Slashdot Mirror
SCO & Linux: If You Can't Beat 'Em
BugBBQ writes "The NetworkWorld Fusion News reports that SCO is going to jump on the bandwagon and produce its own Linux Distro. " The article also has some analysis of what the SCO folks could bring to the scene as well as what extras they have to add.
I would really expect the primary market for this Linux offering to be those same VAR's. Many, many, many of these VARs have been jumping ship of late. Linux has been cheaper, faster, and easier to use out of the box.
--
-- Slashdot sucks.
Which is make Linux attractive to the suits.
i t's-all-three thing we call Linux...
Love 'em or hate 'em, SCO are a "name" to the Men With Big Chequebooks. And being a "name" is far, far more important than having a decent product or any trifling considerations like that.
However, SCO UNIX isn't actually all that bad and has a half-decent, tried and true support infrastructure behind it. SCO also have quite a lot of money.
I would be very interested to try out SCO Linux, just to see what a commercial UNIX vendor makes of this weird now-it's-SysV-now-it's-BSD-now-it's-POSIX-omigod-
--
Peter
A Linux distro with all the quality, completeness and usability of Xenix!!!
"SCO is in a unique position to dominate this [Linux] market," says Tony Iams, an analyst with D.H. Brown Associates, a Port Chester, N.Y., consulting firm. "They own the low-end Intel/ Unix market. They know this space like no one. They have a tremendous set of relationships with resellers and OEMs."
I must have missed something here. Has the science of cryogenics moved on so much that we can freeze an analyst for ten years, thaw him out and get an opinion ?
Stephen Hawking has written another book. It's about time as well.
I think this is where an SCO distro could really make some serious inroads. These are the companies that have a hard time maintaining fully licensed shops running commercial OS's and software.
Now, they will be able to get a stable, affordable solution from a company that they are already familiar with and can trust to provide the support they will need. (Yes, I know Red Hat provides support, but they're still not proven yet at the suit level).
And for mission critical apps, they can get their Unix box and linux support servers all from the same place now. Bring out a desktop distro with decent office apps and you can have the whole organization outfitted from the same place. Well, maybe not that yet... but this is a great start.
Of course I use Microsoft. Setting up a stable unix network is no challenge
I don't think that this was a great surprise to anyone. It was really only a matter of time before they released a Linux distribution. After all, what else is SCO good for if not following the leader? Sun releases free copies of Solaris/x86, SCO shortly thereafter announces the free version of UnixWare. Now that the other major Unix companies are switching to Linux, what can SCO do but tag along?
But SCO is doomed. Unlike SGI and Sun, SCO has no major products outside its operating systems. Any add-ons it offers are likely to be duplicated by Open Source Programs from someone else. What will they be then? A Linux distributer? Good luck.
And, quite frankly, I'm surprised that they didn't consider using BSD. A large percentage of SCO's usage comes from VARs who offer SCO-based products or services. Wouldn't it be better for a business if you used an operating system in which you didn't have to release all the changes you made? But I guess OpenBSD just wouldn't have the same press impact as Linux. Their loss.
..because that's where UnixWare excells. It's scalable across CPUs. Well, I think it's going to be interesting.
Sigged!
My real problem with the hoopla of too many distros is bigger companies complaining about the lack of standards. People complain about companies like the one I work for or whoever only supporting selected few distributions...well...if every distribution wasn't so different it would be an issue. This doesn't bother me...its actually something I like. I like the fact that no one person's machine is like mine. But the more distributions there are trying to make themselves look unique the less they adhere to certain standards which have developed. Hell..what am I talking about. There aren't any standards. ;-) And here is the beauty of that..
The great thing is that those standards aren't as important if the software is open source. So maybe this lack of standardization is helping bring more companies to look into open sourcing their products...which in turn converts them (because they'd see the obvious beauty in open source immediately if they have half a brain)..Wow...what a thought..more distros bring more variety...more variety brings less standards...less standards help people to see the light!
So come on SCO...anyone else want to start a distribution? This is GREAT! I welcome them all!
I wonder what approach they will take to make sure their software is decoupled enough from the rest of the system to insure they don't run afoul of the GPL?
Either they open source (or free software for people obsessed with semantics) everything they have (a radical departure for their corporate culture), or they try to keep a clean boundry between the Linux/GNU system and their own proprietary software.
If they try and keep a boundry and maintain their own closed products, they are likely to do as much work trying to stay clear of the GPL and similiar licenses as they are to put the distribution together... Lets see, this product links to the c library, which is LGPL, so I can do that, but it requires this kernal modification, so I have to release that, but it requires this utility, which is GPL, so I have to include the source for that...
Their web page worries me a little, they sing the praises of open standards, but open standards != open source... both are good but they are apples and oranges.
I am not trying to slam any companies or criticize any of the licenses out there... I am just pointing out that all the current major Linux/Gnu distributions have avoided running afoul of any of the "open source" or "free" software licenses by making everything they add "open source" or "free" as well. If you release everything you add under the GPL or similiar license, you can't be violating the licenses. It keeps it relatively simple.
If they are the first to try to create a hybrid distribution, they will have some new ground to break and some work to do.
The problem they used to address was easier... we have this closed system and we are adding some open source / free software tools. Hard to violate the GPL in this case, just release the source to any GPL software you add. When things are turned around and you are trying to add closed source / non-free software to an open distribution, it is tougher to make sure you have not violated the GPL.
Just some thoughts...
Bill
Mathematically impossible requirements are technically not against policy.
There is a trend towards the merger of old 'closed-source' companies and the newer 'open-source' philosophy. I think this is related to the maturiry of the product. Linux is technically well matured, and its ready for wide enterprise deployment. I see more of this happening in other fronts. Some kinds of application are already very mature, and open source equivalents are catching up pretty quickly. Let us say, if Lotus started supplying support for gnumeric? They were leaders in spreadsheet software, and yet carry a strong brand in that. Other companies could do similar things, with a word processor, for instance.
As for this announcement... companies like SCO are well stablished, have a recongnizable brand, and know how to handle their customer bases. However, it's impossible for SCO to stand against Windows. Being the only commercial supplier of x86 Unix with any life in it, I think it must be time for them to jump into the Linux wagon. I wonder if it's not too late.
SCO has very good administrative tools. Linux administrative tools are the nightmare of operators - people who know how to type, but dont know how to edit a file if told so. These people is responsible for some critical tasks such as backups, system shutdown and restart. Here in Brazil its common to find in smaller companies that the operators for the night turn dont know nothing of english, so its needed some simple interface where they cant be lost.
I think this can be the most successful combination. It will be somewhat weird to hear of 'gnumeric 1-2-3', 'SCO Linux', etc. For people who dislikes anything resembling old-fashioned corporations, that will be the utmost nightmare - their loved open source projects with corporate faces.
Win-win #1
"Building the Linux clustering capacity to be in line with SCO's NonStop Clusters technology"
Linux redundancy is currently limited to a few nodes - SCO claim they will increase this to 12. This will help Linux get accepted into more mission critical applications, where redundancy is a necessity.
Win-win #2
"Beefing up Linux's symmetric multiprocessing capabilities."
We all complained about Mindcraft. SCO deserves credit for deciding to put effort in to fix this.
Win-win #3
"Managing multiple Linux servers as well as applications from a single console as if they were a single system."
Four telnet terms on a screen? Seriously, I wonder to what extent this will be integrated? SGI and others have a similar tool available, enlightenDSM. Linux *really* needs this kind of configuration tool for it to be accepted into the high-end server market.
Win-win #4
"Improving security ..."
;)
Although the article doesn't indicate to what extent this effort will be, it's certainly a good move. Since when did you hear the other vendors claim they are working on improving security issues. Maybe SCO will help fund the linux kernel security auditing project?
Well. Good thoughts SCO. I think they have got the right direction, and im sure 20 years of *nix experience will give some mature input into the "Beloved" OS
The problem with the way Linux is being pushed down everyones throats, is that its forcing the entire market to become EXTREMELY volatile.
Who is "forcing" who to use Linux? I don't recall seeing Linus Torvalds or RMS with a gun to anyone's head. If you don't want to use Linux because it's "already pretty ancient", then there are other "cutting edge" OSes you could choose.
I just don't see how Linux, being free and open, can force you into anything. It's the closed, proprietary software which does that.
Newsflash:
A vendor famous for its remarkably comical marketing dept, SCO shocked the entire unix world by jumping on the linux bandwagon after everyone else had, instead of standing alone and fighting it out as it generally does. "Our marketing dept. ran out of analogies and cliches," said a senior SCO executive on condition of anonymity. "So we jumped on the bandwagon to show we are not a flash in the pan".
SCO's CEO had been quoted a few years ago in Byte magazine ridiculing open source development for linux, using the analogy of holding a cup under a waterfall and waiting for the water to flow.
But having changed its mind, SCO outlined some of the features which would differentiate it from other distributions:
* README files with a neverending flood of press releases announcing "industry partnerships" and "strategic alliances".
* Industrial strength stick-to-it persistence. "Once we make a mistake, we repeat it until we get bored of it or people stop paying attention," said a SCO executive.
* Clueless marketroids included free with each upgrade.
* 20 year old icons, stored carefully in clingwrap in the secret SCO vault.
* Open Source Litigation to harness free legal support for SCO's battles with Microsoft over Xenix (not included with distribution, but free if you buy a $5 "I love SCO" bumper sticker).
If you are using SCO, it defaults to ksh, and you probably also have sh and csh. Korn shell offers plenty of flexibility, but I can see how some people would not like it, espesially if they are not comfortable with vi commands.
Information wants to be anthropomorphized.
The result was that AT&T gave away copies of Unix to universities for educational purposes, which made it very popular in very important places.
After AT&T was broken up, they were allowed to start selling things, and one of the things they did was sell commercial licenses for Unix. In the late 70's, Microsoft was one of the people who bought the right to distribute Unix, and they modified it to run on the 8086 and 68000 and sold it to folks like Radio Shack and other vars, the largest of which was called SCO.
The first license that AT&T used was *very* cheap, and MS was making a lot of money selling Xenix. AT&T was kind of new at this computer stuff and didn't really know what it was worth. As a result, the next version of Unix that AT&T released, they jacked up the licensing fees to the point that MS said pluck yew to AT&T and sold off their Xenix distrubution to SCO.
Microsoft has never liked Unix since that spat, although they did add many Unix like features to their CP/M clone called "MS-DOS".
If AT&T had been more reasonable with their licensing of Unix back in the late 70's, we would all be running Unix now.
SPF support for most open source mail servers can be found at libspf2.
They still have some credibility with solutions vendors -- folks who, e.g., make custom medical records systems and other such things -- but even ISV's are inching away from SCO and towards Linux. Most of the "name" Linux wins lately have been the result of this migration.
But Fortune 500? I know of no Fortune 500 company that would consider buying SCO Unix or dealing with SCO. If they have Unix, they have "real" Unix (Solaris or AIX). Or Linux, in certain special-purpose instances. SCO Unix rates behind even IRIX and HP/UX on my list of "sales into Fortune 500", though the solutions vendors still sell lots of it indirectly.
-E
Send mail here if you want to reach me.
Microsoft was the first commercial Unix licensee. They bought a license to Unix System III.
Now, you must remember that at this time Microsoft was a considerably smaller company. In fact, they only had a few dozen employees, hardly enough to handle their other product lines. So they contracted with a consulting outfit called The Santa Cruz Operation to port Unix System III to the various 16-bit microcomputers that were being introduced. The result was called "Xenix". Radio Shack had a version for their 68000-based business system, I believe Altos had a version for their 8086-based business system, but I don't recall Xenix being sold for standard "PC Clones" by Microsoft at that time -- it was, at the time, a strictly OEM deal, where an OEM wanting Unix had to go to Microsoft, pay money up front for the port, then Microsoft would pass along most of the money to SCO for SCO to do the actual work.
Eventually, Microsoft decided Xenix wasn't going to be particularly profitable, especially with IBM shoving tons of money at them to make OS/2 be foremost on their plate. They handed off Xenix to SCO in exchange for some cash, future royalties on future sales of Xenix that included Microsoft-paid-for work, and a large share of SCO stock (just hedging the bets in case Xenix DID take off).
So anyhow: yes, Xenix originally WAS a Microsoft product. But no, Microsoft didn't write Xenix (or at least not the majority of Xenix), though most of the early Xenix work was a "work for hire" done by SCO for Microsoft (and thus like all such "work for hire" was property of Microsoft). A fact which led to a lot of acrimony and lawsuits in later years.
-E
Send mail here if you want to reach me.
True enough.
Personally, I'm more of a vi person anyway, so I don't consider it an issue. :)
Information wants to be anthropomorphized.
For that matter, almost all shells can grok sh commands fairly well.
So as long as you know where your favorite shell differs from Bourne, you can get things done in most of the other shells out there. If you find yourself using an enemy shell a lot (and it sounds like that it your situation), an ORA animal book can help you find all the spiffy features that are different.
Beyond that, we would just be getting into religious issues.
Information wants to be anthropomorphized.
The clustering, the FNP, Symmetric multiprocessing, etc... Will the GPL apply to these enhancements as well?
Yes. Stuff like that needs low-level kernel hacking, and anything added to the kernel is automatically under GPL (the stuff you mentioned is too low-level to be modules).
Will I retire or break 10K?
Beowulf and NonStop Clusters are two VERY different clustering schemes. Beowulf is sort of like d.net where a controlling terminal sends problems to the nodes for them to figure out and return their result. A real cluster like the sort SCO has been building for a while lets you run programs on it like the cluster were just one large system. Google has a Beowulf-like architecture where a query is sent out to the cluster and they all sift through the various archives looking for your data like 4000 librarians.
I'm a loner Dottie, a Rebel.
A better question to ask is do we really need more than one Linux distrobution.
I'm a loner Dottie, a Rebel.