Guest post by InfoWorld’s Tom Yager writes in favor of Unix in IT, which has been increasingly losing ground to Linux. “Unix matters for a reason that escapes analysts’ notice,” Yager writes. “It’s that little circle with the R in it.” Asking whether IT would rather have a vendor’s promise to interoperate with competitors’ systems, or a contract obliging them to, Yager stresses the importance of The Open Group’s registered trademark of the Unix 03 spec. “The trademark provides IT organizations that need to be sure, without need for digging, that Unix means something, and it does. It means that Unix enterprise solutions work and work together, without regard for the brand on the hardware” — a guarantee of interoperability that is the “product of cooperation among Unix vendors, IT operations, universities, and professional organizations.”
(nt)
The article has nothing to do with “UNIX”, it was written by some Apple fanboy/astroturfer.
The author doesn’t take a paragraph to state that he takes offense if someone insituates that Macs are not PCs(according to him a proprietary platform
).
I thought that PC meant Personal Computer, a word that was in use before IBM released its IBMPC, which much to IBM disgrace was an open platform. Unlike the proprietary Macs.
So that you know, Linux is crap because it hasn’t paid for an UNIX certification.
It doesn’t matter that you have written your code adhering to standards. If you don’t pay for your certification, your system isn’t really UNIX-compatible.
By that logic their system was crap before they paid for one, and the next system they release will be crap until they bribe the UNIX trademark holders again(This is how it works). But don’t let logic get in the way of Apple.
PC… a Proprietary system… LOL
Good one.
When Tom Yeager says:
“[ The Big Three became the Big Four in Summer of 2007, when Apple’s Mac OS X Leopard attained Unix 03 certification. See the InfoWorld Test Center’s reviews: Mac OS X Leopard: A perfect 10 and Leopard Server: The people’s Unix. ]”
he has failed to realise that Mac OS X is based on Darwin, an open-source project based on other open-source projects before it. Darwin is a result of cherry-picking from various open-source projects: FreeBSD, Mach 3.0 and a myriad of others… the two prominent ones being University of California Berkley, and The Free Software Foundation.
How in anyone’s name can this open-source Unix-like operating system distribution equate to a bona-fide UNIX® release? UNIX, as in The UNIX System V Release 4 (5?) codebase that has recently been declared owned by Novell Corporation (ie., they own the copyrights), is a codebase that Apple has nothing to do with since Apple are not licensees of the software. It’s the reason why Mac OS X Server 10.5 costs AUD$598.00 rather than AUD$15980.00 or something.
So from this point on in Tom’s story, you can probably stop reading, since his comments most likely don’t apply. Just because Mac OS X 10.5 (or more strictly, Darwin 9) has UNIX ’03 certification, doesn’t mean that it is a UNIX system—it only means that Mac OS X has reached a level of interoperability on par with other UNIX or Unix-like systems that have achieved the same certification.
Nothing more… and nothing less.
Tom adds, “IBM, Sun, HP, Fujitsu, and Apple also sell Unix.”
Repeat after me… Mac OS X is not UNIX. IBM’s AIX 5 is a derivation from a licensed System V Release 4 distribution. Sun’s Solaris is also one. Dunno how Fujitsu got mentioned… they actually run Sun’s Solaris! And what about HP’s HP/UX 11? And SGI’s IRIX 6.5, before it got shelved in 2006?
Mac OS X is not one of them. Rather, it belongs in a world of its own, being the only open-source distribution (ie., containing no licensed System V Release 4 code) which has attained UNIX ’03 creditation. Note that other open-source projects could have acquired creditation if they really wanted to… but creditation costs engineering time and money… something the open-source community prefer to do without.
I must admit that I just did a simple 15-minute scouring to come up with these hindsights, and I am sure that my research is far from perfect. But surely a supposedly leading IT storywriter (I refuse to say “journalist” here, since journalists are rare these days) could do a tonne more research than I just did and come up with more reliable facts and figures than me!
To the Editors of Infoworld… you’re better off hiring me if this is the best you can do.
Unfortunately what you are saying (even if it is only due to technicalities) is likely incorrect.
Novell transfered the trademarks to the “UNIX” name to the OpenGroup in 1993, and sold the source code (or more aptly the rights to use said code) to SCO. So in all actuality the “UNIX” name belongs to the standards organization that… *drum roll*… provides the UNIX’03 Certification. Given that… once having been certified, OS X does quality to use the label “UNIX”, if for the only reason that it implements everything required by the standard.
Additionally, while loose at best, both BSD and AT&T System ‘X’ were both derived from Version 7… so the argument that some “free”/Unix-like systems are not “UNIX” in general terms is somewhat deceptive.
That said… the certification itself is a big deal and at the same time not really a big deal… It provides a great deal of “assured” inter-operability with other systems… if only people cared enough to code to the standards.
The myth of the Unix ‘standard’ died long ago when we got a dozen completely incompatible Unixes that destroyed any chance that it had of being taken seriously. The proof of the pudding is in the eating, and it failed. The Unix ‘standard’ is worthless.
Why is it worthless? Well, it would be nice if developers actually wrote for ‘Unix’. They don’t and they didn’t, but even when they did they found out that different Unixes behaved differently. They wrote applications for the platforms and the userspace that sat on top of it, and therein lay the rub. The day you see a Mac application being installed and run on Solaris or another Unix is the day Satan skates to work.
Tom Yager is a moron and sometimes no amount of history and being hit over the head will cure some people. I’m not even going to bother with the rest of the inaccuracies and outright lies in the article. I’m just explaining why Unix failed.
Edited 2009-03-26 21:29 UTC
After all those years, and all those attempts at unification, the thing that was finally able to unify Unix was the de facto standard called “Linux”. And even so, Unix came in kicking and screaming.
BTW, while I don’t normally do the Ghani quote thing, because it unwisely treats victory as a foregone conclusion, I will note that the case of Solaris really does parallel the quote. We’re currently in the “Then they fight you” phase. Though if the IBM deal works out, we may be about to win that battle. By that, I don’t mean Solaris needs to go away for it to happen. IBM adding a GPLv2 license to certain parts of OpenSolaris code would be a more satisfying win.
Edited 2009-03-26 21:40 UTC
I don’t even think licence change with OpenSolaris would matter – when it comes to trying to get things into OpenSolaris – its next to impossible. It is the development process that is at fault, not the licence. You can change the licence as many times as you want but unless you address the fundamental flaws of the development process, the same mistakes will keep being repeated.
Linux is successful because it has leadership from the top down; when I look at OpenSolaris it appears to be little more than a bunch of headless chickens running around bumping into each other. Where is the long term plan for OpenSolaris? what do they want to achieve? where is the plan laid out to get them from where they are now to the destination?
To be honest, I was thinking more in terms of possible pollination in the other direction. Solaris -> Linux.
Code reuse between kernels is always an iffy proposition. Often only the transplantation of *ideas* is practical. But it would be interesting to see if, for example, ZFS in Linux might (or might not) fly with licensing issues out of the way. ZFS is more alien than even Reiser4, and by Linux kernel standards, ZFS is a rampant layering violation, as Andrew Morton described it once.
With Ext4 out, and btrfs as far along as it is, I suspect there might actually be little interest at this point. But it would be interesting to see what happened, nonetheless.
It may well be that a change of license wouldn’t actually get much, if any, actual code used. But it would, at least, clear the air about what isn’t in Linux due to licensing, and what isn’t in Linux for reasons of technical taste and/or practicality.
Also, I wonder if an end to the draconian joint copyright situation with current Sun projects like OpenOffice.org might not kick-start a renaissance in that area.
Edited 2009-03-27 02:04 UTC
Well, Ext4fs are also aimed at different markets; ZFS would be more suitable for flash raid where fragmentation is not an issue. The problem with btrfs is that it is behind ZFS when it comes to stability and so on. With that being said, given that alot of these new file systems make assumptions over future storage, it will be interesting to see the viability in the future of them versus what happens in reality.
The reason that UNIX hasn’t faired better in companies is actually two easy things.
1) When you look at the price sheet it costs more (despite the fact that it does a lot more). Managers can’t see the forest from the trees.
2) It is far easier to find people that know how to run the Windows desktop than the UNIX desktop because most people see it every day.
I’m not saying average people are any good at working with/on Windows servers. I’m just saying that managers only look at the lowest common denominator and pick what is easiest to pick unless that pick absolutely won’t work.
Like mission critical stuff. Windows isn’t mission critical no matter how much Microsoft might want to try to say it is. When it comes down to someone getting fired when up time isn’t 99.9%, they don’t choose Windows because they know it will get them fired. This is where UNIX comes in.
And that, simply enough, is why more bosses buy Windows than anything else. That and the fact that MS brass go smooze with CEOs who don’t have a clue about tech and are the ones that too many times make the (clueless) decisions what to buy.