solaris, Iris, AIX, HP/UX, SCO OSE, Dynix, CLIX, et al. The only company I can recall calling it UNIX was Novell's UnixWare, but then Novell got USL from AT&T.

There was also Digital UNIX, which was the new name for DEC OSF/1 after it passed the test suite for SUS conformance. (It was later renamed Tru64 UNIX when there was no longer a "Digital Equipment Corporation".)

So it would cost Oracle nothing.

...as long as they stop calling Solaris "the #1 UNIX OS" (or anything else with "UNIX" in it).

Re: Uh huh

[Guy Harris]

That, and the fact that it runs on the MACH microkernel, which is completely non UNIX, but supports a Posix API layer.

The "POSIX API layer" is 4.4-Lite-derived kernel code plopped atop Mach's platform support, task/thread, and VM code; it's not part of the Mach layer. The Mach code knows nothing about networking or file systems; that's all BSD code (or Apple code). (There's also I/O Kit, which is neither BSD nor Mach, and is what's used for drivers that actually talk to devices, rather than "pseudo-drivers" such as the pseudo-tty driver and the BPF driver. The osfmk directory of the XNU source is the Mach code, the bsd directory is the 4.4-Lite-derived code, and the iokit directory is the I/O Kit code.)

Re:Uh huh

[Billly Gates]

Because Unix is architecturally closer to Linux than Windows, would be the obvious answer. But seeing as there aren't many details given about the role of the servers*, it would be wise to reserve judgement in the particular case, even if it makes little sense as general advise.

*Often times under further questioning admins doing similar switches, the irrefutable answer becomes " exchange integration" or something of the like. Someone offering advise can't very well redesign the entire IT operations over coffee for a company he doesn't work for.

You and everyone forgot the MS Fud and propaganda of the late 1990s with the unstopable Microsoft! Everyone was eager to dump their superior products like Novel NDS for active directory and Netscape for IE 6 because the view was if you did not your customers would not buy your product as by now no non-MS product would ever exist again in the enterprise.

MS made fake Posix and made deals with Digital, SGI, and others to port Unix and VMS apis with win32 ones and once ported keep you there on NT.

Today we laugh at this but the damage was done. How many CAD programs do you see on Linux or Unix? None. The fud worked and a few artist programs like Renderman and a maya port are all that remain.

Since dumping Unix 5 versions earlier 12 years ago it would be a pain in the ass architecturally to port all the win32 code back to Linux. There is some that stayed Unix and most of it has been ported to Redhat and SuSE Enterprise, but workstation wise it is pretty much all gone now.

Sadly lesson 101 in economics and business is Money Talks and Shit walks. Windows one because it was cheaper just like H1b1s won most of the IT jobs today for the reason and why cloud is winning over as well. $$$$

Re:Uh huh

[Billly Gates]

I can give the counter-arguments against using Windows:

* You're guaranteed to suffer every month for maintenance (Patch Tuesday), and require multiple machines not just for capacity-matching, but for redundancy if you want anywhere near the same uptime. In spite of an MCSE/MCSA being cheaper, one competent UNIX admin can maintain 3x the machine count than an MCSE/MCSA can - unless you feel like springing for a lot of pricey add-ons/upsells to keep admin FTE headcount down (e.g. automation via SCOM,SCCM and etc). It doesn't take too much for that SA contract cost to match or exceed the HP one, especially if the Microsoft products have the word "Enterprise" in the product title/license.

* All that aside, I haven't even touched on increased space, power consumption, cooling/HVAC, and etc... the costs scale up almost exponentially in larger installations.

The real catch that geeks and IT managers go crazy over at the end of the day.

Can you verify this with Excel? On paper, they get the same service for 1/4 the price! The bean counters go crazy and now the standard is set at the cheapest offer and you need a real business case analysis to not get Windows or some cloud? That my friend is hard to do unless you already switched. THen afterwards the same bean counter tards will whine about sunken costs and would just add it back to your cost savings to switch back etc.

Re:Uh huh

[Guy Harris]

Unix has three definitions.

1. Legacy Unix, based on the AT&T Unix source code. eg: Solaris and AIX.

2. Certified Unix(tm). eg: Solaris, AIX, and OS X (apparently not included in their decline of Unix numbers).

3. Unix-like operating systems. eg: Linux, *BSD. (*BSD is also somewhat legacy in that AT&T incorporated their source code).

I don't think you can classify BSD as "Unix-like". My understanding that it is considered full-on Unix. And the basis for Solaris, no less.

4) Linux based phones (android)

And 4) differs from 3) how? As far as I know, the low-level parts of Android are the Linux kernel and the Bionic libc, which give you a Unix-like API; Dalvik sits atop those (and possibly other things).

Re:Uh huh

[Billly Gates]

McDonalds and most fast food joints used too or still use it.

See the monochrome monitors at the checkout?

More than likey it is SCO Xenix/OpenServer. It became popular before NT as it was dirt cheap at only $1200 for a real OS that was multitasking and could run on cheap wintel server hardware. Sun and SGI cost $10,000 and up! That adds up when you have 200 stores or restaurants! For a better OS that supported terminals SCO Unix was the cheapest thing around and most practical if you were not running servers or high end number crunching.

Some companies like Home Depot and Autozone switched to Linux and use SCO emulation software or just got reported but many of these systems just work. I remember 10 years ago most HP proliants were SCO certified as well as RedHat enterprise for this reason. Especially if the software is +25 years old but works fine so why change etc?

I would think today it is run as a guest in a VM in 2013 as I wonder about drives and things like that. Does VSphere support greenscreen terminals and serial ports? Many customers now are stuck with ancient XP systems that will age like SCO as well.

Re:Uh huh

[Billly Gates]

All I can say is "I'm glad I don't work there." I can only imagine the nightmares they must be going through moving from HP-UX to a Windows Server cluster.

The very high end servers can cost up to $100k easily if you want hot swapable things and partitioning and so on. If the server was very old as many HP-UX are I can see 8 crappy $1400 servers costing half the price and perhaps a cheaper software product running on IIS/IE 6 (all seem to use this POS at work) could be done cheaper. Especially if the other software package was a green screen terminal one.

I am not saying it is technically the best solution or the wisest, but cost wise the bean counters liked this ... until XP EOL happened and that setup now needs to go too :-)

Re:Uh huh

[smash]

And this is part of the reason it is happening. Newsflash: It's not 1996 any more.

A few things have happened: RAM is very cheap. Windows isn't anywhere near as crap any more (yes it has its warts). The free unix world is still faffing about with per-application specific configuration file formats (sendmail.cf or sendmail.mc vs XML, for example) and wasting time reinventing the wheel every 18 months. There is massive resistance to change in the Unix world for things that need to change (e.g., unified configuration file format that can be understood and processed reliably by a single parser), but no resistance to change for things that break end user workflow for little benefit (Gnome? KDE?)

OS X gets a lot of lower level stuff right, but the free desktop world is too busy skinning whatever window manager is flavour of the month to look like Aqua to notice.

Re:Uh huh

[smash]

vSphere supports fault tolerance between hardware platforms with VMware FT. It runs two copies of the VM in lockstep. So yes, you can do complete, zero data loss resiliency in vSphere with two independent, non-fault tolerant servers, so long as there is resiliency in the network and SAN.

Re:Uh huh

[luis_a_espinal]

Because instead of replacing 1 HP-UX server with 8 Windows servers, it could have been replaced by 1 Linux server.

That's a bullshit script kiddie argument. Requirements (technical and business) dictate what you can or should do. Some requirements would call for a bunch of linux servers. Others for Windows servers.

For many situations, Windows servers do fine, and by the way, I've work predominantly in the *NIX world (web/enterprise/embedded) for most of my career. I'm hardly a Windows promoter, but I don't engage is mindless OS fanboyism, either. I would always prefer Linux over anything else in the absence of specific requirements. But I would never spew RAARR-RAAARH MAH LINUX BEATS YOUR WINDOOZE kind of bs. That's just useless silliness.

Re:Uh huh

[smash]

Not sure if trolling or just retarded...

You're joking right? Try calling your enterprise application vendor with a problem, have them log in to remotely diagnose and find their app running under WINE.

WINE is a good thing. I run it myself - but to expect support from your vendor on a platform they have no control over and do not list as supported means one thing. YOU become application support. YOU are now responsible for every quirk, annoyance or major catasrophe to befall your application and can expect pretty much zero help from the vendor.

Re:Uh huh

[smash]

Pretty much. It's not 1995 any more kids - microsoft have lifted their game significantly. And I'll tell you now that any enterprise running vSphere (suspect: virtually all of them in some form) is running at least a few Windows servers to run vCenter.

Re:Uh huh

[smash]

Windows: You're doing it wrong.

I have had zero unscheduled reboots in the past 5 years for my Windows servers (all running on vSphere). Unix is not immune to kernel patching or operating system upgrades either.

Re:Uh huh

[manu0601]

I modded you down as troll by mistake. I am posting this just to undo moderation.

Re:Uh huh

[manu0601]

But can you build it and use it in your OS X system?

Re:Uh huh

[manu0601]

You are right if you use definitions 1 or 3, and wrong if you use definition 2.

Re:Uh huh

[smash]

Yup, and you can avoid a lot of Windows reboots too by re-starting services. Agreed with the above post pretty much 100%. I'm a Unix (Solaris, Linux, FreeBSD) admin since 1996, and a Windows admin since 2000 (we run both).

Windows does infuriate me from time to time, but all of the above is pretty on the money. Other than spaces in filenames, you can use them, you just need to escape them with \ or use quotes around them.

Not to say that Windows fares too much better if you do things properly with regards to software installation and security; but point being: if administered properly (adequate time invested), Windows is not a fischer price OS. Linux isn't some magical pancea either.

Re:Uh huh

[smash]

Less useful? Serious question: Have you actually run any windows servers in production between Windows 2000 and 2008 R2?

Re:Uh huh

[Guy Harris]

Solaris originally added some System V enhancements to SunOS, but did not change it from BSD.

You're thinking of SunOS 3.2-3.5 and 4.x; the first SVisms were added in 3.2, and a bunch more were added in 4.0 (and RFS was added in, I think, 4.1). The "Solaris" name wasn't used until SunOS 4.1.1 (Solaris 1 = SunOS 4.1.1 + the corresponding versions of OpenWindows and maybe the desktop applications) and SunOS 5.0 (Solaris 2 = SunOS 5.0 + the corresponding versions of OpenWIndows and desktop applications).

As Solaris progressed more and more System V was implemented and added

SunOS 5.0 was Sun's version of System V Release 4. SVR4 was the result of AT&T and Sun combining the SV code base with the Sun code base, with a bunch of things (such as the VM subsystem) picked up from SunOS 4 with some changes (one of the most important VM subsystem changes was the renaming of the routine to look for an unmapped region in the address space from as_hole() to as_gap(); yes, I'm serious, do a Web search for them).

but it was never truly System V compliant.

I.e., it didn't conform to the System V Interface Definition? (That's the only form of "System V compliance" there is.) If so, could you provide a citation for that claim?

Re:Uh huh

[Ungrounded Lightning]

Why add another layer instead of running it under the native OS?

Because:
  - you don't have a port of the service you want to run to the native OS you want to use,
  - the native OS you DO have a port to is a P.I.T.B. from a reliability, cost, security, and maintenance standpoint, but
  - you DO have an API adaptation layer (i.e. the application's code runs "on the iron" so there's no emulation penalty there) that runs on the OS you want to use and does a darned good job of supporting the parts of the API used by the service's code.

Re:Uh huh

[smash]

Solaris x86 disagrees with you.

Re:Uh huh

[smash]

^ that. we just ditched a SCO box a few years ago because long ago our ERP software was installed on it (I'm talking like... 1997?) and the vendor ran it as one of their supported platforms. We migrated to RHEL about 5 years ago. I'm sure there are plenty of others who are running installs of it who haven't migrated yet. The only times it got upgraded were when the hardware was due for refresh.

Re:Uh huh

[InfiniteLoopCounter]

Was an attempt at humor. But executed poorly. I blame it on trying to figure the difference between Unix, *nix, and Linux on an empty stomach and attempting to make sense of the article summary.

Re:Uh huh

[maccodemonkey]

But can you build it and use it in your OS X system?

Yes.

There are even delta patches posted to match it to your current OS X system.

Re:Uh huh

[maccodemonkey]

But can you build it and use it in your OS X system?

Yes.

There are even delta patches posted to match it to your current OS X system.

I should add that you can build it and run it even WITHOUT an OS X system. People used to run the core (called Darwin) on Intel boxes well before Apple supported Intel. There were even distributions with XWindows thrown on top.

Re:Uh huh

[smash]

Well with VMs you no longer need to worry about hardware support - and if they do the job, and can be adequately protected from exploitation... why change? Yes, its a dead platform, but has been for some time. Migrating to something else will be a huge project and in terms of return vs spend I doubt you'll get much out of it?

Re:Uh huh

[hantms]

rest in peace

E: Unable to lock the administration directory, are you root?

sudo rest in peace. Bitch.

Re:Uh huh

[g01d4]

In five years PCs will be rare, because 99% of people who would use a PC in 2010 use their phone instead.

I can't see that happening in an office unless their 'phone' plugs into some kind of docking station and runs some kind of thin client.

Re:Uh huh

[Myopic]

I'm a programmer not a sysadmin, but yeah I've had to program against and deploy software on Win servers in the last decade. Actually, you know, yeah I did do some sysadmin tasks with some of those servers when I was doing the deploys and configuring them afterwards. It was such a pain having to use the GUI tools to change settings instead of copying config files into an etc dir. It was such a pain to find the "restart" button instead of just cycling a daemon.

Many of my complaints are answered by Windows apologists with recommendations to install things that should have been baked into Windows twenty years ago, such as PowerShell or Cygwin. In my opinion if you have to install Unix utilities to make your Windows server useful, then that is an indication you should be using Unix in the first place.

I won't defend the "less useful" comment but I will defend "useless". I don't claim to be open minded. I'm closed-minded to Windows so it's not like a well-reasoned and informed opinion could sway me, so if you want you can just roll your eyes and dismiss me.

Re:Uh huh

[module0000]

Or you can use Xen or KVM, and pay $0 for faster benchmarks and a larger set of features. This reads like a VMWare "certified" something justifying his job or ideal job.

Re:Uh huh

[Provocateur]

More importantly, I wonder if your crowd pronounces it as "E-sexy" or you decided to call it E -S -X-aye.

Re:Uh huh

[UltraZelda64]

That's so funny, so well said, yet so true.

Re:Uh huh

[UltraZelda64]

I'm pretty sure the "Fisher Price" jokes started with Windows XP as a result of its default plastic toy-like theme which included a blue taskbar and window borders/title bars with red "close window" buttons and a green Start button. And that was... 2001? So I don't know where you get 1995 from.

Re:Uh huh

[UltraZelda64]

Isn't that their practiced business model? Undercut everyone, then raise the prices to whatever the "market will bear"?

Hey, isn't that the kind of business of an abusive convicted monopoly? Oh, wait...

Re:Uh huh

[fahrbot-bot]

vSphere supports fault tolerance between hardware platforms with VMware FT. It runs two copies of the VM in lockstep. So yes, you can do complete, zero data loss resiliency in vSphere with two independent, non-fault tolerant servers, so long as there is resiliency in the network and SAN.

Great, so two virtual systems take the place of real one. Not sure that's a bonus... I had 42 Oracle databases running on my HP 9000/T600 with redundant everything, including several redundant F/W LD SCSI channels to an AutoRAID unit and SAN; what's the corresponding dual VMware FT footprint, performance and cost for that?

You can spend your time trying to make one tool work like another or understand the strengths/weakness of each and utilize them each appropriately.

Re:Uh huh

[Penguinisto]

The minis? Yep... they were pretty nifty.

The one I referred to originally was the real deal, refrigerator-sized and IBM-branded.

Re:Uh huh

[fahrbot-bot]

The article is mainly talking about the Unix versions like HP-UX, Solaris, etc...

The article is mainly blowing smoke out its ass.

I'm not sure how stating the point that, though declining, there still exists a market for hardware running Unix for which Linux and x86 systems may not be adequate is blowing any smoke anywhere. Linux/x86 systems are nice and great for many/most situations, but certainly not *all*.

Re:Uh huh

[formfeed]

>>>W
>>>>H
>>>>>O
>>>>>>O
>>>>>>>S
>>>>>>>>>>>>>>>>>>>>>>>>>>>H

Why do you use DOS symbols when the debate is about Unix?

Re:Uh huh

[amirulbahr]

Well it's Woosh-like but not WOOSH®

Re:Uh huh

[multimediavt]

Oh fuck me, I'll bite.

Hmmm, where to even start with WIndows?

Ok, how about having to spend $1500 plus in training to find the damn shit that Microsoft decided to once again obfuscate behind 16 tabs and buttons in lame, third-grade designed dialog boxes just to set one service preference? I can edit a text file that's been formatted the same for ten years (or more) and setup most common Linux services.

How about, every time a new version of Windows Server hits you have to wonder how many new versions of other software you have to buy just to keep your system secure, and/or running? FOSS wins.

Ever heard of a man page? I know Windows admins also have to do their homework before they can weed their way through the 16 levels of dialog box Hell to get to deploying a...well anything!

Good package management handles dependencies without a problem. At least it doesn't break something every few Tuesdays.

Yep, you still need to reboot, if you don't know how to build your own kernel and modularize services so you can reboot less. Plus, shell scripts are your friend for those little annoying service restarts.

Things aren't automatic in Linux because Linux admins don't like opening up gaping security holes, or maybe just want a little more control over what is going on. Different strokes.

You see command line syntax as "limitations", *nix admins see it as not effing up while on the command line. One wrong space between characters can be bad, bad, bad, bad, bad. Just like turning on one wrong checkbox six levels deep in WIndows dialog boxes. What's the difference? A pain in the ass is a pain in the ass.

"I love the undecipherable command-line wizardry. I'm not an idiot." You contradicted yourself there....see man pages. Also, Google is your friend...see homework.

Of course, the point of my rebuttal is to remind you that we all started at zero no matter what OS we chose to use to serve whatever we need to serve. It's when we forget this and forget the merits of all OSes that we truly have lowered the IQ of the room. Right tool, right job. If you want to be a master mechanic you can't just use screwdrivers. 'Nuf said.

P.S. "Error prone manual labor", from a WIndows admin....That's funny.

Re:Uh huh

[multimediavt]

I'm just trying to sort out what this server would have actually been running that one could simply go "Oh well, we're going to Windows now."

Mine sweeper.

Re:Uh huh

[Penguinisto]

Ok, I'll bite.

Cool - allow me to dispel some bits:

-- You still need to patch, or install 140+ dependencies to install one application. Same difference.

RedHat-based: yum -y install (whatever).
Debian-based: apt-get install (whatever).

Best part is, you can put the whole list in at once if you want in one go.

Windows world? It's gotten better (adding roles and features), but it's still reboot hell at times.

-- You still need to reboot. A lot. More than I thought.

Only if you're patching the kernel or glibc. Protip: If you don't want to bother plowing through reboots just because you don't like restarting services, learn to use the telinit command. ;)

-- Things that really ought to be automatic, aren't. I spent a good 50% of the lab doing really fiddly things like cut & pasting iptables rules to open firewall ports. The installer really should have just done that for me.

Clue: Windows application installers don't usually mod the firewall either (unless we're talking MSFT-branded ones, e.g. SQL Server.)

Binding services together and just generally getting things to start up and talk required an awful lot of error prone manual labour.

'fraid you'll have to be more specific than what you posted, because you're not making sense here. What exactly do you mean by "binding services together"?

I love the disclaimer in the training guide: "Linux configuration scripts do not tolerate typos, are case sensitive, and are not possible to validate before running the associated service." Fun stuff. I can't wait to diagnose random single-character problems in 10 kilobyte files when the only error is that one of a dozen services barfed when started.

You mean like when a seemingly random multi-MB .aspx file has a single typo in it, causing IIS to not run, with only a cryptic (and definitely non-intuitive) generic blurb buried somewhere deep in Event Viewer? Or how about a typo in some config file (lurking under a dozen nested folders) causes SSRS to fall over?

Or are you just arsed over case-sensitivity? ;)

Wow, the 70s called and wanted their limitations back: spaces in file names?

Just like in Powershell, you may want to learn to use quotes or escape chars... and with MSFT moving away from the GUI at the server level, you'd better get used to it.

IPv6? In theory, not in practice.

Now I know you're trolling, or are completely ignorant.

GUI config wizards? Nope.

Clue: Wizards are going away in Windows too. Better brush up on Powershell. ;)

Want to make a configuration change to a service without having to stop & start it? You're dreaming!

That's actually an advantage: one can change all kinds of differing network info (IP addy, DNS, NIS/windbind, etc) over ssh (think RDP for grown men ;) ), and not have the machine blink out until you're ready to commit those changes.

An editor more user friendly than vi?

Try EMACS (I kid, I kid...) In all seriousness, there's a zillion of them, with varying opinions. vi on the other hand has the advantage of being universal. I can use it in Linux, HPUX, Solaris, AIX, FreeBSD, etc.

A lot more portable and capable than Notepad, dontcha think?

-- I love the undecipherable command-line wizardry. I'm not an idiot, but how-the-fuck would I know what "-e" does on some random command?

man {so

Re:Uh huh

[Penguinisto]

I don't have to schedule reboots each month. I have *nix VMs (okay, LPARs) that I haven't had to reboot (scheduled or otherwise) for at least six months.

Re:Uh huh

[Guy Harris]

No, Linux not a UNIX It doesn't pass the POSIX standards to get the UNIX trademark, and it never will (due to some significant architectural difference in the kernel and the resulting libc.)

Which specific parts of the Single UNIX Standard cannot possibly be implemented by Linux+glibc due to those architectural differences? Be specific, please.

Re:Uh huh

[deek]

Agreed that Group Policy is wonderful for simple administration of multiple windows machines. You can get something similar to work with Linux, for example using Puppet, but it does take some work to tune it nicely. Powershell is a god-send, and it's about time that Microsoft started to embrace command line administration, although it still needs some tweaking to be as convenient as a unix shell.

There are a few confusing things you mentioned, though.

- 140+ dependencies to install one app? I assume this was done automatically, instead of manually. If not, what the hell distribution were you using?! If so, then I consider this an advantage, because it is a fine-grained system of separate packages, instead of one monolith installer which includes bundled libraries that end up being cruft over time.
- Reboot a lot? I'm still wondering what distribution you're using. I've upgraded whole Debian versions without booting.
- Want to automate firewall configuration? There's an app for that. Again, this can vary with distribution.
- Getting things to start up and talk? More distribution based behaviour.
- Configuration scripts can certainly tolerate typos, depending on the service. With many services, if you do make a typo, that setting is ignored and a warning is printed with the line number. No need to search a whole 10kb file. Unix services are generally very good at telling you exactly where and what the issue is. I only wish Windows services were as simple.
- Spaces in filenames? Works fine, though you _must_ encase filename variables in double quotes when doing scripting work. It's something that people often forget. Case-insensitive user names? Yep, you're right. Unicode text? It's there. I can type UTF-8 at the prompt without issue. IPv6? In theory and practice. GUI config wizards? Yes, depending on your distribution. Text based config wizards? There are some, but agreed, not many. Configuration change without having to stop and start? Look up the HUP signal. An editor more user friendly than vi? They're there (pico for example). Beats me why you wouldn't learn an excellent text editor like vi anyway. It's like not learning to touch type when using the keyboard. I install vim on most windows servers I administer.
- Those "-e" command line options are just shortcuts. There's often a long version of the option which is more descriptive. There's a long history in Unix of using shortcut options. Apparently, it's more efficient.

Most of the issues you mention vary greatly depending on the Linux distribution you use. If you have to complain, do so about the distribution you're using. Don't label them under the all-encompassing name of "Linux".

Re:Uh huh

[Guy Harris]

Really? That's all you got?

Where in your list is Tru64?

In the "dead" list, given that DEC^WCompaq^HP aren't coming out with new releases and given that it runs on an instruction set architecture of which no more implementations are being made and it isn't being ported to new architectures.

Irix?

In the "dead" list, given that SGI aren't coming out with new releases and given that the only instruction set it supports these days, MIPS is now targeted for various flavors of embedded computing rather than general-purpose computing, and it's not being ported to other instruction sets.

UNICOS?

These days, it's called "Linux".

SCO?

Wow, they're still around, not that they're players in the same "enterprise server" market as Oracle/HP/IBM and their respective proprietary Unixes.

Limiting yourself to Unixes you've encountered is a pretty lame standard.

Limiting yourself to Unixes that are still around, if the topic is the decline of "Unix", is, however, a rather reasonable standard.

Re:Uh huh

[dbIII]

Was an attempt at humor. But executed poorly

I think that's how the renaming started out too, especially since right up to that day any time somebody asked RMS a question about linux his reply was "never HURD of it". Funny the first six times, the next hundred not so much.

Re:Uh huh

[bertok]

I could reply in detail, but you missed my core point, so I'll pick out just a couple of the more relevant ones to reply to:

vi on the other hand has the advantage of being universal.

Which is what we were told in the lab too. But how sad is that? I shouldn't have to "make do" with a shitty text editor that's saddled with the lowest-common-denominator limitations of the 37 year old systems it was originally developed for, in 1976!!! This is the problem with both UNIX and Linux: they haven't really changed, at their core, for three or four decades. Sure, there's GUIs and whatever on top, but a soon as you want to build real systems, it's time to roll the sleeves up and get elbow deep in decades old crap, with all the limitations and inefficiencies that implies.

Sure, there might be some sort of masochistic pride in learning how to use a text editor that basically no first-time user can even exit without a cheat-sheet, but I have better uses for my time.

Type "top". When it comes up, type the letter "M" (for memory). Five keystrokes, and you get it updated continually.

Did you not read the bit where I said that the whole single-letter option thing is insane, because nobody can possibly guess what commands mean based on just one symbol?

The whole two- or three- letter commands with one-letter options isn't a good thing, it's a legacy from the ancient times when tab complete didn't exist, and "terminals" were typewriters. In that era, every character saved improved administrator efficiency measurably.

Here's a hint, in case you've been living under a rock for the last four decades: those times are over. I have a 1920x1200 LCD screen, and my phone gets 10 megabits. I don't need to be shaving a few hundred bits off my commands so that it will transfer faster over a 300 baud link, because those exist only in museums.

So back to your "example": So great, I can sort by "memory" by pressing "M". Awesome. Here's the columns I could sort by in Windows:

> Get-Process | Get-Member *memory*64 | select -ExpandProperty Name

NonpagedSystemMemorySize64
PagedMemorySize64
PagedSystemMemorySize64
PeakPagedMemorySize64
PeakVirtualMemorySize64
PrivateMemorySize64
VirtualMemorySize64

First of all, that command line made perfect sense to you, right? You can understand what it means, without having to look up any of the commands in help.

For Linux, I have no idea how what memory statistics are available for a process, but Google came to the rescue with what looks like the likely set:

vsize - The size of virtual memory of the process.
nswap - The size of swap space of the process.
cnswap - The size of swap space of the childrens of the process.
size - The total program size of the process.
resident - Number of resident set size, this includes the text, data and stack space.
share - Total size of shared pages of the process.
trs - Total text size of the process.
drs - Total data/stack size of the process.
lrs - Total library size of the process.
dtp - Total size of dirty pages of the process (unused since kernel 2.6).

Answer me this: Which one is "M"? What are the letters for the other ones? Do you know off the top of your head? If you see some random "top" command-line, would you be able to immediately identify every single option from memory? Can "top" sort by any of those columns? What about every other Linux command with a single-character options? Have you memorised all of them too?

PS: right after I finished typing all of that up, I actually read through the Wikipedia "vi" page. I love this bit:

"Joy used a Lear Siegler ADM3A term

Re:Uh huh

[Forever Wondering]

There were the unix wars awhile back, and I knew a few people at AT&T related companies who would frown whenever BSD was mentioned. The definitions of what is true Unix has changed over the years.

I was in two different mc68000 based startups. The first used Unix v7 and a C compiler that had int's at 16 bits. The second company used Sys III [and later Sys V] and a C compiler that had int's at 32 bits. The latter was much easier to program.

This second company had its CPU card and disk controller designed by a guy who had his own startup. Virtually identical hardware, but he used BSD. His company survived/thrived but the one I was in didn't. He attributed it to the selection of BSD: Many people had used BSD on a Vax and preferred that.

There was also a technical performance difference. Unix/v7/Sys3/Sys5 filesystems used a 512 byte block size. BSD used a 4KB block. I benchmarked this and the BSD system got 4x the performance [remember disk controller was the same]. I was able to compensate somewhat by putting full track read caching into the disk device driver and got a 3x bump. But, this was a "best effort" tweak against a better architectural choice [of BSD]

Basically since V6/V7 things have branched often but subsequent merging was less common; sometimes only the ideas merged rather than the actual code. There may be a common ancestor but the amount of that common ancestor that is still present is extremely small.

At that time, AT&T was extremely aggressive in enforcing the licensing, so not much code could be shared. With POSIX standards (and others like IETF RFC's), code sharing isn't as necessary/desirable as it once was. Compatible implementations are more important.

For example, Linux uses glibc. glibc has seamless multithread support using NPTL. FreeBSD's libc needs a special call to be issued at the outset to say "enable threading". After that, there are a bunch of:
    if (threading_enabled) { ...
    }
    else { ...
    }
This was FreeBSD's architectural choice (e.g. they felt the overhead of the thread lock primitives warranted this approach). My personal experience has been that the thread lock overhead is so low for the uncontended case that glibc's way is better. OTOH, glibc's passion for complex linking (e.g. weak symbols, weak aliases, symbol hiding, etc.) makes me crazy.

YMMV

Re:Uh huh

[CadentOrange]

I think I'm going to start calling it "E-sexy" as that sounds way cooler.

Re:Uh huh

[UltraZelda64]

Exactly! If GNU wasn't ESSENTIALLY Unix then RMS wouldn't have bothered to name it that! I mean, come on, Unix is right there in the name. Why name it that as opposed to giving it a whole new name of it's own that speaks to what it IS rather than what it ISN'T unless he knew that Unix was so much a part of it's identity that he had to try to define it as not Unix in it's name?

Yep--it's even in the original announcement: it's a "Free Unix!" and a "new Unix implementation".
https://www.gnu.org/gnu/initial-announcement.html

I'm pretty sure the name was intended to be somewhat of a joke, and with their love of using recursive acronyms to come up with "clever" names (just look at the Hurd...) that's probably the case. But in a way the name is kind of ironic, and likely on purpose, because they weren't allowed to actually call the system "UNIX" and I'm sure they knew that, while at the same time it gets the point across quite well that it's not "really" UNIX--it is a clone.

Re:Uh huh

[smash]

Its double the cost of x86 hardware in terms of RAM and CPU. CPU is generally abundant; you already have multiple hosts in your cluster anyway. You can limit it to VMs that really need to be up, only.

Re:Uh huh

[rastos1]

-- I love the undecipherable command-line wizardry. I'm not an idiot, but how-the-fuck would I know what "-e" does on some random command? There is just no way without trawling through man pages using a command-line reader with no mouse support and keyboard shortcuts I don't know. Compare this to a sample PowerShell pipeline "Get-Process -Name 'n*' | sort -Descending PagedMemorySize". You'd have a hard time finding an IT engineer that can't figure out what that does.

Once a year I hit a task that makes me think "this could be probably best done with PowerShell". Last time it was: list installed packages and their versions. So I fire up PowerShell window and ... have no idea how to solve the task. So I start googling and find out that what I want is Get-WmiObject. That can list a lot of things including the installed software. That is decided by the parameter "-Class" which should be followed by a string. The allowed values of that string can be obtained by running "Get-WmiObject -List". There is 1024 of them! And can be probably more depending on your system. Of course the documentation does not list them. The right one in my case was probably "Win32_Product". But I have no idea whether it really lists all that I want or only software from MS or only software installed using .msi, or 32-bit software or software that made the Right Thing(tm) during installation, ...

So perhaps the linux commandline arguments look undecipherable to you. But I'll take them over PowerShell any day.

Re:Uh huh

[Darinbob]

One big advantage of BSD in the early days was its vfork() system call and it took a long time for AT&T versions to adopt a similar feature.

Re:Uh huh

[jalopezp]

I'm pretty sure the gp was just being coy.

Re: Uh huh

[IRWolfie-]

No it isn't open source. They haven't released their code in a long long time.

Re: Uh huh

[Midnight Thunder]

MacOS X's core OS is open source. You can download the kernel and recompile it and swap yours in if you want to, and all the standard user space stuff is basically FreeBSD.

Also, it is a certified UNIX 03 operating system, so it is more "UNIX" than Linux, which is what I assume you're comparing it to.

Just to provide a link to make life easy (source is available up to 10.8.4 and includes BSD licensed stuff. Code for the UI level is not provided.):

http://opensource.apple.com/

and some documentation to backup what you are saying:

https://developer.apple.com/library/mac/documentation/MacOSX/Conceptual/OSX_Technology_Overview/SystemTechnology/SystemTechnology.html

With regards to Darwin, there are two related sites:
  - http://darwinbuild.macosforge.org/
  - http://www.puredarwin.org/

Re: Uh huh

[Midnight Thunder]

No it isn't open source. They haven't released their code in a long long time.

10.8.4 is the latest and is available: http://opensource.apple.com/release/mac-os-x-1084/

Note you should expect a few months delay between release of a new version of MacOS X and release of the open source components on this site, but up until now they have always delivered.

Re:Uh huh

[Billly Gates]

Ok, I'll bite.

I'm a Windows admin, but I just went to a training course to learn about a high-end enterprise product that runs on top of Linux. I've dabbled with Linux-based stuff before (proxies, VMware, ESX, etc...), so it's not exactly new territory, but I figured it's 2013, it'll be interesting to get a glimpse into the current state of the "Linux Enterprise" world.

My experience was this:

-- You still need to patch, or install 140+ dependencies to install one application. Same difference.
-- You still need to reboot. A lot. More than I thought. I suspect that it is possible to avoid most of them though by judiciously restarting services, but the effort is much higher and the outage level is practically the same, so what's the benefit, really?
-- Things that really ought to be automatic, aren't. I spent a good 50% of the lab doing really fiddly things like cut & pasting iptables rules to open firewall ports. The installer really should have just done that for me.
-- Binding services together and just generally getting things to start up and talk required an awful lot of error prone manual labour. The lab guide was liberally sprinkled with warnings and "do not forget this or else" sections. Lots of "go to this unrelated seeming file, and flip this setting... because.. just do it or nothing will work."
-- I love the disclaimer in the training guide: "Linux configuration scripts do not tolerate typos, are case sensitive, and are not possible to validate before running the associated service." Fun stuff. I can't wait to diagnose random single-character problems in 10 kilobyte files when the only error is that one of a dozen services barfed when started.
-- Wow, the 70s called and wanted their limitations back: spaces in file names? You're risking random failures! Case-insensitive user names? Nope. Unicode text? Hah! IPv6? In theory, not in practice. GUI config wizards? Nope. Text-based config wizards? Not many of those either. Want to make a configuration change to a service without having to stop & start it? You're dreaming! An editor more user friendly than vi? Eat some cement and harden up princess!
-- I love the undecipherable command-line wizardry. I'm not an idiot, but how-the-fuck would I know what "-e" does on some random command? There is just no way without trawling through man pages using a command-line reader with no mouse support and keyboard shortcuts I don't know. Compare this to a sample PowerShell pipeline "Get-Process -Name 'n*' | sort -Descending PagedMemorySize". You'd have a hard time finding an IT engineer that can't figure out what that does.

I keep hearing about the supposed efficiency advantage of Linux, but I just don't see it. Given a Hypervisor, PowerShell, and Group Policy, Windows administration a piece of cake in comparison.

Go admin an IIS server and your opinion will change rapidly!

I play with some php scripts on some VMS on my home machine but the admins I have talked too love the CLI and the non registry of Unix. There are some things were the only option is to do a reformat and install because of something you did with Windows 6 installations ago. Changing an XML file wont fix it as it is event driven and the registry is a black box of cryptic hex decimal keys!

With Apache or cgnix you just edit the files and resstart the service. Done!

Yes shell scripts are a pain to work with but you are in charge of a complex operation and have full control. It is like the old joke with the unix plane being a full f-16 you need to assemble while the Windows one is done as a crappy flight 737 with glitches with the everything bolted on so the pilot just works around the glitches.

When playing with a full setup for $100,000+ worth of equipment such skill is needed is needed. Maybe not to run Word but customization as pain as it is required. Even under windows you need to know registry entries and word seldom used win32 apps in the /system32 directory in your admin work and advanced VSphere knowledge before a recruiter will even talk to you.

Re:Uh huh

[Danathar]

And there is the problem. How do you define the market for "UNIX". There is a decline, but it's in respect to a specific segment of the UNIX market, not the UNIX market as a whole.

If you factored in every Iphone and/or IOS device which you could easily categorize as a UNIX based system (since it's based on Mach and BSD right?) then the stats would be completely different.

Re:Uh huh

[fatphil]

> A failed CPU on a VMWARE box would cause the machines to migrate over to a standby, and be up and running before anyone even knew.

Is that a feature? Personally, if a CPU's potentially gone haywire, I don't want to just resume a potentially corrupted system state on another CPU. Imagine if all TLB loads caused the line they read from to be blanked. (I've seen worse from dodgy CPUs.) The safest behaviour on such a failure is to do nothing.

Re:Uh huh

[dkf]

It's powered by the users own sense of self importance.

Did the BMW patents expire already?

Why do you think there's no iCar yet?

Re:Uh huh

[Lonewolf666]

Makes sense.

The right time to switch would be when a non-trivial change to the applications on the VMs becomes necessary anyway. At that point, you might as well go all the way to port the applications to some other system (Linux?).

Re:Uh huh

[fatphil]

In the 80s, I was using "the VAX cluster". In what way is a "cluster" new?

I'm no IBM fanboi, but I fully understand it when the system z-heads laugh at linux re-inventing things they were doing 2-3 decades earlier.

Re:Uh huh

[MachineShedFred]

Or they are already switching to Thin Clients and using desktop virtualization (Citrix, VMware Horizon, etc.).

I know that we're looking at it, and we don't jump into anything that is cutting edge.

Re:Uh huh

[drinkypoo]

Those who understand Unix are doomed to reinvent it... functionally

And those who do not understand software development are doomed to try to compete with Linux.

Re:Uh huh

[drinkypoo]

The safest behaviour on such a failure is to do nothing.

The safest behavior is to drop back to a known-good checkpoint on another node, something like a complete system shutdown and image. What vmware is doing is the best compromise between that best case and the case where you really need to continue performing work. But if you know that you have problems that demand that level of paranoia, you are free to process your workloads through a job-processing system on systems which will panic and explode rather than risk compromising output.

Re:Uh huh

[Shatrat]

I would agree that you can't generally say "1 HP-UX server can always be replaced by 1 Linux server", but I don't think it's a stretch to say "1 HP-UX server that isn't doing anything that can't be ported to Windows, could be even more easily replaced by Linux". Also the fact that they multiplied their hardware footprint by 8 is a pretty good indication that Windows wasn't a good fit for whatever application they were needing.

Re: Uh huh

[Rational]

Looks like according to some people here, the "UNIX market" must necessarily exclude Apple, because 'teh openz' or something.

Re:Uh huh

[K. S. Kyosuke]

Isn't that their practiced business model?

I'm puzzled, are you talking now about the prices of Microsoft software, or HP-UX? I sort of thought that everyone is doing that.

Re:Uh huh

[Rob Riggs]

A) I did not say a cluster is new. B) You did not have a VAX cluster so much as you had a DEC cluster. C) I have no problem at system-z heads laughing at Linux re-inventing things . Whatever features system-z has -- it (and your DEC cluster) has one key misfeature: vendor lock-in.

Re:Uh huh

[mcalchera]

Posting to undo a misclicked moderation.

Re:Uh huh

[MightyMartian]

It's Unix-like, which is close enough these days.

Re:Uh huh

[slash.jit]

dont forget Mac OSX!

Re:Uh huh

[Just Some Guy]

I wouldn't have phrased it against "old guys like you" that way, but you're pretty much right. I'd much rather scale horizontally than depend on vertical hardware upgrades being able to handle increased loads.

Re:Uh huh

[cundare]

And so is the Mac OS...

Re:Uh huh

[Pippinjack]

Surely you mean GNU/WOOSH?

Re:Uh huh

[Guy Harris]

Solaris 1 implemented System V like init

No, it didn't. Solaris 1 was built atop SunOS 4.1.1 through 4.1.3u, which had a 4.3BSD-based kernel and a 4.3BSD-based init.

Re:Uh huh

[s.petry]

Wasn't Solaris 1 when they implemented /etc/init.d/ and /etc/rc?.d/ structures? Testing my memory depths on this one :) Thanks!!

Re:Uh huh

[Guy Harris]

This is really bass ackwards. Android has the same GNU userland as most other Linux distros or did you not know you can run ls, rm, ps et al. on your phone?

Command-line GNU userland, maybe. libc userland, libc GNU userland, not so much.

Re:Uh huh

[Guy Harris]

Wasn't Solaris 1 when they implemented /etc/init.d/ and /etc/rc?.d/ structures?

No. As the person who 4.3BSD-ified the SunOS init in SunOS 4.0, I know that the 4.x init was 4.3BSD-based. (The only tricky bit was preserving binary compatibility for pre-4.x binaries that used ttyslot() and therefore had, in the binary, code that expected /etc/ttys to look like the V7 /etc/ttys; the SunOS 4.x init would look for /etc/ttytab, which had the same syntax as 4.3BSD's /etc/ttys, and parsed it as such, writing out a legacy-format /etc/ttys.)

Re:Uh huh

[benjymouse]

Yep, installed packages are available through the Windows Management Instrumentation - where you'll find rich information about everything related to system management, from system board, network adapters, performance counters, drivers over printers, volumes, disks to installed software, updates, running services etc.

Yes, this model offers much more information that you'll ever get from a Linux system unless you have dip into some serious log file archeology.
The WMI classes are not part of PowerShell, but PowerShell will seamlessly integrate with the WMI/CIM model. Consequently, the WMI classes are not documented as part of PowerShell. However, they are very well documented. To get a little feel for what type of information can be obtained for a Windows system, look at the documentation for WMI. There are also CIM classes which should work cross platform with any system that adheres to DMTFs WBEM protocol and CIM model. Yes, they are available for Unix/Linux as well.

Your example is actually very interesting. You managed to figure out that Win32_Product was *probably* the WMI class you wanted. Great! here is how you will progress from there: run the command and pipe the result through the Get-Member cmdlet to get an overview of the available information (properties, methods).

Get-WmiObject win32_product | Get-Member

From the answer (cannot be shown here due to ./ junk character filter) you can see that Win32_Products has (among other properties) a caption and a version. Hence, you type (or press up-arrow and edits the prev line):

Get-WmiObject win32_product | Format-Table caption,version

And voila you have your installed products with versions, nicely formatted in a table. This workflow is actually very, very common with PowerShell:
1) locate the command,
2) pipe the results through get-member (which has the standard alias gm because it is used so frequently),
3) chose properties,
4) re-run the command but pipe the results to one of the export-, out- or format- cmdlets.

Yes, you can export the list as comma- or tab-separated values instead by piping the results through Export-Csv:
Get-WmiObject win32_product | select caption,version | Export-Csv myproducts.csv

The GPs point was exactly about the discoverability and consistency of PowerShell. There is no overloaded use of -e options. All options has meaningful names, but can be abbreviated as long as there is no ambiguity. Output from cmdlets are self-describing and you can discover both properties and methods.

And PowerShell works with objects. If you inspected the output from the Get-WmiObject win32_product | Get-Member, you would notice that the returned objects each expose an Uninstall method. See if you can figure out what Get-WmiObject win32_product | where name -like "*Java*" | foreach Uninstall will do.

Re:Uh huh

[nukenerd]

@AC, 22:01 wrote :-

In five years PCs will be rare, because 99% of people who would use a PC in 2010 use their phone instead.

Fantasy. Among many factors why this will not be true, just consider the fact that the IT-using population is ageing, and anyone over the age of about 45 has deteriorating eysight. In 5 years that is anyone over 40 today. You seriously expect those people to look at web pages (or anything else) the size of a large postage stamp?

Re:Uh huh

[seant2013]

Agreed OMG nux'es are attacking the bigger Unix'es !!! Give me a break shesh!

Re:Uh huh

[s.petry]

Awesome stuff! Now when time allows I need to go back and read. I started to think "wow, this seems like so long ago" and then I realized that it was..

Re:Uh huh

[benjymouse]

Go admin an IIS server and your opinion will change rapidly!

I play with some php scripts on some VMS on my home machine but the admins I have talked too love the CLI and the non registry of Unix. There are some things were the only option is to do a reformat and install because of something you did with Windows 6 installations ago. Changing an XML file wont fix it as it is event driven and the registry is a black box of cryptic hex decimal keys!

You may be good at playing Unix. But your ignorance of Windows does not mean a Windows admin would be equally clueless. Clue: IIS does not rely on the registry for website or server administration. The registry contains very, very basic configuration parameters. You may have experience with (or talked to someone who used) IIS6 and the metabase. That was still not the registry, though.

Apache httpd's config files are one of the best examples of how convoluted file based configuration can get. It is a pseudo-xml line-oriented format where every module use it's own way to refer to the same concepts, such as headers and cookies.

And yes, IIS sites are now fully configurable through XML config files, which are built using an extensible model (config sections and config groups).If you do not like XML files there is also a full configuration API exposed as .NET as well as PowerShell cmdlets.

You can now type
IIS:
which will change your PowerShell "working location" to IIS. Yes, the IIS management model (sites, application pools, bindings etc) is exposed as locations. You can now even cd to your site:
cd Sites/MyCompanySite
From there you can list the content using ls. From the Sites node you can create, remove, rename and copy websites as if they were files, using the same cmdlets you would use for files. What could be easier than that?

Oh, how would an Windows admin figure out what he could do with the webserver? He lists the commands of the webadministration module:
get-command -module webadministration.

IIS can be administered through text (xml) files, .NET/COM API, PowerShell, the management console plugin and graphically through the IIS administration tool. How again did you administer an Apache? through cpanel? (yikes)

Re:Uh huh

[Rutulian]

but I just went to a training course to learn about a high-end enterprise product that runs on top of Linux

Uh huh, sure. Unless your "high-end enterprise product" runs on a 10 yr old copy of Slackware, what you've written is complete baloney.

Compare this to a sample PowerShell pipeline "Get-Process -Name 'n*' | sort -Descending PagedMemorySize"

As I'm sure you had to learn how to use Powershell, you will likewise have to learn how to use the command line in linux. And when you do, you will know what -e does for the most important commands. Btw, the majority of common linux commands have long-name switches along with their shortened counterparts (ex: cp -a vs. cp --archive), as well as quick --help pages to refresh you if need be.

Re:Uh huh

[rastos1]

To discover how something works I turn to the documentation. If it wasn't for a random page on the Internet I would have no chance to discover that "Win32_Products" is the class that i need. Yes, PowerShell is a powerful tool, but there is too much of various objects that I might encounter. I will need object X today and Y in a month and they have nothing in common - except perhaps the formatting/selecting cmdlets. In Linux I learned at about 20 utils 15 years ago (binutils/textutils/fileutils)|. Just a basic form of those - and I can quickly find out any advanced options for them in a man page. They all work with files and text stream. And I can accomplish every task with them by just re-using the same utils. I don't have to discover what do they return.

P.S.: It seems to me that for some reason you assume that -e option for one program (e.g. grep) has something to do with -e option for some other/unrelated program (e.g. ps). Why would one expect that is beyond me.

Re:Uh huh

[Billly Gates]

I will take your word at it. I never met an IIS admin who felt competent using it as they always complain you change something and it breaks another component and nothing just works as clearly which to me gave the impression of no distinct seperation of things like application pools and bindings.

I worked iwth IIS just briefly for OWA during Exchange setups so I will take your word at and assume they were all disgruntled former Unix admins.

I really do hope you are right and no registry is used to tune IIS or any MS products so much. In the past MS was notorious for it! One of the great things about using a Mac is you can backup and move a whole program by just copying it! Try that on a windows PC?

Re:Uh huh

[Forever Wondering]

vfork was definitely useful in its day [before COW]. I was curious as to whether vfork still had any advantages over fork in a modern system, so I did a little search. vfork can still be faster than fork [even with COW]. It's particularly useful on non-MMU systems (e.g. embedded).

But, it's considered a [small] security hazard due to race conditions with the child: privileged process does vfork, child lowers its privilege level [in prep for execve], but then child gets a signal. The signal can corrupt the parent and/or pin it (you have a non-privileged [child] process suspending a privileged [parent] one). vfork is used by some libc's to implement posix_spawn, but it's use is controversial unless some precautions are taken.

BSD's most lasting/widespread contribution may be BSD sockets. Prior to 1989 we had the [ghastly] OSI/ATT Streams/TLI. In 1989, BSD became unencumbered by licensing issues [from AT&T] and could release the socket code/specs. Thus, the internet in the [technical] form that we know today comes from BSD.

Re: Uh huh

[IRWolfie-]

I am aware of that list. The operating system itself is not on that list, specifically the kernel as well. Consider that OpenDarwin shut down for the express reason that they couldn't get the code off Apple, I don't see what you are talking about.

Re:Uh huh

[0ld_d0g]

One of the great things about using a Mac is you can backup and move a whole program by just copying it!

Except all of the programs that cant be copied and need to be installed via installers (or install kernel extensions)

Try that on a windows PC?

Turns out its so easy that Microsoft let you do that since Windows 1.0. Tons of windows apps work just fine by copying them around. The most popular example I can think of is Steam. I first installed it ~2006 and I have just copied the Steam folder everytime I upgraded the OS. (Steam installs a service but it detects that you've moved it and re-installs it if its missing). For more apps just Google 'portable apps'. e.g. - http://portableapps.com/ The Application Bundle in OSX is just a folder. I don't see anything technically special about it.

Re:Uh huh

[Bill, Shooter of Bul]

In all honesty, I don't understand all of your post. Maybe you are trying to make a larger argument about how Microsoft was successful in persuading everyone to write Microsoft specific aps by providing examples that don't really make sense in the domain of server software? Point kind of taken, I guess. But I don't really see how that applies here as they are moving in 2013 from Unix to Microsoft. I don't think Microsoft in 2013 is as compelling as it was in 2003. Without knowing their specific requirements its tough to really judge.

Re:Uh huh

[malvcr]

This is a focus problem.

UNIX (branded) : One very big and expensive software license for a very complex and capable "enterprise-level" operating system running on an even more expensive and complex computer.

UNIX (derived) : One "or many" -- even counted on the thousands, small and more limited computers performing specific tasks on cheap computers, covering not only what the UNIX-branded was capable of doing but much more things, with less cost and more flexibility. And with this specialisation comes the expansion, because they even can run a UNIX-derived operating systems in mobile devices, covering practically all the world.

In my own perspective, UNIX was as important for current UNIX-derived operating systems, as DOS was for current Windows. They are impossible separable parts of the same evolution story.

Re:Uh huh

[cgi-bin]

Am I the only one who read that as NOOB-FAP?

Re:Uh huh

[smash]

And I've got windows VMs that haven't had a reboot in over 6 months as well. What's your point? The only need to take the machine down is for scheduled updates that can't be mitigated by other measures in the meantime. Given the vast majority of issues appear to be browser problems or GUI issues when logged into the machine (and my servers are pretty much never logged into) then I can quite happily not run windows update every patch tuesday and still remain safe.

Re:Uh huh

[jerryjnormandin]

This is all FUD. Everyone knows that windows does not scale well. I get to play with the IBM Power7 Platform and Linux at work. I have got to say my favorite platform is still the IBM Power Platform scales extremely well. I've got 20 HMCs, 196 frames, and around 800 LPARS and VIO servers. LPAR2RRD rules!

Re: Uh huh

[samkass]

I am aware of that list. The operating system itself is not on that list, specifically the kernel as well. Consider that OpenDarwin shut down for the express reason that they couldn't get the code off Apple, I don't see what you are talking about.

The entire "operating system" isn't there, but the XNU kernel and UNIX user space are all there up to the latest MacOS X point release. Enough to get a bootable OS with shell and a full suite of UNIX utilities. Honest. Here's some instructions for building the kernel that you can then swap in. Then you can download all the userspace packages and build and swap them in as well. What isn't provided is a nice set of changelogs, package installers, open bug database, etc. But the code is available and BSD licensed.

Re:Uh huh

[jwhitener]

Well, and it isn't just VMWARE. There have been other strategies and virtualization techniques allowing for the use of low cost linux boxes in highly available environments for quite a long time now.

I don't know anyone, who has the choice software-wise, staying with Solaris or HP-UX nowadays.

Re:Uh huh

[jwhitener]

"Linux adds some of the stability to the bulk hardware, but the bulk hardware is simply not as good"

We've been moving away from Solaris / HP-UX 'big' servers to lots of VMWare clustered linux servers for a long time now. Uptime is identical if not better, and it is way cheaper. Physical servers basically start to be seen as hot swappable. The apps don't care if one dies or is moved.

The only people I know who are still using Solaris or HP-UX are those that can't move off those platforms because of specific apps that have no linux equivalents.

Re:Uh huh

[s.petry]

You may not work with very large applications either? One of he reasons we run many many Sun/Oracle systems is SMP applications that require massive CPU count. I have yet to see a PC that supports 128 physical processors. DMP does not work for everything, and SMP is still vastly superior in performance.

Re:Uh huh

[morgauxo]

"To begin with, GNU will be a kernel plus..."

Ha Ha!

Re:Uh huh

[jwhitener]

Probably not on your scale. I work with a few services that are a couple million hits a month each. Max concurrent sessions around ~5,000 for most of them.

But that is basically what I was saying. Unless your application demands it, moving away from Sun/Oracle/HPUX makes sense now. In terms of pricing and stability.

Re:Uh huh

[s.petry]

Still varies depending on what your requirements are. If I need 100% uptime, a small IBM P will be cheaper than a cluster of Windows Boxes. Assuming the applications can run on a P series. It's cheaper to run numerous small web servers on a couple of small Oracle/Sun boxes and use Solaris zones. No license costs for Zones, and VMWare gets pretty expensive when you look at capabilities that require fail over and true redundancy. Lets not forget that you are probably stuffing a F5 or something in front of Windows boxes to get HA, and that needs to be included in your costs.

Claiming that Virtualization fixes hardware woes is simply untrue. You still have them, and license costs to boot in the case of VMWare. It's easy to use fuzzy math to make claims that some things are cheaper. Fuzzy math does not always match reality, and in some cases completely ignores reality. I believe I demonstrated that above in the overview of the actual costs of migrating to Windows.

Re:Uh huh

[s.petry]

I should add that if you had said Linux instead of Windows it would have been a more favorable point. Windows requires MS licensing and maintenance, some type of Anti-virus, patching, administration software and more people to man the work. It still may not have been true mind you, but closer than Windows would be.

Re:Uh huh

[Penguinisto]

Yes, this model offers much more information that you'll ever get from a Linux system unless you have dip into some serious log file archeology.

You have no idea at all what /proc is, do you? If you did, you'd never make such an ignorant statement.

Re:Uh huh

[Penguinisto]

And I've got windows VMs that haven't had a reboot in over 6 months as well.

If I didn't give a damn about security and the server didn't do anything important, so would I. ;)

Re:Uh huh

[Archangel Michael]

What you propose, is completely possible within VMware as well. All you have to do is snapshot your machines to the tolerance you need. But then again, how does one know which snapshot is the last good one? There is a risk in ever action you take, you're just minimizing that risk. How much risk you can afford is the line you spend mitigating against the damage done.

And since we are talking blade chassis, the blades themselves do not have fans. The Chassis has fans, and fan redundancy, so you can lose a fan and not have issues with airflow. And if you're that mission critical, you have enough chassis' to handle one being offline while you perform maintenance. And you have redundancies for network, air conditioning, power, and a duplicate "disaster recovery site" for when an airplane crashes into your building, in a different geographic location.

Disaster planning is not difficult, it is expensive. And if you're doing it on a shoestring, you're going to screw up somewhere on something you overlooked, so don't; you're not qualified to do disaster planning.

Re:Uh huh

[fatphil]

> You did not have a VAX cluster so much as you had a DEC cluster

Bullshit.

I had access to a "VAXCluster". That was DEC's name for it, what right do you have say that it was called something else? Sheesh.

How long has Netcraft been confirming BSD dead?

[ron_ivi]

How long has Netcraft been confirming BSD dead?

Re:How long has Netcraft been confirming BSD dead?

[morcego]

BSD confirmed Netcraft is dead.

Re:How long has Netcraft been confirming BSD dead?

[gstoddart]

I have no idea, but Netcraft has been confirming (or not confirming things) have been dying so long that Netcraft themselves have become a punchline to me.

Because, other than their periodic confirming that something is dying, I have no idea of who the hell they are or why I'm supposed to care about what they tell us.

When I see "Netcraft confirms it", it's just another bad internet meme to me. Are they actually relevant to anything?

Re:How long has Netcraft been confirming BSD dead?

[i kan reed]

The reason it became a meme is because someone submitted that headline, while netcraft itself was in no position in the industry to confirm much of anything. That's the joke. Someone actually thought "netcraft confirms it" meant anything at all, and it was funny. And reiterated about whatever the next few articles were. Meme became ensconced.

Re:How long has Netcraft been confirming BSD dead?

[gstoddart]

That was kinda my point ... Netcraft confirming something has been a meme on Slashdot as long as I've been using Slashdot.

And that's at at least 2 or 3 weeks I think. ;-)

Re:How long has Netcraft been confirming BSD dead?

[i kan reed]

I was explaining the origin. I'm confused how that can be your point.

Re:As if ...

[WarJolt]

Nope...Unix engineers don't have to worry about being publicly humiliated if they checkin bad code. Mostly because the public doesn't really care.

I thought OS X was Unix

[cyfer2000]

If OS X is Unix, what do you call iOS. And if we take Linux as a kind of Unix, how about Android? Or maybe the title should be written as "the steady decline of Unix Server License sale"

Re:I thought OS X was Unix

[blueg3]

I don't know if iOS is a certified Unix, but OS X is. Linux is not.

Re:I thought OS X was Unix

[MightyMartian]

You can put Linux userland on most (unlocked) Android devices, and really Dalvik is, for the most part, just a userland layer itself. That's like saying someone who install Debian for the specific purpose of running Python or Java apps, somehow he's not running Linux.

I don't suppose there's a hard or fast definition, but if the kernel provides Unix-like services in a Unix-like way (in other words, supports Unix system calls) and is at least capable of running Unix or Unix-like userland utils, then I'd say its Unix. I know that might mean by a very loose definition that Cygwin makes Windows a Unix-like operating system, but as Cygwin is basically a compatibility layer that runs on top of then kernel like any other userland application, I think I'm on reasonably solid ground to state that Cygwin does NOT make Windows a Unix-like operating system, any more than running a Nintendo 64 emulator on my Windows box makes my Windows box a Nintendo 64.

Re:I thought OS X was Unix

[dtgm]

If OS X is Unix, what do you call iOS. And if we take Linux as a kind of Unix, how about Android?

Or maybe the title should be written as "the steady decline of Unix Server License sale"

No. The title is right. You are just trying to generalize something that is specifically, and legally defined. You can argue that other systems even, DOS are similar to or like UNIX, but you can't say that they are UNIX.

Mac OS X, client and server are UNIX because they satisfy the Single UNIX Specification. They have to satisfy this specification because the Open Group said so. The Open Group owns the trademark for UNIX. The SUS is an extension of POSIX drafted by the Open Group. If you say you are a UNIX system, it means you satisfy the owner's requirements. If you don't, you are doing an illegal thing and the Open Group can take you to court if they chose to. This isn't evil. What is happening is people are trying to generalize UNIX into a broader definition. It is the trademark owner's responsibility to fight these attacks against their property. If they don't, the name falls back into general use.

Having standards are important. USB would be pretty terrible if companies made USB devices that worked similar to other USB devices. You may have to purchase a USB controller that worked for that device, which really gets away from what USB is all about. The same is true with UNIX. UNIX machines typically run enterprise type software. For instance, Oracle bought Sun; Oracle produces enterprise database software and Sun produces the OS they run on, Solaris SPARC. Eventually Solaris ran on x86, but that is getting offtopic. What's important, is these machines need to have high uptimes and allow different manufacturers to compete with each other to prevent a company from getting locked into one manufacturer of UNIX machines. In theory, you should be able to replace an IBM AIX with Sun Oracle Solaris or HP-UX.

See here for versions of UNIX that satisfy the 03 revision: http://www.opengroup.org/openbrand/register/xy.htm

Re:I thought OS X was Unix

[Savage-Rabbit]

If OS X is Unix, what do you call iOS. And if we take Linux as a kind of Unix, how about Android?

Or maybe the title should be written as "the steady decline of Unix Server License sale"

OS X has a Unix certification, iOS as far as I know does not. Of course this entire discussion would become moot if somebody were to foot the bill and get Linux Unix certified.

Re:I thought OS X was Unix

[dtgm]

I just wanted to clarify what I wrote above that using "Unix-like" to describe a product is illegal. It's the same as if Burger King said, "Our company is McDonald-like" Basically, if you are trying to name-drop, or profit from another trademark's popularity, to increase your own, you are infringing on their rights.

Re:I thought OS X was Unix

[unixisc]

Reading the headline of this story, it seemed to be more about the old school UNIXes from yesteryear - the Solaris, Irixes, Unixwares of the world - they're not refering to OS-X, even if OS-X is legally UNIX since it satisfies SUS. What you wrote above is correct, but what's being used here is more a colloqual usage of the term 'UNIX', which nowadays is sometimes understood to mean the traditional, proprietary, mainly non-Intel editions of UNIX, rather than an OS that receives a certification from the Open Group for satisfying the SUS.

The generalizations are not 'attacks', if there was prior usage before UNIX got owned by the Open Group and had new rules laid down for when it can be used. For instance, SunOS may not satisfy current SUS standards - would you then say that it was not UNIX? Current OSs, such as Linux or OpenBSD or Minix may not call themselves 'UNIX' unless and until they've passed the SUS and been certified by the Open Group. But then, should we then resurrect the term 'POSIX' to cover all OSs that broadly fall under the UNIX umbrella, such as having shells like Bourne, kernels like Linux, Minux, BSDs or Illumos, or other common features?

Re:I thought OS X was Unix

[Guy Harris]

If OS X is Unix, what do you call iOS. And if we take Linux as a kind of Unix, how about Android? Or maybe the title should be written as "the steady decline of Unix Server License sale"

No. The title is right. You are just trying to generalize something that is specifically, and legally defined. You can argue that other systems even, DOS are similar to or like UNIX, but you can't say that they are UNIX.

Mac OS X, client and server are UNIX because they satisfy the Single UNIX Specification.

OS X client is irrelevant to the Network World article, as it's talking about servers, and OS X server is pretty much irrelevant to the server market they're talking about, so "The Steady Decline of Unix Server License Sales" or "The Steady Decline of Unix Servers" more clearly states what's happening than does "The Steady Decline of Unix".

Re:I thought OS X was Unix

[Patch86]

Depends on how you define Unix. OSX/iOS are based on FreeBSD & NetBSD. If you're talking about official, formal Unix OS's, neither Linux, the *BSDs or OSX would count as none of them are entitled to use the Unix trademark; only Solaris, HP-UX and AIX are, that I can think of.

If you're talking about POSIX compliance, OSX does count, but Linux and (oddly) the *BSDs don't. Even more bizarrely, you might count Windows as POSIX compliant due to Cygwin and the like.

If you're talking about "descended from Unix", you'd be able to count *BSD and OSX, but not Linux. If you mean "functionality based on Unix", you'd have to include Windows, which has whole vast swathes of BSD code in it, not to mention stuff it came by through it's VMS heritage. And if you mean "works a lot like Unix", you could probably reach a definition which includes Linux, *BSD & OSX, but excludes Windows.

Or to put it simply- it's a really complicated question unless you're really clear about your terms...

Re:I thought OS X was Unix

[serviscope_minor]

I don't know if iOS is a certified Unix, but OS X is. Linux is not.

Honestly who the fuck cares?

To reuse an analogy from above, that's like claiming that all the quacking waterfowl that seem to like being fed bread in the park are not "certified Ducks(R)" because because the man from The Duck(R) Group(TM) has not been paid to come along and certifiy the ducks. The other reason to not care is that the duck certificiation man is not, in fact, an ornithologist and would happily certify a goose as a Duck(R)(TM).

Re:I thought OS X was Unix

[UltraZelda64]

I wouldn't be surprised if Apple just kept the cash and didn't bother to get iOS certified. I can imagine it now; "Certified UNIX" in Apple's iOS advertising material, and people all around the world mock it and laugh at the concept of an OS designed for an overpriced trendy toy tablet/phone being advertised as certified UNIX. Why would Apple even waste the money? On the other hand, getting their main desktop/workstation/server OS "certified" makes more sense, and that extra bullet point in their advertising probably does actually help out a bit.

Personally, I think it's all a big joke--the entire concept of UNIX as a trademark certification. It requires a wad of money just to be considered which is the biggest problem (hey, you can buy your way into anything even if your product itself is not deserving of it... just wave the $$$). This leaves countless potential operating systems out even if they have technical advantages, leaving only a select few legally capable of carrying the prestigious UNIX name, and in many cases they really don't have anything truly special to make them worthy of that honor.

Re:I thought OS X was Unix

[Guy Harris]

To reuse an analogy from above, that's like claiming that all the quacking waterfowl that seem to like being fed bread in the park are not "certified Ducks(R)" because because the man from The Duck(R) Group(TM) has not been paid to come along and certifiy the ducks.

Well, in fact, legally speaking, they're not Duck(R) waterfowl. The point is that, given that they walk enough like a Duck(R) waterfowl, and quack enough like a Duck(R) waterfowl, people are willing to treat them the same way they treat Duck(R) waterfowl even though they cannot legally be called Duck(R) waterfowl. (I.e., people are willing to use Linux distributions as a substitute for Unix(R) systems because, legally speaking, even though they're not Unix(R) systems, they're enough like it that the differences aren't worth spending the extra money for a Unix(R) system.)

The other reason to not care is that the duck certificiation man is not, in fact, an ornithologist and would happily certify a goose as a Duck(R)(TM).

Well, yeah, I suspect most sites with Unix applications - oops, sorry, that's "applications for Unix(R) systems" :-) - are probably not going to port them to z/OS (the certification for which is for a standard that's almost 20 years old, BTW; are there any geese that have been certified as a recent version of a Duck(R)? The only recent Ducks(R) I see are all stuff that could reasonably be considered Real Unix), given that if their apps assume they're using ASCII, they're in for a rude surprise....

Re:I thought OS X was Unix

[Guy Harris]

Depends on how you define Unix. OSX/iOS are based on FreeBSD & NetBSD. If you're talking about official, formal Unix OS's, neither Linux, the *BSDs or OSX would count as none of them are entitled to use the Unix trademark

Wrong.

Re:I thought OS X was Unix

[Shag]

Depends on how you define Unix. OSX/iOS are based on FreeBSD & NetBSD. If you're talking about official, formal Unix OS's, neither Linux, the *BSDs or OSX would count as none of them are entitled to use the Unix trademark; only Solaris, HP-UX and AIX are, that I can think of.

UNIX® is a registered trademark of The Open Group (ditto for Motif, and the X logo is also their trademark), who seem fairly certain that recent versions of Mac OS X are, in fact, certified as being UNIX, and entitled to use the trademark.

Re:I thought OS X was Unix

[Patch86]

I stand heartily corrected.

Re:I thought OS X was Unix

[blueg3]

It doesn't matter, except that it's a pretty handy measure of whether or not something "is Unix". Mostly it's useful for saying: if you consider Linux to be "a Unix", you must also consider OS X to be a Unix also.

This branding / certification nonsense goes back pretty far, and to my knowledge, GNU/Linux has never been a certified Unix: hence the annoying *nix appellation.

Re:I thought OS X was Unix

[tomxor]

While OS X (Darwin) has bits of FreeBSD and NetBSD, it can't count those for source inheritance of UNIX as they are based on 386BSD which had functional inheritance of AT&T's UNIX and none of it's source. Although AT&T and co made sure to include plenty of 386BSD and derivatives code into their UNIX after. The difference being that while there is (at some points in history at least) portions of same source between them, it's all from the BSD's to the AT&T descendants, the other way around wouldn't make FreeBSD free anymore. Darwin is UNIX only because of POSIX and not because of BSD but because for some reason Apple found value in adhering to that specification.

Try

[Impy the Impiuos Imp]

What if we re-wrote it in "D"?

A distinction without a difference

[Fished]

The distinction betwen "Linux" and "UNIX" is virtually meaningless. All of the traditional proprietary unixen are massively customized from the original System V/System 7 sources over the past thirty years -- such that it's hard to say that they have a common core even. The only real difference is a marketing difference.

So, say it with me!

Meh.

Re:A distinction without a difference

[maswan]

And *nix in the form of, say, Oracle Solaris or IBM AIX is more restrictive than the GPL. Linux is just one branch of the unix family.

Re:A distinction without a difference

[TheCarp]

agreed, however, there are some distinctions that may be interesting here, the Unix/Linux one is really putting it in the wrong spot. The big difference is hardware. Linux runs on many different types, though, typically "PC Server" hardware, X86, whatever you want to call it.

Commercial Unix has come to mostly be, tied to hardware vending. Nobody, to my knowledge, goes out and buys an off brand system and installs HP/UX on it. You COULD do that with Solaris but generally, you are either going to buy Sparc hardware, or you are going to run Linux. Have you ever thought "I am gonna go get some HP blades and put AIX on them"?

I think their real problem is that, this sort of high end hardware sales suffers when PC hardware is at the point that people are junking physical servers because the per server utilization is under 10% and they can get more bang for their buck out of virtualization.

Justification of a new hardware system is getting harder as less workloads really need it. Sure, some of them even offer virtualization, but what do you need LPARs for when you already have virtualization technology and are training people on that? Any workload that can be done in linux on a VM is just going to go that way because it fits in better with the direction the rest of the infrastructure is going; and means being locked in with one less vendor.

Re:A distinction without a difference

[unixisc]

The distinction betwen "Linux" and "UNIX" is virtually meaningless. All of the traditional proprietary unixen are massively customized from the original System V/System 7 sources over the past thirty years -- such that it's hard to say that they have a common core even. The only real difference is a marketing difference.

So, say it with me!

Meh.

I fully agree w/ this. While the commercial and proprietary differences b/w UNIX and Linux may be there, much of the same expertise that long term UNIX veterans are familiar with would be available to users of Linux. The only things different - the differences b/w Linux distros ain't as insurmountable as the differences b/w different UNIX distros, such as Solaris and HP/UX (w/o factoring in the differences b/w SPARC and PA-RISC), the OS is not closed source like the UNIXes were, and they don't cost 2 arms and a leg the way things like AIX or HP/UX used to.

It's true that all the UNIX companies are either dead (Sun, Intergraph, SGI) or have abandoned UNIX (IBM, HP/DEC) (yeah, I know AIX and HP-UX is still alive, but their vendors no longer aggressively promote them over Windows Server). On the Linux side of things, there are Red Hat, SuSE and Oracle, while iXsystems covers the FreeBSD guys. But if one wanted UNIX, it's a lot easier to get that now than it ever was - either go w/ one of the Linux distros, or if one wants the old school UNIXes, there are still choices: the xBSDs for the BSD side of things, or OpenIndiana from the System V side of things. Yeah, a few things are dead - OpenIndiana will be similar to Solaris, but not Unixware, if that's okay w/ the user.

But I agree - it's misleading to say that 'UNIX is dead' when it implies that Windows is the only game in town. That's only true on the desktop and laptops, but not on tablets, phones & servers.

Re:A distinction without a difference

[stox]

There was no System 7, you are referring to R&D Unix Version 7, which System III was the commercial fork of. System V was really a fork off of R&D Unix Version 8.

Re:A distinction without a difference

[Guy Harris]

It's true that all the UNIX companies are either dead (Sun, Intergraph, SGI)

One of these things is not like the others, given that Oracle bought Sun and are still selling their servers and their Unix. (SGI's still around, too, but good luck buying a MIPS-based server running IRIX from them.)

Re:A distinction without a difference

[unixisc]

Did Sun ever sell either SPARCstations or SPARCservers w/ Linux on them? How about Oracle - can one buy a SPARCserver w/ OEL on it, or is OEL only available on Sun x86 systems?

Re:A distinction without a difference

[Fished]

You're right, I just got crossed up. I'm OLD and it's been a LONG time.

Overlooking the obvious

[Frogg]

The article completely neglects the fact that OS X is a fully certified Unix, and, whilst OS X might not be overly popular in the server market, it certainly has a very large percentage of the desktop market. So yeah, perhaps the old-school companies that provide Unix OSes for servers may be in their 'last days', but Apple's OS X has brought Unix to the masses via the desktop, so Unix certainly isn't going to die any day soon.

Re:Overlooking the obvious

[danlip]

And by "very large percentage" you mean 7%.

(hey, I'm a big OS X fan myself, but describing its market share like that is pretty baffling)

Re:Overlooking the obvious

[Above]

Perhaps the AC doesn't know, the Open Group, you know, keeper of the Unix specification, has a certification program. Apple participates, and has the certificates to prove it.

Mac OS X Version 10.8 Mountain Lion Certificate, from the people who own the Unix(TM) specification.

Re:Overlooking the obvious

[samkass]

OS X is a fully certified Unix

You keep using that word. I do not think it means what you think it means.

"Certified" is not a general intensifier. It has meaning. Don't sprinkle it on your sentences like some verbal MSG.

Yes, it does. Are you asserting some meaning other than having been certified as a Open Brand UNIX 03 registered product, which MacOS X has been since version 10.5?

Re:Overlooking the obvious

[cusco]

Unix to the masses??? Apple could have put VMS or the Windows core under the hood, and "the masses" wouldn't have known the difference. Or cared much for that matter. OS X is the pretty interface to almost everyone who uses it, I can count on one had the Mac users that I know who have any idea what BSD even is.

Re:Overlooking the obvious

[Darinbob]

7% of a massive market is pretty large.

Re:Overlooking the obvious

[Anonymous Coward]

Unfortunately, OS X stopped tracking the POSIX C API. There are different levels of "UNIX03" certification. The current edition of the "UNIX03" standard is version 4, issue 7. Version 4 has been out for awhile, given that it's not up to issue 7.

OS X meets v4 for command line tools, but only v2 for the C API. The *BSDs and Linux are light years ahead of OS X's C API. In a few years, OS X's C API will be as relevant as Windows', which was POSIX certified in the 1990s.

Apple in general appears to be giving up on Unix. They've deprecated much of the Unix API in iOS, including simple things like fork(), and have stated that they want to move all application developers, including desktop app developers, to the iOS API.

Re:Overlooking the obvious

[serviscope_minor]

Yes, it does.

Nah not really.

Are you asserting some meaning other than having been certified as a Open Brand UNIX 03 registered product, which MacOS X has been since version 10.5?

No, he's asserting that no one cares and it's a meaningless intensifier, because do you know what certified Unix with a capital U (TM) really means? It means certified not quite compatible with the market leader (Linux).

Re:Overlooking the obvious

[dfghjk]

But it is not a "very large percentage".

"...whilst OS X might not be overly popular in the server market, it certainly has a very large percentage of the desktop market."

Re:Overlooking the obvious

[Guy Harris]

OS X meets v4 for command line tools, but only v2 for the C API.

Same for Solaris 11, same for AIX 6, same for HP-UX 11i V3. Does this indicate that those OSes have not adopted any of the subsequent API changes, or does it indicate that the Open Group haven't updated their conformance suites so that you can't test anything better than v4 for the command-line tools and v2 for the C API?

(And do "V4" and "V2" refer to Issues 4 and 2, respectively, of the SUS, or do they refer to something else?)

The *BSDs and Linux are light years ahead of OS X's C API.

What are some of the significant things present in the C APIs of Linux and the *BSDs that are not present in the C API of OS X, and which, if any, of them were introduced in issues of the SUS later than 2?

Apple in general appears to be giving up on Unix. They've deprecated much of the Unix API in iOS, including simple things like fork(), and have stated that they want to move all application developers, including desktop app developers, to the iOS API.

Where have they stated that (in a way indicating that "all" includes not only the folks writing Mac Applications(TM) but also the people doing UN*X apps that work on OS X)?

Re:Overlooking the obvious

[unixisc]

It was there as long as the original 68k Macs were there. It was never there on PowerMacs - for people who had to have UNIX there, Apple provided either AIX from IBM, or MkLinux. And once OS X was ready, it became the default UNIX for both PowerMacs as well as subsequent x86 Macs as well. Incidentally, did Apple ever offer OS X in a server configuration?

Not suprising

[BigDaveyL]

One can get Linux or *BSD on commodity hardware for a fraction of the cost.

Re:Not suprising

[Em Adespoton]

GNU doesn't really apply to *BSD -- any BSD that wants to pay the fee can pretty much become UNIX.

So does this really mean that there are fewer licenses being sold?

Re:Not suprising

[unixisc]

Non-sequitur: all the current BSDs - FreeBSD, OpenBSD, NetBSD and any other derivatives are FOSS, even if not GPL, and so BigDaveyL is right - one could get, say, FreeBSD at a fraction of the cost of HP/UX, either on the same Itanium hardware, or on cheaper x64 boxes. For that matter, even the Linux Foundation could, if they cared, get Linux certified by the Open Group. But neither have thought that it's worth it.

Re:Not suprising

[Em Adespoton]

Non-sequitur: all the current BSDs - FreeBSD, OpenBSD, NetBSD and any other derivatives are FOSS, even if not GPL, and so BigDaveyL is right - one could get, say, FreeBSD at a fraction of the cost of HP/UX, either on the same Itanium hardware, or on cheaper x64 boxes. For that matter, even the Linux Foundation could, if they cared, get Linux certified by the Open Group. But neither have thought that it's worth it.

Exactly -- the real thing that's dying is the certification and the licensing, not the technology.

Re:Not suprising

[unixisc]

Yeah, what would be the value added for licensing today? In the 90s, it made sense, since all the vendors wanted to make that point that their OS was a UNIX edition, as opposed to being a proprietary corporate OS such as MVS, OS/400, VMS/OVMS, OS/2, or variants. Today, there is no such need - they can either go with a distro of Linux, or w/ one of the BSDs, and they'd be pretty standard. So why burn precious cash on licensing fees and other such rights, which don't have any tangible benefits anymore?

Moronic analysts

[wiredlogic]

Errol Rasit, research director at Gartner, concurs that the primary cause of Unix weakness over the past decade is migration from the RISC platform to x86-processor based alternatives, which can run many Unix workloads, usually at attractive price/performance ratios.

x86 has been implemented on a RISC based core ever since the PentiumPro. RISC won. It didn't wither away. That transition made possible a performance boost allowing Intel to compete against the home-grown processors of the traditional Unix vendors who lacked the cash to invest in fab advancements needed to match pace.

Such are the fools pandering their vaunted "analysis" to the media these days.

Re:Moronic analysts

[Em Adespoton]

Errol Rasit, research director at Gartner, concurs that the primary cause of Unix weakness over the past decade is migration from the RISC platform to x86-processor based alternatives, which can run many Unix workloads, usually at attractive price/performance ratios.

x86 has been implemented on a RISC based core ever since the PentiumPro. RISC won. It didn't wither away. That transition made possible a performance boost allowing Intel to compete against the home-grown processors of the traditional Unix vendors who lacked the cash to invest in fab advancements needed to match pace.

Such are the fools pandering their vaunted "analysis" to the media these days.

Sorry, but it didn't win. RISC didn't get clobbered by CISC or vice versa; rather, they both got consumed by VLIW. VLIW pipelining made the debate over instruction set complexity meaningless, as you get custom sets based on which pipeline is used, due to long instruction chains. You could argue that at the core of each VLIW chip you have a RISC; but you could also argue that the result is really an extremely CISC. It's kind of like arguing about Toyota vs Ford, when in reality, they both have components made by Honda and Mazda, as well as each other these days.

So Errol Rasit's observation is valid. There was a migration -- I know, because my old 32 and 64-bit RISC code is a headache to port to x64, unless it is abstracted. The current registers however handle old CISC x86 code just fine.

Re:Moronic analysts

[RamiKro]

Last I've checked, VLIW isn't even in desktop CPUs. And Intel's Itanium failed EPICly. In embedded, VLIW are purely an Intel thing and minor ARM dominance.
As for your "old 32 and 64-bit RISC code", can you post a snippet? I've only seen MIPS64 so far but I'm a little curious about ARMv7's if that's what you're referring to... Or is it something different altogether?

Re:Moronic analysts

[unixisc]

VLIW? Itanic was the biggest fiasco in Intel's history. The very premise was flawed - the assumption was that the industry would accept disruptions in software compatibility b/w each generation. Moreover, the savings in real estate as a result of eliminating the dynamic analysis units and moving them to the compiler - that turned out to save precious little. RISC remained the sweet spot b/w CISC and VLIW. Oh, and VLIW doesn't do pipelining - everything is tossed to the compiler.

RISC architectures - as defined as CPUs w/ instruction sets that were not CISC - did die away - w/ the sole exception of ARM. PPC and MIPS had some presence in the console and set top box markets, but other than that, RISC was pretty much cooked. Yeah, the x64 from both Intel and AMD did adapt enough RISC practices to make the debate moot, while in the case of Intel, being a couple of process generations ahead more than covered any theoretical advantages that RISC had over CISC.

Errol Rasit is correct if he's using the term 'UNIX' in a colloqual sense, to collectively describe the proprietary POSIX OSs of yesteryear - Solaris, HP/UX, AIX, Unixware, SCO OSE, Irix, OSF/1, Ultrix, et al. Those OSs did die when their hardware platforms lost to x86 based workstations and servers - even Itanium couldn't displace the x86, and this was Intel's third attempt at it.

Re:Moronic analysts

[Em Adespoton]

Last I've checked, VLIW isn't even in desktop CPUs. And Intel's Itanium failed EPICly. In embedded, VLIW are purely an Intel thing and minor ARM dominance.
As for your "old 32 and 64-bit RISC code", can you post a snippet? I've only seen MIPS64 so far but I'm a little curious about ARMv7's if that's what you're referring to... Or is it something different altogether?

I was actually referring to PowerPC 970 bog-standard code. You're right though; I'm mixing my desktop and mobile chipsets.

Re:Moronic analysts

[RamiKro]

It feels like a chicken and egg thing. The researcher has been asked about the decline of an operating system, and offered the change in architecture as the main reason.

It doesn't matter if Intel's x86 is technically more RISC or more CISC in this context. Or what minicomputers and personal computers were at the time. The point is that x86 out sold DEC hardware.

We could debate the facilities the architecture has to offer, and mention the abstractions x86 systems (like Windows) responded with... But all that matters is that once you try to port the system to another architecture, you end up with an inappropriate api model and a legacy of language and system design choices that were done in a different decade, for very different needs, and for completely different capabilities.

I'm told we're headed into an age of extreme parallelism and possibly non volatile MRAM. I will be mighty disappointed to see a POSIX system ported to something like that. But no doubt it will happen and no doubt it will kill off the closed source systems that can't offer effortless cross compatibility at little to no price. And big companies are just not that flexible.

Re:Moronic analysts

[amorsen]

x86 has been implemented on a RISC based core ever since the PentiumPro.

CISC has been RISC with a translation layer from the very beginning. The whole point of RISC is to get rid of the translation layer by ditching all the complex instructions which require it, then implement the rest directly in hardware.

Re:Moronic analysts

[dfghjk]

There is nothing about modern processors that are in ANY way "RISC based" at their core. If that were true, modern RISC processors wouldn't need similar architectures. Sour grapes.

"Such are the fools pandering their vaunted "analysis" to the media these days."

Look who's talking.

Re:Moronic analysts

[dfghjk]

VLIW is an embarrassing failure that was predicated on the assumption that compiler technologies would evolve to do things they couldn't. Modern processors couldn't be more removed from VLIW concepts if they tried.

Re:Moronic analysts

[Guy Harris]

But all that matters is that once you try to port the system to another architecture, you end up with an inappropriate api model and a legacy of language and system design choices that were done in a different decade, for very different needs, and for completely different capabilities.

Or, if you're just talking about instruction set architectures, you end up with an operating system whose API model works Just Fine for both instruction set architectures (and ISAs other than the one ported from or the one ported to) and a legacy of language and system design choices that work Well Enough for both.

So presumably you're referring to system architectures, not instruction set architectures such as SPARC, MIPS, POWER/PowerPC/Power ISA, Alpha, and x86, as they really don't differ, at least in the way Windows or UN*X use them, in ways that make much of a difference.

Re:Moronic analysts

[Guy Harris]

CISC has been RISC with a translation layer from the very beginning.

Or kinda sorta VLIW, if you mean "CISC has been implemented in microcode atop not-so-CISCy microengines from the very beginning", although it's also been implemented purely in hardware (in, for example, the IBM System/360 Model 75, the GE 600 series, and the Honeywell 6000 series, although that was in an era before the term "CISC" had been coined).

The king is dead, long live the king

[dwheeler]

"Unix" - as they define it - is going away. But what's really happening is that old implementations of Unix are being replaced by modern implementations and re-implementations of Unix.

Servers are increasingly using Red Hat Enterprise Linux, Fedora, Debian, Ubuntu, etc. On the client side, the #1 smartphone (by popularity) is Android, based on Linux. The #2 smartphone is iOS, based on Unix. On the desktop, Macs are running MacOS, also based on Unix.

Re:The king is dead, long live the king

[unixisc]

Not just that - even if one wants to be a purist and argue for the original UNIX, then too there are good choices. If one wants an FOSS BSD, there are FreeBSD, OpenBSD, NetBSD and Minix. If one wants an FOSS System V, there was OpenSolaris, which is now OpenIndiana.

UNIX is in no danger of dying, since theoretically, anybody who understands the code can pick it up and either continue to make improvements, or fork it.

What's this?

[joh]

Neckbeard teasing? Or what?

Unix (in some incarnation) is running the world. It runs on servers, on embedded systems and basically all tablets and smartphones (both Android and iOS are Unix).

I cannot believe I'm wasting 30 seconds on this. Die, Slashdot, die.

Re:What's this?

[Type44Q]

Die, Slashdot, die.

There's an easy solution for that (hint: your post isn't it). :p

Re:What's this?

[Provocateur]

neckbeard confirms, slashdot is dead, joh (pronounced "yo")

Idiot

[Tough Love]

This guy seems to be blissfully unaware that FreeBSD is Unix. With Apple selling millions of handsets, Unix is obviously not in decline. Just squeezed out from one role (by Linux) and taking on a new one.

Obviously, this might change in the future, but from the moment, Unix is doing to opposite of declining. Troll article. If there is a story in there somewhere, it is the rise of Linux in the server room.

Re:Idiot

[toddestan]

Maybe you are blissfully unaware that iOS is not UNIX(tm)? And given the lack of attention that Apple has been giving its computer business, OS X is going to be on the decline real soon.

Re:The 90s called, they want their Windows back.

[ebh]

Unless, of course, your main application has gone EOL, and its replacement (and all its competitors) only runs on Windows.

System V

[Anonymous Coward]

My experience was on AT&T Unix System V. I used to jokingly refer to it as REAL UNIX with a hint of faux snobbery and a straight face.

While working on a Linux system, I was using some command line utility (doesn't matter) and the command kept wrapping. Ran it - errors. Retyped - errors. Retyped - finally worked.

Anyway, a skilled Linux user was watching me, typing away and then running my command - the syntax worked like it was a AT&T System V UNIX, BTW.

Said Linux dude said, try this - and he proceed to do the same thing with the same program but with like one or two flags and then the args.

It worked.

There have been quite a few time savers (I won't call them improvements) built into Linux.

I can't blame them - some of the most common things that we did in Sys V were overly verbose.

Anyway, wanted to share that - gotta go; there's a Matlock marathon and it's Pizza and Banana pudding night! Betsy has got the hots for me and she so young - 68! I'm gonna have a GOOD time tonight!

Re:System V

[RabidReindeer]

There have been quite a few time savers (I won't call them improvements) built into Linux.

A rather wry observation that I made when working regularly with Solaris 8 was that a lot of the system commands resembled the Linux commands of a decade earlier. Before they added the extra options. Ditto the usages of the /etc directory. One of the best things about Linux is that it has virtually no binary files under /etc.

Branded Unix might be on the decline...

[Anonymous Coward]

...but if you think about it, Unix-like, to include Linux, Mac OS X, iOS, and Android, is on a meteoric rise. Especially in the ARM world.

I'd mod the OP Flamebait

[dickens]

Between OS-X, IOS and Android, this discussion is more than a little comical.

Re:I'd mod the OP Flamebait

[H0p313ss]

Between OS-X, IOS and Android, this discussion is more than a little comical.

Not really, the article is quite specifically talking about Unix. Linux and iOS and OSX are not Unix. Much in the same way that Volkswagen does not makes (horse-drawn) wagons.

Re:I'd mod the OP Flamebait

[Rosyna]

Not really, the article is quite specifically talking about Unix. Linux and iOS and OSX are not Unix.

Mac OS X is UNIX. iOS and Linux are not.

Re:I'd mod the OP Flamebait

[Guy Harris]

Between OS-X, IOS and Android, this discussion is more than a little comical.

Not really, the article is quite specifically talking about Unix. Linux and iOS and OSX are not Unix.

Depends on how you define "Unix".

If you define it the way The Open Group does, i.e. "a trademark that we license to organizations that prove, using our validation suite, that their OS conforms to the Single UNIX Specification", then OS X is, in fact, Unix, but iOS or Android aren't.

If you define it the way TFA does, i.e. "a big enterprise server OS in the UN*X family from those traditional makers of UN\*X big iron that remain", none of them are Unix. That's the definition that's relevant here, so, as you note, the discussion isn't comical in that regard.

Re:I'd mod the OP Flamebait

[H0p313ss]

Not really, the article is quite specifically talking about Unix. Linux and iOS and OSX are not Unix.

Mac OS X is UNIX. iOS and Linux are not.

The micro-kernel is not unix and the userland came from Unix.

It is unix-like, it is not unix.

Re:I'd mod the OP Flamebait

[Guy Harris]

Not really, the article is quite specifically talking about Unix. Linux and iOS and OSX are not Unix.

Mac OS X is UNIX. iOS and Linux are not.

The micro-kernel is not unix and the userland came from Unix.

As did the BSD part of the kernel (the kernel is more than "the micro-kernel"; "the micro-kernel" is the stuff in the osfmk directory of the XNU source, the BSD part is the stuff in the bsd directory, and there's also I/O Kit in the iokit directory).

It is unix-like, it is not unix.

The BSD part of the kernel, and much of the Darwin userland, is as much "unix", in the sense of being derived from AT&T code, as are the *BSDs. (It's also "unix" in the sense of passing the Single UNIX Specification test suite, but you presumably already knew that and weren't using "unix" in that sense.)

The alternatives got better

[sl4shd0rk]

I was soooo glad when we finally decommissioned our last Solaris box. It's not that Unix got worse it's just the alternatives got better. Also the proprietary RISC based hardware underpinning much of commercialized Unix lost out to cheaper PC commodity stuff. Again, it's not that RISC sucked, it's the fact that the lazy proprietary paradigm couldn't figure out how to evolve past the "Screw, em. They're locked in. They _CANT_ switch" model.

Re:The alternatives got better

[acscott]

Vendor lock-in? The interwebs made Unix or it's posterity significant: Being able to quickly respond to the needs of the many made it relevant again horizontally as well as vertically. The openness fostered a new "software age" that was aligned with network effects created by the telecosm of the internet. The price of efficiency made it cost-prohibitive to effectiveness: the joy of creation, creativity was sparked and found new soil where evolutionary pressure weeded out the lack of joy in freedom. Where collusion failed, cooperation succeeded. Subtle difference just as a program is sensitive to a misplaced byte.

UNIX - GNU ?

[pD-brane]

The decline of UNIX is, among other things, attributed to "the abundance of Unix-specific apps that can now also run on competitor's servers."

There is this thing called GNU, which has the explicit goal to replace UNIX. So it is not that Unix-specific apps can only run on other systems, the whole system is replaced. And though largely backwards compatible, improved as well (and free of course).

The abundance of Unix-specific apps

[wisnoskij]

"the abundance of Unix-specific apps" is killing Unix?

NetworkWorld sensationalist headline.

[Bill_the_Engineer]

IDC Enterprise Server Group simply reported that revenue generated from the sale of commercial Unix is declining and IBM is overtaking HP in that segment of the market.

NetworkWorld took that little factoid and turned it into "The last days of Unix" article despite the fact that the actual article mentions Linux as being a competitor. I'm sure BSD is also taking a good chunk of market too.

Re:*NIX is most dominant now

[Em Adespoton]

Windows is POSIX compliant, although not a certified UNIX. Other than Linux/GNU, there are only a few niche OSes out there these days that AREN'T (or couldn't be, if someone applied for certification) UNIX.

B'cos of the CPU?

[unixisc]

Replacing HP/UX with Windows? Do they have any CPUs in common any more? It's only possible if one is replacing HP-9000s or Integrity servers with x64 servers from either HP themselves, or someone like IBM or Dell.

HP/UX only ran on PA-RISC previously, and now only runs on Itanium. Server 2008 did run on Itanium, but is no longer supported there: Server 2012 is a Wintel only platform. So if one wanted to replace HP/UX with Server 2012, then one would have to move out the Itanium servers, and bring in x64 servers as replacements. Long term, not a bad idea, but short term, it would be a major loss on the Integrity servers, which ain't cheap

UNIX has be dying slowly for 20 years.

[Anonymous Coward]

The real decline is that the group that produced the source code product at AT&T was effectively disbanded in 1993 when AT&T sold off UNIX System Labs to Novell. Some of the slack was picked up by SUN, HP, and others, but in the past 20 years, none of the features and improvements that were planned--at the source code distribution level--ever happened.

Hold on....

[LoRdTAW]

The Steady Decline of Commercial Unix - FTFY

Most of the the big Unix vendors have either switched to Linux or offer Linux as an alternative (eg IBM). Apples OSX since Leopard has received official "Open Brand UNIX 03" certification. iOS is not mentioned and most likely is not certified as the certification is unnecessary. But iOS is still based on OSX which is Unix certified and before certification, Unix like. Open Solaris was the only truly open source Unix but Oracle put a stop to that. Now OpenIndiana and illumos have replaced them and I don't believe they can carry the Unix brand.

Unix like operating systems such as GNU/Linux, and to a lesser extent, BSD have replaced commercial Unix operating systems. They both provide two of the most critical parts of Unix: POSIX and X windows. From there many programs originally written for a major commercial Unix vendor be it IBM's AIX or SGI's IRIX can quickly be ported to Linux or BSD with minimal effort. Just look at what Linux can run on:
* Embedded systems with tens of MHz and a few megs of ram to the worlds largest supercomputers with thousands of nodes.
* Just about every every high powered ARM embedded electronics hobby board runs Linux such as the Raspberry Pi, Beaglebone UDOO and others.
* Linux is also pushing into hard real time markets previously dominated by QNX, LynxOS and VxWorks. National Instruments now has an ARM version of their CompactRIO running real-time Linux. Previously they used an embedded Power CPU from Freescale running VxWorks.
* The Linux kernel is the foundation for Android which is dominating the smartphone and tablet market.

Re:Hold on....

[amorsen]

You cannot get iOS Unix certified as you cannot even get to something which will run a standard /bin/sh script, and the Unix C API is unavailable. You can jailbreak iOS and install a Unix environment, but out of the box it is not a Unix.

Mod down original article

[tekrat]

Never mind the submitter, or Slashdot for even carrying this story.
Mod the Original Article as flamebait.
Whoever even bothered to write the article in the first place needs to lose his license to write tech journalism. Author is clueless. Sure, *pay-for* unixes are dying. HOWEVER, Free/Open Source Unixes are thriving.

Re:Mod down original article

[yelvington]

Whoever even bothered to write the article in the first place needs to lose his license to write tech journalism.

If you start applying standards to tech journalism there won't be any left.

Nonsense

[gweihir]

While true UNIXes are in decline (a true UNIX shares source-code with the original UNIX), clones (Linux) are very much alive and that is what counts. Even some true UNIXes (free/open/netBSD) are not doing too badly. There are even more interface-compatible systems that follow the UNIX philosophy. In a nutshell, the only "OS" today that is not UNIX-like and matters is the Windows isle of incompatibility.

Re:Nonsense

[rubycodez]

BSD "not doing too badly"??? hah that stuff is everywhere in embedded systems from appliances to elevator controls, most people probably have some BSD running devices at home and not even know it

Re:Nonsense

[gweihir]

I know. Visibility of the xBSDs is not so good though. But they are definitely not going away.

An Un-Story so Move Along, Move Along

[mendax]

As several people have already noted here, this story is essentially a lie, or at least an exaggeration. Linux (for all intents and purposes) *IS* Unix without the trademark. That is one of the reasons why Linux grew to be so popular. A large number of people wanted a Unix but didn't want to pay for it. Just because the operating systems that can legally be called Unix are shrinking in usage (with the notable exception of MacOS and it's close cousin iOS) does not mean that Unix is dead. Unix and that which would be called a Unix in a trademark-free world is alive and well and is exploding exponentially. It's in every pad computer, e-book readers, most smart phones, my Sony Blu-Ray player, airliner entertainment systems, and many, many other places I can't think of at the moment.

It is very simply to explain: Money

[rabun_bike]

Once you could replace a Solaris box with 10 Linux boxes and the business finally believed Linux could run their business servers and not just the printer spooler the game was over. This started really happening 12+ years ago and continues to this day. The same goes for HP-UX, AIX, SCO, you name it. These machines were elegant, high performance devices in many cases (expect SCO) but they were unbelievably expensive. Some of the hardware was exotic such as the Digital Corp. and SGI hardware. Just as counting machines were displaced by Mainframes and Minicomputers, the Mainframes and AS400s were replaced with UNIX, and the UNIX machines will be replaced with x86 / Linux machine which may one day be throw out by ARM / Linux machines. It will not be unprecedented.

I have heard from many admins over the years that they loved the tools, loved the support, of the UNIX vendors but the reality is the cost. Google had no real desire to build their search technology on commodity gear but when you look at the scale and cost the math just doesn't add up to use anything else. This transition has only accelerated with VMware and the Intel and AMD chips with vitalization acceleration embedded, and commodity servers with staggering hardware performance. The market has matured, Linux is accepted, and the scale of deployment in virtual (cloud) environments makes total economic sense. The next step are other pieces of the infrastructure that I cannot commit on due to NDA. But it is coming very quickly.

In Engineering - Unix is nearly done

[toast-]

In the Engineering CAD world, Unix has nearly run its course. All companies have dropped Unix support for the newest versions and only some maintain Linux/OSX versions for newer unix-like machines. Most are Windows only. Automotive companies, which are notoriously slow in technology adoption have mostly abandoned UNIX

Ford will retire their UNIX workstations (HPUX) for suppliers and customers in February 2014. These are largely HPUX 11.11i.
Unigraphics NX stopped UNIX support (HPUX, AIX, etc) as of NX 6 but opened support for Linux and OSX as of 8.
Dassault systems CATIA supported HPUX, AIX (6.1+) and Solaris on V5 - but as of V6 in 2011 they have ended UNIX support and are Windows only.
Pro Engineer quit most UNIX except Solaris until Pro Engineer / Creo 4.0 - at present they are Windows only.

Re:In Engineering - Unix is nearly done

[ShoulderOfOrion]

IMO there are two primary reasons for this 1) high-end CAD software is incredibly expensive and has licensing schemes that work better in the Windows ecosystem and 2) the 3-D graphics drivers for Linux, particularly OpenGL, suck big time.

I do wonder if porting to Windows will come back to bite the companies in the future, as Microsoft seems to be intent on destroying Windows via the touchscreen interface--an interface which is anathema to most CAD engineers. Then again, Gnome seems to be following Microsoft down that abyss, and I don't see any sign that the Linux OpenGL drivers will be getting better any time soon.

Re:In Engineering - Unix is nearly done

[Guy Harris]

As others have pointed out, OS X is UNIX.

And it's one of the two UN*X platforms that most of the CAD software listed in the Wikipedia "Comparison of CAD software" page runs on if it runs on any UN*X at all; the other is, not entirely surprisingly, Linux. Most of them, according to that page, at least, only run on Windows, however.

Re:In Engineering - Unix is nearly done

[Guy Harris]

IMO there are two primary reasons for this 1) high-end CAD software is incredibly expensive and has licensing schemes that work better in the Windows ecosystem and 2) the 3-D graphics drivers for Linux, particularly OpenGL, suck big time.

And 3) the traditional UNIX workstation vendors stopped making UNIX workstations, so "engineering workstation" now means "high-end PC", and, of the UNIXes from the traditional UNIX workstation vendors, the only one that runs on PCs is Solaris.

Re:In Engineering - Unix is nearly done

[lexman098]

Good thing I went into digital IC design then. Our CAD vendors don't even support Windows at all. Long live Linux for getting real shit done.

winner

[Tom]

Most ridiculous article ever.

Unix is running the world. There's a stupid, fragile, overpriced PC OS from some small indie software maker in some backwater place near Seattle, but everything else is Unix. Whether it's Linux, OS X, BSD, iOS, Android, or one of the surviving big Unix players, there's really not all that much out there that isn't a Unix.

Heck, even most of the RTOS are Unix or Unix-like these days. Gaming consoles seem to be the only area of computing not dominated by Unix. Funny, gaming is also the only reason I still have a Bootcamp partition. :-)

Re:winner

[rubycodez]

As rabid a Unix/Linux/BSD fan I am, I'd still maintain IBM and Unisys mainframes run the world, because most the world's money is in them. Outside of that, it's largely Unix all the way down from supercomputers to network appliances.

Re:winner

[Tom]

Contrary what they want you to believe, money is largely irrelevant. If it would all disappear tomorrow, we'd have a few uneasy days, and quite a bit of chaos, but unless there's a mass panic, everyone would pretty much just keep on going until someone figures out what to do about it, and so nothing much would really change.

But if all the major embedded OS were to crash tomorrow and refuse to reboot, most of our world would come to a halt.

Or in other words: If Unix became sentient and moved away to another galaxy, we'd be fucked. Your phone wouldn't work, half of the world's computers wouldn't work, but you'd hardly notice because all the others are down as well because most of the power plants would be gone, along with most of the factories, transportation networks, traffic controls and probably the water and sewage system, too.

Re:So the old timer Unix unix, not the new time *n

[unixisc]

I'd hardly describe FreeBSD as dying off. Dunno about OpenBSD or NetBSD though

Re:So the old timer Unix unix, not the new time *n

[rubycodez]

the number of embedded systems running openbsd or netbsd (from everything from printer to elevator controllers) goes into the millions

so ubiquitous they're invisible

No surprise

[maliqua]

solaris ruined by oracle who even cares if it exists anymore, HP-UX has been a disaster longer than i have been alive, SCO decided to burry itself IRIX was too niche oriented.

UNIX isn't losing half the team didn't show up now they have to forfeit

Re:No surprise

[ediron2]

I hate Oracle as much as the next geek, but Sun really shit the bed all by themselves w/r/t solaris, before Oracle bought them.

Full disclosure: I adore ZFS. I hop between Debian and Solaris and OS-X and windows. I just could never make a compelling economic argument for sun boxen after about 1998, and certainly not as desktops to keep people well-versed in the idiosyncracies of that specific *nix flavor.

Re:Darwin doesn't count?

[Guy Harris]

Gotta be some metric boost there, right?

Not in the enterprise server market, which is what TFA was talking about.

MAC anyone?

[erp_consultant]

Sure, pure UNIX use is declining on servers but OSX is UNIX based and that seems to be doing fine. Linux is too. I think that the decline of UNIX on servers is more about the architecture it is tied to rather than UNIX itself. Linux, of course, can run on Intel CPUs which tend to be much cheaper. UNIX has always been a rock solid OS in my experience and it will continue to live on, behind the scenes, for a long time to come.

Depreciation

[sjbe]

Thing is, its amortization schedule was roughly 2 decades at least if I were to guess.

Just being pedantic here but properly speaking it would be depreciation, not amortization. Amortization is for intangible property. Depreciation is for tangible. And under MACRS depreciation schedule (typically required for taxes) most if not all of the depreciation would have occurred in the first 7 years.

Re:Learn to read

[mendax]

Uh, I beg to differ with you. While Apple did kill its Xserve line of rack-mounted computers, it indeed still markets server computers. Apple offers a server configuration of the Mac Mini. Mac Minis may not have been designed to be servers, they make great servers, can be rack-mounted with a bit of third-party equipment, and they are much, much cheaper than purpose-built servers. You get more bang for the buck. Apple also pushes the Mac Pro as a server despite its size and lack of pizza-box shape.

The real subject is...

[jcdr]

The last non-UNIX OS share is declining.

IDC is simply ordered by the last non-UNIX OS (Windows) owner (Microsoft) to redact some PR to distort the reality. There solution is to count only the past generation of UNIX and to ignore all the new generation of UNIX, like all the Linux distributions, all IOS, all Android, almost all top supercomputer on the planet, the vast majority of routers, the vast majority of recent TV, the majority of web servers and data centers, and a lot of more specific applications and embedded systems.

Where it counts...

[phreakincool]

Websites, databases, firewalls, DNS, application server, file-servers - most businesses run some form of UNIX or Linux. And wouldn't dare think otherwise. This article is stupid. I work for a company where we would literally lose millions for every reboot of certain mission critical servers. UNIX ain't going nowhere.

Re:Uh huh - Mainframes

[riverat1]

They said back in the 1990's that mainframes were on the way out yet IBM keeps pumping them out. I don't see Solaris disappearing from our operations until long after I retire.

WHOA!!!

[madcat_sun]

Seriously this is one of the greatest lies I ve ever seen!!!!! Someone pay to post here??? What happen?!!!!!! Even if this is not certified unix, linux kernel/ ios kernel are sons of unix, as well the HPuX, the IRIx and the solaris are.

I wish

[iamacat]

Doesn't the world deserve a modern OS after 50 years of legacy? All the files can be memory mapped, pipes can transmit objects with rich behavior rather than just bytes, /bin/sh is overdue for a state of the art replacement with optional GUI, GPU-based computing should be deeply integrated into everything.

And of course legacy stuff can be still made available with high degree of standards compatibility, but discouraged for new projects.

Re:I wish

[Guy Harris]

Doesn't the world deserve a modern OS after 50 years of legacy?

Which OSes that were available in 1963 are still available (or have a direct descendant that's available), other than the Burroughs MCP?

Re:UNIX LIKE != UNIX

[Guest316]

BSD is a fork of original AT&T UNIX, so yes, it still has some claim to the title. Linux, however, is only a UNIX-alike, and rather a kludgey mess of one at that.

Trouble first started when loud advocates started promoting Linux as the bestest thing evar, saying they wanted to replace MSWin in the server room, yet managed instead to displace commercial UNIX. Then corporate agism took its toll, as more experienced admins were replaced with fresh-faced kids whose only experiences were running MSWin and Linux at home and feeling that this gave them all the knowledge they needed to be enterprise admins. This lead to IT being relatively handy with Linux and abysmal with anything else, and in turn those non-Linux systems not being administered correctly and failing when they wouldn't under competent administration.

The fact that so many posters here are indignantly declaring Linux to be UNIX just speaks to how bad the situation has gotten. Can't say I regret having retired a few years back when I see that this is the environment I'd be facing if I were still working in tech today.

Dinosaurs

[goodmanj]

The dinosaurs didn't die out. Many of them went extinct, but the rest evolved into birds, a group which is more diverse, ubiquitous, and arguably more interesting than the dinosaurs ever were.

Unix helped OS X adoption in higher ed ...

[perpenso]

Unix to the masses??? Apple could have put VMS or the Windows core under the hood, and "the masses" wouldn't have known the difference. Or cared much for that matter. OS X is the pretty interface to almost everyone who uses it, I can count on one had the Mac users that I know who have any idea what BSD even is.

Its underlying Unix nature and compatibility help OS X adoption in higher education. This is why the console and X Window system were part of the standard install rather than optional installs.

Not all Unix users know or care about [insert Unix flavor here]. They have a specialized app, they know the app, the don't know or care what is hosting the app.

Re:Z series eats the midgets (HP Sun etc)

[Guy Harris]

To boot IBM now lets one run Linux on mainframe class hardware (the Z system). Actually the interesting trick is IBM in the mainframe line has run microcoded machines since at least the 370 line.

Since the 360 line - all but the Model 75 and Models 91, 95, and 195 were microcoded.

As a result to get a new architecture, you just write new microcode.

You could add new instructions with new microcode, but you aren't going to get Shiny Newer Faster Machines (at least not past a certain point) with new microcode on top of the Same Old Boring Hardware, and, with Shiny Newer Faster Hardware, you're going to need Shiny Newer Microcode to run atop that hardware to implement the instruction set. (And, with recent generations of hardware, most of the instructions are, I think, implemented in hardware, with some instructions implemented by trapping to PALcode^Wmillicode.)

iHonk

[Tablizer]

drop in the number of people who use Kleenex when they blow their nose

Merge the tissue and the OS: Kleenix

I blew my nose with Winex, but got a Blew Screen of Death.

The amount of nines

[dutchwhizzman]

The amount of nines in UNIX claims is too damn high!

Really, to get more reliable than Linux with your "UNIX" system, you need to do an awful lot of things very good. Doing the same things good on Linux will probably get you very similar availability statistics. When you are designing a high availability platform you need to design software, middleware, OS, iron, networking and storage to work as an integrated system with fail over and load balancing on all levels. Relying on a single box with a single OS instance on a RAID-5 disk set isn't going to get you the numbers you are after. It doesn't matter who makes the hardware or the OS, it's going to fail and it will hurt your availability.

Linux has for many years been just as reliable as the traditional UNIX flavors. The real difference is in the people that make all the mistakes and shortcuts that will hurt availability. With traditional UNIX, everything used to be sacred and partially because of the costs, a lot of investments were made into solid operation and design of systems. "Oh put it on Linux, it's not important and it's cheaper. Just use some old written of Windows server hardware for it, it'll take the load just fine." has been the mantra of many shops for a lot of years. This sort of attitude would result in Linux getting lower reliability numbers. Not because of the OS, but because of the way it was deployed. Once more and more companies started to use Linux seriously to get a competitive edge, or just because they didn't have the budget for the legacy UNIX stuff, it turned out it was actually just as good.

I remember taking about 10% of the local web hosting market 14 years ago with a linux-only company. All the competition was running on IRIX and Solaris and just couldn't match our prices. They had big money behind them and weren't making a profit, we were. The entire company was only funded once at the start and the entire growth was funded from profits. You can't do that if the OS isn't good enough to work competitively. We did about 99.8% on single instance web servers and we probably would have done 99.9% if we didn't have a single uplink causing several hour long down times a year at the time. This was 14 years ago and it's only gotten better.

NASA and several other space agencies are using Linux for mission critical applications. I've worked on a satellite data crunching platform that had 6hr service windows that would lose measurements forever if that window got missed or data wasn't backed up after processing within 24 hours. Due to cost, there was no spare hardware and storage was on RAID5. Some servers were single instance and failover was manually. This platform is now over 10 years up and running and it hasn't missed a single bit of data.

These examples were from before there was "enterprise grade" high availability built into Linux distributions. It was from before raid6 or iSCSI storage networks and such. Linux has had enough reliability for a long time and it's gotten better and better. The amount of nines in the claims of traditional UNIX has little to do with the OS or the hardware, but much more with the way you work with it. Lately, I've seen expensive traditional UNIX systems behave like an unpatched windows95 PC. Single instance, EOL hardware, EOL OS, EOL software will do that to you. Spontaneous reboots, no vendor support, data loss, intermittent failures, the works. Even though they bought everything "first class", it was giving them all the trouble that Linux was supposed to give you if you believed the FUD the traditional UNIX vendors would like you to believe. It's not (so much) about what you buy, but it's about how you use it that matters most.

Re:The amount of nines

[fatphil]

Most commercial providers of linux virtual servers offer less than 2 nines in their SLA.

And most of them *fail* to provide that many months in the year. They just rely on the fact that nobody's going to sue them for failure to provide their contractual obligation.

Your satelite anecdote data sounds similar to what I look after. However, 6 hours would be a luxury most of the time. (I'm trying to work out how to schedule replacing several vital servers currently, we don't have the budget for full redundancy, and I'm targetting 30m as the downtime. It's gonna be a stressful evening.)

Re:The amount of nines

[Blue23]

I think I agree and disagree with what you say.

In the "traditional Unix" department, I'm strongest with AIX. It's IBM's flavor,a dn it only runs on their hardware. They do a LOT to make the HW and SW work hand in hand. They have a lot of options to prevent problems or reduce them to non-disruptive (for the end user). They have a lot of admin options that don't require reboots. For virtualized systems the default setup had "dual VIOS" (Virtual I/O Servers) so you can muck with them including rebooting while keeping your virtual machines ("LPARS") purring and happy. If you need big boxes, especially if you are scaling UP instead of OUT, they'll give you a level of uptime that's better than linux can. Becuae (a) less administrative disruptions, (b) less unplanned outages causing disruptions, and (c) only runs on dedicated hardware so it's avoiding all the extra issues that come with that.

However, if you're got an application that scales OUT, Linux can give you better total uptime simply because between hardware and OS licensing you can afford to have more of them. If I'm looking at 50 IBM POWER systems running AIX, vs. 120 linux blades or pizza boxes, and taking any one of them out still leaves your apps available with just a little hiccup to the users who were on that host, that's hard to beat.

These examples both assume that you know what you're doing. Anyone can muck up a machine, being more expensive doesn't protect against that.

Redhat

[LQ]

Our company sells software products for *nix platforms. Five years ago we only supported HP and Sun but started looking at Redhat. Now the majority of our customers are on Redhat. Costs are lower and performance is so cheap these days that you don't need fancy hardware. And the "you really ought to port ot AIX" lobby has totally died out.

Re:Redhat

[unixisc]

How difficult it is to port to, and support the different major distros, such as RHEL, Debian, Gentoo and Slackware? How about FreeBSD and NetBSD? Are there differences that would cause something developed for and ported to RHEL not to run on, say, Mint?

Re:Redhat

[LQ]

How difficult it is to port to, and support the different major distros, such as RHEL, Debian, Gentoo and Slackware? How about FreeBSD and NetBSD? Are there differences that would cause something developed for and ported to RHEL not to run on, say, Mint?

It's probably not difficult at all but it's just that our products are tested, delivered and supported on a small set of platforms. Can we run full system test on every distro?

Re:Redhat

[unixisc]

Not every distro, but how about the major big ones - and let the others that are derivatives simply follow? Like build it on RHEL, Debian, FreeBSD and NetBSD, and then let any derivatives of these be accountable for any incompatibilities/differences. Like say you support it on RHEL, don't bother testing it on OEL or CENTOS. If you support Debian, don't bother about KNOPPIX or Ubuntu. If you support Slackware, don't bother about Salix. And so on. Essentially, the major ones, and then you can avoid supporting their derivatives, which is what the bulk of them are.

Re:Learn to read

[Guy Harris]

Uh, I beg to differ with you. While Apple did kill its Xserve line of rack-mounted computers, it indeed still markets server computers.

Then, given that, in the sentence before the sentence you're presumably disagreeing with, he said "...this article is confining itself to analysis of the "enterprise server market."", presumably he should have said "Apple does not market enterprise server hardware.", which is certainly true.

Re:Uh huh (Metamods read this!)

[evilviper]

WOW. If my comment isn't squarely ON-TOPIC, I don't know what is.

Story is "The Steady Decline of Unix" and I'm talking about the Unixes that have lost substantial market-share.

Lovely.

I call BS!

[EmagGeek]

Did anyone bother to check to see whether Netcraft confirmed this?

Ambiguity !== Sensationalism

[tomxor]

The annoyance and confusion displayed in replies to this article all stem from the fact that mentioning 'UNIX' on it's own without clarification can be ambiguous. Though a reasonable meaning can be assumed in this case. So here is some clarification in a digestible order for those unaware of the complexity involved in the use of the term... For the short version skip to the bottom.

UNIX

UNIX is the name of AT&T's original OS and is unofficially used to refer to it's many source derivatives through commercial vendors. However most of these can also officially be called UNIX, see POSIX bellow...

UNIX is a trademark that is currently owned by The Open Group. This organisation maintains the POSIX specification, and certifies systems that conform to the specification allowing them to officially use the UNIX trademark. The purpose of POSIX is to maintain compatibility between Unix variants: It is possible for any system (source derivative of AT&T's or not) to comply to this specification, become certified and use the UNIX trademark.

So in the strictest of sense UNIX refers only to POSIX compliant systems (This is at least what The Open Group would like), this will mostly include systems that are also source derivatives of AT&T's Unix. However it also includes those that are not, a usefully complex and deceiving example is Apple's OS X base system "Darwin" which is POSIX compliant and can be officially called UNIX (and has been)... It would be easy to assume that this is through some kind of source inheritance (which it isn't even though it is... read on): Darwin like most has a complicated ancestral tree, but can be described loosely as a combination of FreeBSD, NextStep and the Mach Kernel (XNU), each with their own complex history. Looking closer at FreeBSD this is where the potential inheritance appears, however firstly: FreeBSD is Unix-Like and does comply with POSIX, secondly: it is not a source derivative of AT&T's Unix (even though it is)... just to prevent follow up posts here's why: The three famous BSD's are all derivatives of 386BSD, the precursor to this short lived BSD was an attempt to create a system functionally identical to AT&T's but without any of AT&T's source, thus being free, they replaced almost all of the files and then removed the remaining to produce 386BSD. However! AT&T's Unix and variants have all used various portions of 386BSD and it's derivative's code many times, which means they share source but not the way around most people think, which is why FreeBSD is free or course... This is the part that gets many people confused about FreeBSD and why many end up calling it UNIX, many official POSIX compliant UNIX variants (including AT&T's) contain code directly from FreeBSD, however it came from FreeBSD, and more importantly FreeBSD is not POSIX compliant... ok i think i've hammered that in enough now.

UNIX-Like

As mentioned above a system must conform to the POSIX specification to use the UNIX trademark, those that do not but have similarities are bundled into the UNIX-Like category, these should never be called just "UNIX"... not because it's infringing on the official trademark, but because it doesn't conform to the specification and creates confusion, a line must be drawn somewhere, that's the whole point of POSIX. Linux doesn't conform to POSIX and it doesn't try to, it has it's own specification, there are many similarities in the spec and that is why they are called UNIX-Like along with FreeBSD, Minix et-cetera. Unfortunately they are quite commonly referred to as UNIX anyway, and this ads to the already confusing scenario above.

The Steady Decline Implicitly

The article means UNIX (as in derivative and POSIX) on the server platform where it has long been, for the few that are not commonly used on server platforms can be ignored due to context. Linux and co are not included, that should be clear by now.

Re:Ambiguity !== Sensationalism

[tomxor]

What does being grammatically correct have to do with being knowledgable in a specific part of computing history. For your information I know the difference and i've always had difficulty remembering to only use the apostrophe on one, if you want to discredit me entirely on that basis then i'm not interested in your opinion. I'm not an expert but the guy who assembled this is... would you discredit him as an expert if you found an inevitable grammatical error in his writing somewhere. I hope you understand rhetoric... do not reply otherwise i will assume troll... i also hope my excessive use of ellipsis do not melt your grammatically pedantic brain... or do i... go away useless person

Um...

[jschmerge]

Linux *is* Unix.

WHOOOOOOSH!!!!!

[bigsexyjoe]

WHOOOOOOSH!!!!!

Most notably, confirming that BSD is dead. I think you might be onto something with "just another bad internet meme."

Re:so long aix

[Guy Harris]

Sure there is still iSeries and AIX running on Power, but remember just a few years ago iSeries didn't run on Power.

Where "a few" means "18"; AS/400, as it was called then, switched to (an extended form of) PowerPC in 1995; "Power ISA" is the current name of the instruction set architecture formerly known as PowerPC.

Re:As if ...

[unixisc]

Only if you are using the term 'UNIX' in the same sense as the article, and not including projects like OpenBSD. Otherwise, the various *BSD projects do have checks & balances to make sure that they're not checking in bad code. And while they may not have leaders as aggressive as Linux, Theo has quite a matching reputation in his own right

Re:*NIX is most dominant now

[unixisc]

When did POSIX and OpenGroup merge? POSIX was, and still is, an IEEE standard. The Open Group was an organization formed by merging 2 consortiums - X Open and OSF. The first owned the X Window standard, and the second owned the competing UNIX standard to USL and Sun.

Re:so long aix

[unixisc]

Well, IBM can always run Linux on POWER/PowerPC, since they don't make money selling AIX, they make it selling POWER servers. So it doesn't matter what OS is on it. IBM could take RHEL, put it on their POWERs and have their software - Tivoli, Pure, et al just run on that. And they could either let Red Hat maintain the OS, outsourcing that job to them, or like Oracle, they could do their own distro. So Kruemcke is right that it's irrelevant to IBM whether Linux beats AIX or not. POWER can still have a significant edge over Xeons.

Re:Actually ....

[toddestan]

So if 7% is "HUGE", then what words do you use to describe Windows' marketshare?

There's a difference?

[doccus]

There's a difference between OSX , QNX and all the other "NIXes? Other than tweaks here and tweaks there, and different GUIs, seems to me that UNIX is really having it's biggest boom ever....

read the article for a laugh

[tom arnall]

these people are gobbledygook merchants. the article is the most mindless thing i've ever read.

Why?

[Dissident]

I know why people write silly articles like this but why are we wasting our time discussing it here? :)