Slashdot Mirror
Ransom Love's Answers About UnitedLinux
1) LSB
by Anonymous Coward
From the description of UnitedLinux it seems to me that it is simply a commercialized LSB. How is what you are offering different from the LSB project?
Ransom:
LSB is a specification and does not describe a complete distribution. UnitedLinux is combining the efforts of four major Linux providers around the world to create a common, best of breed server implementation (full Linux Server distribution). It will be a full implementation of the LSB standard, but it will go beyond and include components not currently defined in LSB.
LSB does not attempt to solve the business model around Linux. UnitedLinux is not only providing packaged bits and bytes of a distribution, but also giving developers a global infrastructure of support in local languages and channels to sell their products on a global basis. UnitedLinux solves nearly all of the impediments facing the commercialization of Linux, such as market fragmentation, the availability of applications (by simplifying certification of hardware and software solutions), and a valid business model for Linux (Supportable business quality product through limited binary distribution and 12 month maintenance agreement with every license sold).
2) What will you give back to the community?
by dbarclay10
In a completely selfish vein, what will you give back to the community? Caldera doesn't have the greatest track record (I can think of a few specific cases but I'll omit them here for brevity) for providing some return to those people who have coded the _VAST_ majority of Linux, GNU, and everything else. Aside from, of course, providing jobs for developers.
Ransom:
I am a little frustrated by this question as it implies that somehow Caldera or any other Linux company is making millions of dollars off of Linux. Every Linux provider has spent far more on promoting Linux than they have ever received. Not even Red Hat is profitable and a lot of their revenues are generated from non-Linux technologies. Millions of dollars have been spent in recruiting applications, advertising, and tradeshows to promote Linux, not to mention the millions spent in employing engineers as well as the innovations that have been given back to the community. The actual development cost of producing a product is only about 20-30%; marketing, sales and support constitute the majority.
I am also troubled by your impression that Caldera's contributions to the community have been scarce. Caldera architected and paid Red Hat to enhance the original RPM. We were the first to address a graphical desktop; the first to provide a graphical installer; the first to provide a management system. I could go on and on. All of these innovations were given back to the community. You can see a more detailed list at www.caldera.com/developers/community/contrib/.
Please don't get me wrong - I firmly believe that what is produced by the community is vitally important. But it is not what creates a product or a solution that businesses have to deploy. Caldera has been a company for profit. Profit enables continued investment in Linux. UnitedLinux is for profit so that there is money that can continue to be invested in Linux itself. I do not believe in a Linux model that requires ongoing charity to survive.
That said, Caldera and all the UnitedLinux partners will continue to provide all of the changes and enhancements that are made back to the community. I believe that all of the participants have a very good track record of so doing. In addition to providing the source back, UnitedLinux will offer development programs that will provide continued access and updates for the serious developer.
3) Source and binary distributions
by RGRistroph
There has been some confusion on your statement in the UL teleconference to the effect that while source code would be available to meet the requirements of the GPL, "binaries would not be freely available." Could you clarify what that means? Is it possible that UL will distribute only source, or only distribute source and binaries to it's member companies? (Who will then be responsible for making sure they meet the license requirements on software which is in their distributions?) Surely UL or it's members don't intend to distribute binaries compiled from GPL code and assert the recipient can't re-distribute them?
Ransom:
The binaries that are certified by the major ISVs and OEMs will not be made freely available for distribution by anyone. This is to limit the support liability for these companies and to ensure a high quality, consistent product around the world for support purposes. The UnitedLinux product produced is not just a binary, but 12-months of maintenance. That maintenance is for a single system and therefore has limited distribution. The source code for the server will be made freely available for all in compliance with all of the Open Source licenses.
There will be programs for developers who need access to the binaries and they will include options for ongoing updates and patches to ensure continued certification compliance. Our desire is to make UnitedLinux easily available for serious developers, and give them means to make the development process easier.
4) Commercial Development
by Marx_Mrvelous
It seems to me that a group like UnitedLinux could bring a lot of commercial development to the Linux platform. Are there any efforts to bring companies who have so far neglected developing for Linux due to support costs, like most hardware venders, into UnitedLinux?
Ransom:
Certainly. By enabling one certification for hardware and software and then facilitating a global distribution of the solution through established global channels with support in local languages around the world, UnitedLinux should increase the number of hardware and software participants supporting Linux.
5) Future of Linux
by micro-colonel
Where do you see the true future of Linux being? Will it remain mostly in the enterprise and web server market, or do you think that it will also make large gains in the desktop market? Also, to what end does the goals of UnitedLinux fit into your predictions for the future of Linux?
Ransom:
Linux has great potential in moving beyond the web server market into the mainstream of the application server market. There is a lot of work that still needs to be done, however, to allow Linux to be a dominant application server platform. The objectives of UnitedLinux are to take that first step: enable Linux to be used by mainstream businesses. Accordingly, the initial effort is focused at the server.
I also believe that web services will become the dominant method for outsourcing IT. As the Internet becomes the primary business platform and Internet client interfaces become dominant, Linux will continue to make inroads into the client. (Take, for example, the fact that we now spend more time in email than in an Office suite.)
Finally, one of the keys to desktop penetration of Linux is in ease of management and provisioning. Making Linux easy to configure, deploy, manage and interoperate with Microsoft alternatives will greatly enhance Linux's acceptance at the desktop. Clearly, several of the UnitedLinux companies are addressing these issues individually. UnitedLinux may address this after enabling Linux to be used as a mainstream application platform for business. Another key, of course, would be the number of applications with which end-users are familiar. With the current balance weighted towards Microsoft, the need is for new software that makes end users more comfortable with and consequently more accepting of Linux on the desktop.
6) Business Model...
by powerlinekid
Mr. Love, I'm curious as to how you'll make money from this? By not giving away binaries it seems as if your group is trying to sell Linux, and probably service and support with it. Now you appear to be in competition with Red Hat (on server) and Mandrake (on desktop) who both give their software away. Red Hat makes it's money from service contracts and Mandrake from special software for paying customers. I guess my question is how can you compete against them, when they are just as good and give it away for free or cheaper? What is the incentive you will give consumers to actually purchase your software as opposed to downloading isos from other companies?
Ransom:
It should be noted, first off, that Red Hat has moved to a model on advanced server where they are not giving away the binaries and they are charging around $800+ for their advanced server product.
Going forward, there will only be two platforms certified by the major hardware and software vendors, Red Hat and UnitedLinux. For Linux to move from the peripheral of the business network into mainstream application server market, businesses must be assured that their platform is certified and will work with other applications and hardware solution in their environment. What the UnitedLinux customer is paying for is 1) the assurance that his applications will work together, and 2) the ongoing maintenance and support of that certified platform. The restriction on binaries is to ensure product quality and consistency of the brand for hardware and software vendors and for the quality of support within the business organization. I believe that Red Hat is moving to a similar model with business customers. The majority of the value will be in product assurance and maintenance. Both of these are of tremendous value to the business customer.
As mentioned earlier, binaries will also be made available to developers, but they will be through programs that can keep them updated and in sync with all changes. More information on this will be forth coming.
Desktop derivatives will be made available by each of the respective Linux companies. Since they will not carry the UnitedLinux brand and do not need application certification, their binaries may be available based on the individual company's policies.
7) Documentation
by forgoil
Will there be some form of initiative to work together on online documentation for both end users and developers? For instance making sure that there is up to date information on all applications and APIs in a common format (for example XML, that can be used to generate info, man pages, html, etc)? I personally don't think the distributions as a whole are well documented enough, and I think it would be one area where everybody would gain from co-operating.
Ransom:
We will be offering developer programs to address many of these needs and your suggestion here is a very good one. Clearly not having to duplicate efforts will allow us to create a much higher quality combined product and this is an important area to improve.
8) Who certifies compliance?
by Rogerborg
Who will certify compliance for each vendor provided distro, and who will pick up the pieces when (not if) an application appears that borks on one or more of the distros? If it's UnitedLinux, is each vendor prepared to pay to fix snafus committed by the others? If it's the individual vendors, what happens when one of them screws it up and wrecks confidence in UnitedLinux?
Ransom:
Every company will be shipping a common CD that will include a complete Linux distribution including installer and desktop. This is the UnitedLinux aspect of the distribution. All the additional value-add will be on separate CDs. Consequently, there will be a common quality check on the base components. The testing of the value-added components will be the responsibility of the individual companies.
9) Patents
by Rogerborg
Given the ongoing uncertainty over whether Red Hat's actions regarding patents will actually match its rhetoric, what is UnitedLinux's position on patents? Specifically which of the following will you do?
- Eschew patents altogether.
- Obtain your own patents.
- License, trade or buy outright patents from other companies.
- Oblige your members to hand over or license patents to UnitedLinux or to all other members.
- Match Red Hat's current stated intent and express a non-binding intention to stay enforcement for a given type of open source development as long as it is convenient for you to do so.
- Agree to explicitly license your patents at no cost, for a limited time or in perpetuity, to a given type of development (as sharply distinct from merely staying enforcement and leaving a Sword of Damocles dangling over developers' heads).
- Obtain and reserve the right to use patents freely against any target, as any other commercial software companies (e.g. Sun, Microsoft) would do.
The four companies have not discussed their position on patents. The official statement will have to be forth coming.
We live in a day when patents have not become a tool to protect, but a weapon to wield. Since the patent office lacks the technical expertise to discern between what is valid and what is clearly an attempt to blackmail, and because innovation continues at a rapid pace, this is a serious problem for our industry and one that will need to be addressed.
10) On the Relationship between Companies
by the-banker
How is the membership into the United Linux group going to work, and how much flexibility will there be? Can any distribution join? Are there significant costs to becoming a member? Can members set their own policies with regard to per-seat-licensing? In sum, how much freedom do the member companies have in how they market, contribute and license United Linux?
Ransom:
The membership will be open. Any Linux company will be able to join, but they will have to pay the fee to become a joint owner of the UnitedLinux LLC. That fee is to offset the cost of development and marketing the UnitedLinux product and brand and to have the upside potential of profits.
There is no per seat license for UnitedLinux. The restriction is per server and it is the 12-month maintenance fee for that system. Other companies can set their own pricing, but they are under obligation to deliver the same product and maintenance deliverable per system and will be responsible for the fees back to UnitedLinux. There may be several ways of participating with UnitedLinux. The details on membership and different options will be forth coming.
The binaries that are certified by the major ISVs and OEMs will not be made freely available for distribution by anyone. This is to limit the support liability for these companies and to ensure a high quality, consistent product around the world for support purposes.
The question itself is typical from the community that has grown to embrace Open Source, but the answer is pure PR fluff. Keep an eye out for this technique; you'll see it in play whenever cable rates go up or a business is about to make it harder for you to cash a check.
Try not. Do or do not, there is no try.
-- Dr. Spock, stardate 2822-3.
If the quality is higher than the free alternatives, there will be people willing to buy for that extra reliability.
Stop the brainwash
But it is not what creates a product or a solution that businesses have to deploy.
I hope the dark irony of that statement doesnt escape this audience. The fact is GNU/Linux was created exactly BECAUSE of this concept. Sad, Ransom Doesnt Get It.... or he thinks we are all stupid.
questioning fees or the idea of paid binaries:
If you don't work in a large corporate environment with some heavy iron or reliability/QoS contracts with customers, shut your yap now and read on.
Allow me to explain.
Current primary linux distros (RH, Mandrake et. al.) provide a set of binaries and source for multiple target platforms.
Each of these "should" work, or will "probably" work on a given platform, and the free updates, or access to source from which you can upgrade your system will also probably work; none of these distros go through any heavy-duty regression testing or certifications for uptime guarantees or reliability.
What Mr. Love is getting at, and the idea behind UL, is to provide a serious platform for linux in the server market.
By maintaining strict control of the binaries (both Linux base and applications) UL can fully test and certify that "out-of-the-box" the software and server combos are completely compatible. No guesswork. No hoping that source compiles and runs ok the first time on platform x,y or z.
They can also begin to provide guarantees on the availabilty of critical applications.
That's what business customers are going to pay for. They'll demand nothing less. To get something for free in a quality-controlled environment is useless if you aren't 100% certain that it will all function as-promised.
I am a little frustrated by this question as it implies that somehow Caldera or any other Linux company is making millions of dollars off of Linux. Every Linux provider has spent far more on promoting Linux than they have ever received. Not even Red Hat is profitable and a lot of their revenues are generated from non-Linux technologies. Millions of dollars have been spent in recruiting applications, advertising, and tradeshows to promote Linux, not to mention the millions spent in employing engineers as well as the innovations that have been given back to the community. The actual development cost of producing a product is only about 20-30%; marketing, sales and support constitute the majority.
If marketing and sales are that high a percentage of your costs, you're doomed. What he means is that millions of dollars have been wasted promoting linux to the wrong people. As much as I like free beer and booth babes, you don't need to spend millions of dollars selling a free product to a bunch of geeks who already have it, and don't intend to buy support. Furthermore, they should take a lesson from microsoft, and have virtually zero support costs. All support should be billed by the hour. Somebody needs to stop giving this guy money.
The worst of it though is that he feels that we, the developers out here writing the actual code, owe him something because of all this "promotion" that Caldera has done. It's like the guys who used to wash your windshield at the toll booths and then ask for a tip. Did we ask for this "promotion"? No, so go away. At least now we know not to hold our breath when looking for community support from UL.
Linux will be sold the same way all other platforms are sold: by the applications. When there are applications you need that run on linux, you get linux. I see it every day as the purchase orders come in. No distribution provider out there looking to make a buck understands yet.
I have to give Ransom credit for his good answers. Especially in the area of whether Caldera will make billions off Linux without giving back to the community. Ransom makes it pretty clear that Caldera's development of Linux, even though done for profit, is in itself something big that's being given to the community--it's the effect that takes place when a product has a lot of support from reputable companies more than which lines of code Caldera writes and puts wherever. Even in Caldera didn't write a single apostrophe of code but only packaged other people's hard and free work and sold it for profit, they'd still be opening a market window for the Linux community that wouldn't have existed otherwise.
I believe that for Linux to really be successful, both in terms of code quality and widespread acceptance, a team of companies needs to get together with a specific purpose to fulfill: To develop a desktop version of Linux similar in functionality to Windows XP or Mac OS X. In other words, other than a few options here and there, it wouldn't really have the possibility of configuring every detail. The bare-bones system, which basically means a desktop with a few icons on it for basic functionality (and no applications) would basically be integrated into the kernel. When the computer boots up, it boots directly into a nice graphical desktop, kind of like the way BeOS did, except Linux will actually stick around. Most of the configuration would be built in, heavily reducing the number of "system" files. No desktop system needs to drown in its own configuration. I would say that a really good bare-bones desktop Linux OS shouldn't take more than 15 to 20 megs of space, and that includes all the beautiful graphics it would have. To make this system useful, a number of applications made specifically for this version of Linux would be included, increasing the size to about 60 or 80 megs. This might include the standard "fun" stuff that consumers look for, like video editing, audio editing, image editing, and the more mundane crap, like word processing and whatnot. Obviously, major formats would need to be recognized. Only when something that's actually easy to use becomes available, Linux will skyrocket in popularity and Windows will go into the dumpster. Yeah, most of the code for this exists. Someone just needs to recognize why Linux tends to be so damn complicated, and rearrange these pieces into something that's really well designed and coherent.
As a disclaimer, just so someone doesn't say, "You stupid fsck! If you think Linux is too hard, go back to Windows and enjoy crashing all the time. Otherwise, learn how to use the damn thing!" I've been using Linux for years now, since an old Yggdrasil distro I picked up at some computer trade show, and a 8 CD set of crap they downloaded from a bunch of mirrors and stuck on some CDs. (Back in those days, those CDs were really handy, especially considering that most folks had a 14.4 connection. And that was considered fast! Even 56k was far off! So that was a loooooooooooooooooooooooong time ago.) Anyway, I've since realized that Linux is too fragmented. There's way too much duplication of effort. The Linux file system "standard" is totally wrong, IMO, because everything's in the wrong place. (/home should be under /usr, for example.) And the configuration is a nightmare! There are literally hundreds of configuration files strewn all over the system! So I since moved on to FreeBSD. All ya'll Linux folks out there can learn a shitload of great stuff from the BSD folks. Just look at how much of the system configuration for a FreeBSD system is in a single file (rc.conf). In Linux, this would be scattered across 10,000 different files in a monster directory structure under /etc/rc or something like that, as far as I remember. As you can easily tell, I haven't touched a Linux system in ages, and there's a reason. That's why I say this desktop thing needs to be made. Linux can kick Windows' ass in server stuff. Now it needs to start kicking Windows' ass in the desktop, because only when Microsoft loses that battle, they'll no longer be able to embrace and extend, a practice that fscks up all our code because we have to be compatible with their bugs, flawed design decisions, and programming features probably designed by lawyers and marketeers rather than programmers. Ooooooooooh well. Oh yeah, I had all kinds of Negra Modelo last night. It was fun. (If Cerveceria Modelo made software, their operating system would kick everyone's ass!)
The GPL doesnt ask for any of which you are demanding from businesses. Indeed, some of the other Open source licenses (notably, the BSD license) are even less demanding than the GPL. Noone forced the authors of open-source software to pick a specific license - they did so of their own volition. One of the consequences of that choice is that other companies can make money off of that work. How is this immoral?
Redhat (and some other companies) have done more than what is required off them by the license. I wont go into their motivation - which is irrelevant to the point I am trying to make - that is : just because some companies may not go the extra mile, doesnt make them immoral. If you cant stand the consequences of the open-source licenses, dont pick them to license your code in.
There is no such thing as luck. Luck is nothing but an absence of bad luck.
Heed the advice that has been around for years now - Give Away the Software (that means code and binaries, in a usable form), and Sell Hardware, Support, and Maintenance
... but you still need to buy their hardware, support, and maintance to get it.
That model doesn't work unless you're an established player (eg IBM) and are well known. IBM sells custom hardware (RS stuff, not x86 crap), support, and maintance. They formerly (and still do) sell the OS (aix) as well. Now they'll give it away
Ask LNUX how well selling hardware with a free OS worked. Or RHAT how much money they make selling support for a free OS.
Provided that more "traditional" variants on the Mac were still available (just as other Linux distros will remain available), I don't see what would be wrong with a move like that. The whole idea would be to get the Mac into a new market, not to sell that model to the people who already use Macs to do audio/video work. If Jobs were to try to sell a supercheap audio-less "BusinessMac" to people who run spreadsheets all day, I would applaud.
I think Love is up to the same thing. He's not interested in you "Linux community" people. You'll go on running Slackware or Debian or whatever. Corporate jargon, on the other hand, is the language that the people running NT know, and those are his prospective customers.
As copyright owner of this comment, I authorize everyone to defeat any technological measure which limits access to it.
One of the most frustrating things about using some commercial products under GNU/Linux has been the Red Hat centric methods of distribution: binary RPMS designed for Red Hat that may or may not work with other RPM based distros, and will require quite a bit of hoop-hopping on non-RPM distros such as Slackware, Debian, and all of the excellent Source Based Distros.
Not all products do this
Now it looks like United Linux is trying to reduce the choice of anyone wanting to run a commercial product on their system to two choices: Red Hat or United Linux. How likely is a product released as a cross-distro compatable tarball going to be to get "certification" vs. a Red Hat/United Linux RPM. It looks like the behometh forming here has every intention of dictating standards and shoving de-facto norms down the community's throat, rather than taking the consensus-based approach we have used up until now. If this impression is correct, this is anything but a positive development for GNU/Linux or free software and will likely be quite detrimental to the communities which surround and support them.
Having tried Red Hat, Suse, Mandrake, Debian, Source Mage, Gentoo, and others, I can unequivocably say that, in my place of work as well as in my home, the kind of prior restraint on my choice they seem to be aiming for (with their 'certification' requirements) is such that it will eliminate commercial products we would have considered purchasing from contention altogether (We are a Debian shop, currently transitioning to a source based distro for performance/reliability reasons. Red Hat and United Linux are not, and never will be, under consideration for deployment here).
It would be unfortunate if this were to become the norm
This reduces choice, and is a bad thing for the Linux community IMHO. The result is more likely to be less willingness to deploy commerical products if a marginally usable free alternative exists that happens to be usable on the platform/distro of choice (whatever that may be). Contrast this to current conditions, where many vendors are sufficiently neutral that one can deploy their product anywhere, and the free community drives most standards, not Red Hat and Caldera.
I started reading this interview with a very positive "I'm glad to hear the negative rumors are likely wrong" and by the end of the interview have reluctantly concluded that this isn't going to be a positive development for GNU/Linux at all, and many of the worst fears expressed by others earlier are quite probably entirely well founded.
Worse still, the answer given to the patent question was downright chilling...thank [deity] it only affects the United States, and not (yet) the entire world. This is one way they could very effectively steal our work and cut off our access to the products of our own labor, perhaps even in spite of the GPL.
The Future of Human Evolution: Autonomy
I'm gonna have to disagree.
If you want to swim in the financial waters, you need to speak the financial language.
Being able to sling around financial buzzwords no more implies technical incompetence then being able to sling around Linux buzzwords implies programming competence. (Not to mention that it is not a given that technical competence is necessary to run a technical business; it's just so few CEOs, or indeed people, can stand to trust the tech people enough to make good decisions just based on their word.)
While it can annoy people like us to slog through a language with a lower informational density (caused by the higher density of words meant to signal things to the listener's subconcious like 'conformity' and 'if I sound like this, I must know what I'm doing', annoying to us but CEOs who don't use this language aren't CEOs for long), the true test of the value remains the content of the speech.
On that measure, this CEO did quite well, and your harshness is incredibly unfair, and your moderation probably undeserved. They equally justifiably think the same thing about people who are only concerned about tech, and never the business. ("What actually puts food on the table? Do the techies think they can eat their server?")
Look past the surface.
Sounds an awfull lot like "One Microsoft Way".
One window manager, one mdia player, one web browser, one video app....
Slackware forever. Honestly, what else would you trust when it absolutely positively has to be stable, secure, and easy
Yes, the parent was modded as funny (rightfully), but...
People who talk like this can only relate to two things: other people who talk like this
Welcome to the real world. The real world does not consist of coders who ignore everyone else in their insular little universes and think that everyone should be as smart as they are. There's a reason that there's more MBA's than engineering degrees out there -- more people think in a management/marketing perspective than an engineering one.
Is Mr. Love spewing a lot of horseshit? Sure. But it's horseshit that other horses will sniff, recognize and find comfort in rather than the bullshit that a lot of the Linux community spouts. Bullshit is just confusing to horses - it smells kinda similar, but the nuances are missing and it makes the horses nervous.
Insert sheep or lemmings for horse if that makes you feel better.
It still doesn't change the reality that Linux is still far, far away from serious corporate acceptance. IBM, Redhat, and other companies are making headway here (hooray), but it's largely through the same kind of crap that Mr. Love is talking. It's not based off some ethical superiority, and it sure as hell isn't because businesses think OSS is a good model for their core logic.
We'd love to move our systems over to Linux here... but upper management starts looking nervous every time it's mentioned. So instead we're on AIX (gack) and horribly overpriced PPC boxes. Maybe once Linux gets some more positive press from companies like Redhat, IBM, Caldera, etc. we'll finally be able to move to something better than AIX.
What I read from the subtext of this article is that the major Linux distribution companies have lost confidence in the GPL. He doesn't come right out and say it, but it's clear that they are fighting the license. Without violating the GPL, they are trying to make it inconvenient for users to redistribute the software and avoid the per-seat licensing. He also says that he not does believe in a Linux model that requires ongoing charity to survive. This sounds like a dig at Mandrake.
He is right, though. Charity is a terrible business model. The only time it works well is when you campaign for some worthy cause (e.g. feeding starving children in Africa) and then spend 90% of the money on "administration". Here's an interesting tidbit from the other side of the political spectrum. Did you know that Ayn Rand (a laissez-faire capitalist and staunch opponent of taxes) believed that if income tax was abolished then the rich citizens of America would voluntarily donate money to establish an army? Can you imagine a world in which armies were established and paid for solely by robber barons? I mean, they'd probably be sent off to fight any country that threatened to mess with our oil supply. Oh wait, that already happens.
-a
How to rationalize theft.
If you want to swim in the financial waters, you need to speak the financial language.
Slashdot isn't financial waters. Any good speaker knows that you pick your language to match what the audience will respond to. If you can't speak techie, then you probably can't think techie. If can't figure out when it's inappropiate to speak financial, you probably can't figure out when it's inappropiate to think financial.
What? GNU/Linux was created for business deployment? Where do you get that?
Posts like this reinforce my belief that most Slashdot posters are computer-crazed teens with no business experience. They can't see the big corporate picture. But in truth, businesses don't want to have to learn every detail about Linux. They want to get on with their operations. Computers and operating systems are a means to an end.
I think Ransom acknowleged the importance of software generated by the open-source community. But until packaged in a business-ready (read: dumbed-down, simple, hassle-free, supported, whatever) distribution, the hundreds of open-source projects we're talking about really don't do much for business.
Obviously there are other distributions out there, some might cost less. Ransom feels that there's a market for a business distribution. He's probably right, that's where the money is.
I fixed the formatting error. Thanks.
As for straightupness, remember that we're talking about it in corporate CEO context here, not about a normal person. For a CEO, Ransom is pretty okay. He uses buzzwords because it's part of his job, but the use here is light compared to most.
(An aside: Peter Wayner and I have registered BuzzwordInstitute.org so we can issue "buzzword compliance certificates" and such; it's a joke, one we'll get going as soon as we have time to actually make the site. We can use some help if you're up for it; we're both very busy. We'll give you some sort of impressive title, even. "Executive Director in Charge of ______" or some such. Peter pointed out that inflated titles are part of buzzwordness, necessary "for executive retention purposes.")
Anyway, I'll admit that I prefer interviews with people like Moshe Bar, who simply speak their minds, but some of the corporate activities are important and it's worth knowing what the Linux corpses are up to even if you have an RMS shrine in your dining room before which you genuflect 5 times daily.
- Robin
How is this essentially correct? He's wrong! Un-Right! Incorrect! You're basically saying that "well, it wasn't first, but it was the first one 'of consequence' and the 'best one' (even though it wasn't first)." Well hell then! "Microsoft had the first gui!" (of consequence, of that quality, etc..etc..)
You're not even splitting hairs here!
Many of the people in this discussion are talking in broad generalities about how you *think* corporations are going to react to UnitedLinux; let me tell you how *I*, a sysadmin for a company that uses both Linux and Windows 2000 servers, gauge this interview.
First of all, Mr. Love is entirely correct about what corporations want, which is a standard, consistent server platform with regular upgrades. We currently use Caldera OpenServer for file sharing and DNS, and we love it. However, we're nervous about utilizing this platform for more than these two functions.
Example: we want to move to an LDAP mail and message server, akin to Exchange. So, I hit the boards and forums, check SourceForge (of course), and come across 8 - 10 server products that *sound* like they fit our needs. My problem is this: how can I tell that this is going to work as advertised on our systems, and how do I know that it's not going to bork everything else we're running? Sure, it's all *supposed* to run properly and play nice with the other boys in the garden, but everyone knows that server platforms, regardless of manufacturer, have interop glitches. Yes, it's true that, if I had all the time in the world and was being paid 200% more, I could download each of these products, set up a staging environment, and try to hammer out the bugs myself.
Or, I could install Exchange and be up and running with a high confidence level in about 3 hours. I know, I know- you all think Exchange is a steaming pile; but the reality is that it's quick and easy to install and administrate. Why? Because Microsoft has farms full of paid developers making sure that it is.
If UL can provide my company with the peace of mind Linux does not currently provide, it will have made a new customer. In turn, the rest of the community will benefit because the installed base will expand, and perhaps another developer gets a job. There's a whole lot of sweetness to be passed around!
One more thing: I would like to remind all die-hards here that corporate profits are the *only* thing supporting Linux development right now. Every developer who works on any open source project gets his beer/rent money from a job someplace; even though he doesn't get paid (most of the time) for the time he spends on a project, he only has the luxury of spending spare time because he has a job to cover the expenses. Why not support UL, so that maybe a few more Linux geeks can have a good job doing something they love?
Perhaps you didn't see the net loss of $17.6 million for the company.
Whether a company is making profits or losing money, the employees still get paid (be them coders or CEOs).
Caldera is not making money. The employees are, but then again, would you work for a company if you were losing money working there?
--- I used to moderate, then I read the -1 articles and decided having to filter through them was not worth it.
Suse has had nearly this policy for a while. I know they haven't given out ISO's for a while, I'm not sure on binary packages. I'm going to guess that some of the companies are going to give out free ISO images that are based on the UnitedLinux source, but aren't "UnitedLinux". So you can't claim it is certified.
People will still want the free desktop OS for workstations and whoever continues to provide that has a huge marketing/PR advantage over everyone else. However, that advantage can be quite costly to provide. So it might all come out in the wash.
RedHat does essentially this in the Advanced Server Platform. Go try and find those ISO's. They aren't anywhere on RedHat's site. I've hunted high and low for them. It is very similar to the Personal/Professional release execpt for some of the kernel patches they apply, and a few 3rd party tools.
The next step after this is that RedHat isn't bothering to get Oracle certified under all of the Personal/Professional Releases, only on the Advanced server. If isn't certified, Oracle shouldn't support you. I would imagine with all big proprietary products will be the same. It isn't cost effective to re-certify RedHat for all the big name products every release. So from a purely business perspective it makes a lot of sense.
So RedHat can give away free workstation copies but get you to buy a copy of Advanced server for all of you're server machines. I'd imagine that UnitedLinux will end up doing something similar somehow.
Kirby
You use the word fragmentation with absolutely NO idea of the history of the word.
The Fragmentation of UNIX (FoU) has nothing to do with whether the debian installer enables scsi-emulation for you or not. You are perfectly capable of reading documentation and enabling it YOURSELF. FoU also has NOTHING to do with whether you understand how to read an XFree86 conf file.
I suggest two things:
(1) Read on the history of UNIX before you speak of it.
(2) Pay for support if you want a company to hold your hand through the most basic of computing tasks.
I always wondered why Redhat and Mandrake and the ilk always made binaries available for download for free. People would sometimes scream and cry about SUSE not offering ISO's at EXACTLY the same time as the packaged distro for purchase. People would scream bloody murder but forget that the GPL is a SOURCE CODE license, not a binary license. The binaries are yours to do with what you want. If you want to sell your brand of Linux for 1 million dollars per seat, you can. However, you cannot deny anyone access to the gpl'd source code. If they wish to recompile the entire distro and sell it for $1 or give it away, there's nothing Caldera can do to stop them. And there's no reason for Caldera to worry about that.
They don't care. They want to do business with people that want to exchange money. Companies forge relationships by the exchange of money. That's how business is done.
They like it.
Open Source coders build relationships by exchanging code.
They like that.
The GPL allows for both and encourages both. It's really pretty simple and elegant... any engineer can appreciate that can't they?
Toddlers are the stormtroopers of the Lord of Entropy.
Apparently the corporation you admin for isn't very big. In almost any organization that depends on it's computer resources, a fee per server wouldn't even raise a question. Would I pay for UL when I could get Red Hat for free? Only if there was some credible reason to do so. But I assure you, If UL was more stable, or robust, or whatever, I would be the first in line to pay for. Most professional sysadmins would.
When I worked for small businesses, I ran Linux on their servers to save money. Now that I'm no longer in that situation and have a little money to spread around, I still choose to run Linux because it's the best tool for the job.
load "linux",8,1
Anyone may distribute the binaries if they want, but neither Caldera, or anyone else, is under any obligation to do so.
"The restriction on binaries is to ensure product quality and consistency of the brand for hardware and software vendors and for the quality of support within the business organization. "
Heh... Well, that could be an argument against the distribution of _modified_ binaries not labelled as such, but not one against the distribution of _unaltered_ binaries. The only reason I can see to restrict the distribution of unaltered binaries is to obtain more money.
Now, keep in mind that all GPLed bits of UL will be freely available to the public in short order. (After all, the UL members can't enjoin their customers from distributing the GPLed bits, so even if the UL members won't do it directly, the sharing that is the basis of the linux community will ensure that they sneak out somehow...)
Anyway, the UL base seems to have a good chance of reaching the segment of the enterprise market looking for well-supported linux server distributions (that Red Hat is trying to reach with their Advanced Server product, for example). They tend to buy distros that contain a lot of non-GPLed value-add anyway, and will spend big chunks of cash on wider (more "end-to-end"-integrated) products.
But, it's kind of odd to try to go after them with the "standardization" angle, isn't it? I mean, they don't care very much about standardisation as a tool to integrate products from different vendors, because they don't want to spend any time on that anyway. Reaching a minimal cost to fit the requirements isn't really very important to most enterprise customers, when it comes right down to it, as long as the cost is below a subjective "too high" level. So wouldn't it be better for UL to pump out as much distribution lock-in as the customers can handle?
Ransom Love had his anwers posted into a widely-watched public forum (i.e. Slashdot), so he had to be careful what and how he says what he says. Anything he mentions publicly can affect UnitedLinux and/or Caldera in the eyes of the business world. Adapting to the audience is not really possible, at least not to a large degree.
You need to install an RTFM interface.
Comment removed based on user account deletion
Not only that but companies _want_ to pay someone else to do that kind of maintenance. Why? because a fee is a _known_quantity_ - you can budget around it. The last thing the comptroller wants to hear is "Oh, I don't know how much the TCO will be, who cares it was free! Now pay me some value, for an unkown number of overtime hours, to keep it running".
"Sanity is not statistical", George Orwell, "1984"
I agree with the original poster and I also agree that Ransom's motives weren't exactly PR free. Find your happy place somewhere in the middle; welcome to reality.
-AC
This isn't hard to accomplish. The GPL specifies that the source code is "the prefered form of the work for making modifications to it." All you really need to make a difficult to duplicate build system is a new language.
Just because you have the source, doesn't mean you have a compiler for it.
People forget that OSS is a development model, not a business model. Linux has proven the effectiveness of the OSS development model. (So the OSS stuff is just snow, okay?)
... sorta like how Dell, Gateway, and Micron all support Microsoft. Now, the question is how much of a hold will Love keep on this? If he keeps too loose a control, the distributions will fragment (losing effectiveness), if he keeps too much control, the distributions will quit and won't join him.
Now...I think Love has correctly interpeted the business market and recognizes that unless some radical changes are soon had Caldera (and a number of the other small distributions) are going to be out of the market. Right now, RedHat has the mind share and Mandrake the market share. (Debian, much as I like it, is a non-issue for most businesses and is, for the most part, a speck on the business horizon. --- which is, imo, a good thing.)
So, Love proposes (to other soon to be out-of-business distributions) commodizing the OS. This isn't a bad thing (especially for Love if he can get them to agree). This allow other distibutions to differentiate on other things
It'll be interesting to see if he can pull it off.
But then again, considering his alternatives, he didn't have much of a choice.
There is little or no information that needs to be graphically conveyed to a human installing a linux distribution.
There is a tiny amount of information that is more easily assimilated through graphics; mostly status bar and partition size stuff. This can be readily represented with ASCII character-based graphics.
The system requirements of a graphical installer are higher than the system requirements for a non-Xwindows server - which is a significant portion of the linux user base!
Money spent developing graphical installation routines is money wasted. Linux does not need and has never needed a bitmapped installer. The resources would be better spent on something like the Gnome or KDE projects, where pretty pictures have some justifiable use.
I am always amused when some trade rag rates one distribution higher than another because of a bitmapped installer. Shows that the reviewers have the priorities of a five-year old - "ooh, pretty!" and are not competent to judge distro quality.
It's really pretty hilarious to hear the /. crowd talking badly about people talking in buzzwords! Have a look around - most of the people here seem incapable of having a conversation with ordinary people. (Most of who are more intelligent, although less technically inclined that the self-righteous crowd that hangs out here.)
Believe it or not, there *is* a real world out there, and people live and work in it. It's that world that produces the surplus that gets wasted on salaries for ungrateful Stallman slaves.
It's crap like this that really makes me wonder if supporting Linux is a good idea. (after many years of supporting Linux, I am seriously pondering whether it is wise to continue to do so.)
If Ransome Love had any sense, he'd recognize this crowd of backstabbers for what it is and go pursure his plan based on BSD, where at least he has a chance at being successful. Linux is doomed to play along the edges of the market so long as the "community" insists on bludgeoning to death anyone that tries to bring it sucessfully into the real world.
"The future's good and the present is nothing to sneeze at." - Roblimo's last
If all the 'distros' will be based off a common 'universal' cd ( the first disk he spoke of ) and then VAR cds added to the set to refect each companies slant..
Why isnt the *first* disk to be available if its
not got any 'special' stuff that must be supported, as he says is the reason for not releasing binaries..
Ever hear of 'unsupported' download versions?
Sure they need to make money too, but what if i dont care about the VAR parts, but would like to be compatible with this 'standard'...
Does that mean i can download the source to recreate the 'base' system totally?
---- Booth was a patriot ----
The real reason is that under the GPL you are allowed to sell the binaries, but after you have done so, you are compelled to distribute the source (to those customers) for free.
... uncomfortable. So by selling the binaries they are catering to that market.
It is interesting that his explanation is basically unintelligible, or possibly just silly, when there really is a sensible reason.
A subsidiary reason is that for many of their potential customers the thought of doing their own compiling is
Another person has brought up the possibility that the binaries may also have the United Linux trademark embedded in them, which would separately restrict the right of any third party to redistribute them. A further barrier to competition.
I think that there are many obvious reasons for their choices, and his obfuscation is merely an indication of how trustworthy Mr. Love is.
I think we've pushed this "anyone can grow up to be president" thing too far.
Ransom gave this link for contributions from caldera: http://www.caldera.com/developers/community/contri b/
Lets take a look at these.
AIM benchmarks: Well if you are not satisfied with cat /proc/cpuinfo now you can clock your system with benchmarking software that caldera has provided for you.
Caldera Open Administrations System (COAS): these are supposed to be contributions to the entire linux community not a small crappy GPL'ed program they use to promote their own bastardized distro?
CScope: it is amazing how they take credit for this, as SCO (Santa Cruz Operation) released this openly before Caldera had anything to do with them
Java: Now you are taking credit for Java?!?!?
Linux Kernel: and I quote "Caldera has contributed several Linux kernel enhancements, including Windows support, IPX support, NFS, and more". Well I know that I have excellent "Windows" support in my linux kernel, how about you. It looks as if the ipx support is genuine, but NFS has been around longer than caldera.
Lizard: Yet another installation system. but this one seems to be very locked down, with not mention of OS licensing, or any available downloads. Again isn't this supposed to be contributions to the Linux community?
Netscape: Well sure you saved netscape! OS is the only thing that has saved netscape to this point.
NCPS for linux: more netware stuff. Well being that ransom did have a lot of dealings with novell, I guess contributing to this dying networking technology seemed like the right thing to do.
NKFS: yet more dying netware support.
OpenSLP: Well another BSD style license, for I program I'm sure everyone uses everyday.
RPM: quote: "Working with Red Hat, we developed the first package manager." BS! The first and best package management system was the good old tarball. RPM has cause me more trouble than it has avoided me.
Webmin: yet another BSD style licensed project "started" by caldera.
WordPerfect 6.0: Well thanks for paying for the copy of WP for linux that I also paid for. (someone made a lot of money off this deal). WP for linux was so bad I had to revert back to staroffice after using wp for a week. some things are best not shown as a badge of honor.
UDI: vaporware. looks like it was started back in 99, and 3 years later, it is still as un-heard of.
So what we have here, is some loose attachments to BSD style licenses, and a bunch of programs written for the dying netware protocols.
Ransom, look at what the linux community has actually DONE for you, such as what the foundation for your entire company is. and your ticked!!
This guy dosn't stop, so he must be stopped.
I have Win2k boxes under my control with 400 days of uptime, but average about 95% availability (bandwidth, software defects from 3rd parties, etc).
huh? how do you apply all of your security patches on win2k without rebooting??
i have a mandrake linux machine with almost 500 days uptime, and the only thing it needs is a kernel upgrade. oh, and this machine runs NIS/NFS, samba, apache, sendmail and cvs for a 20+ man company... so this machine isn't exactly idle.
Anyone who claims to offer five 9's availability for an applicatio without taking alot of your money first is probably smoking something - it requires redudant power, physical locations, bandwidth, cooling/heating, best of breed software and administrators, and a healthy dash of luck.
well, i agree that hardware plays a large part in total availbility time, but so does the software (OS). we have recently installed vmware's GSX server on an IBM netfinity server. we have 3 virtual win2k servers running under linux. the other day, one of the win2k machines locked up after 3 weeks of uptime, so the virtual machine had to be restarted. when our IT department heard about this, most of the windows admins were like, 'cool, 3 weeks without a reboot.. we rock!'
in other words, the windows admins were not surprised that win2k only lasted 3 weeks before a dump.
on a side note, vmware's GSX server completely kicks ass, and i highly recommend it to all linux admins that have whiny windows IT depts breathing down their necks.
After showing them usuage charts, they realized. Our customers all use the system during normal business hours, all in EST time zone, all very predictably. After a few late stragglers NO ONE had ever logged in between the hours of 1:30 am and 4:30 am. Ever. It had never happenned. So that became maintenance time, everday. Instead of rushing off closely timed scripts to make maintenenace work, I spread it out, did a bit more maintenance than needs to be done daily.
jesus christ, daily maintenence? you really are a windows admin, aren't you? in linux, it's sometimes hard to get a service configured just right (sendmail anyone?).. but once you do, you never have to mess with it again.
using windows, it's easy to configure anything, but then you find yourself reconfiguring over and over and over..how disgusting..
The bottom line is that very few places need five nines. Very, very, very few. And you're not going to get it for free (or for $800 from redhat).
no, you won't get it for free, but i guarentee you can get it cheaper using linux than with windows. that really is the bottom line.
Who's the sucker? The company "stealing" 5000-man-hours code created by volunteers or the volunteers not understanding what the GPL is?
If you, the author of your code, don't like "Big Companies" stealing code, then change your license. Attach a license that more or less says, that no one can use your code commercially(for profit).
IF you don't realize this or don't care(and think GPL is the holy grail of "Open Source"(tm)), then I don't give a flying fuck who's using your GPL'ed code as long as they comply with the license.
As an aside, don't cry about UL not releasing binaries either; as they are adding a "cost-to-use" factor. For example, not many people would install linux, if you had to compile the kernel, compile 100s of packages, configure and write scripts, etc. all on your own.
A few ideas tossed into the mix.
Isn't this just trying to make linux and the whole gpl'd software world into a proprietary product, with all the hassles that come with it? Why would I want to install a closed source binary anyways? My experience with linux and binary applications has been universally negative. Is the solution somehow to try to limit the open and moving free software, where source is available, hence no need for binary compatibility? Now we make our systems like the proprietary software systems, to try to attract some applications that will not work well, not permit local fixing, and probably will not exist for long due to the above. Orphans once again.
Isn't this what we all have fled from, putting up with the odd inconvenience, but enjoying overall a much better experience?
In other words, it is a good idea from a business stand point, but it won't work because these firms are going up against better funded, deeper outfits. The key is changing the rules. Linux has, now they want to change them back.
My view is that the barrier to linux in large corps. is simply due to it not being around for long enough. I don't mean in absolute terms, but a good complete workable system. The graphical interfaces are just now arriving. Has linux been effective and in common use as a server for 5 years? Just about. I was conversing with the admin guy at a small local auto parts manufacturer. I asked what he thought of linux. He has worked with unix before, SCO, etc., so wasn't afraid of that. He said you gotta stay with what has stuck around for a long time, otherwise it costs you a pile of money. Hence microsoft.
This initiative will add nothing to that. Possibly some warts caused by different distribution quirks, but not much else.
Time has been very good to linux/etc. I believe it will continue to be.
Derek
I'm kind of curious as to where the leap happened between
and
I would assume at some point, your organization took Exchange and ran some tests against it to ensure it would manage to deliver what it claimed. Perhapse you even talked to Microsoft and got some information as to what to expect out of their product and good implementation strategies. And you would have sought out reviews and opinions from industry news sources and technical discussion forums.
Many of the same strategies apply to investigating Open Source applications for the Enterprise. One advantage to Open Source is that its community and developers are rather exposed to the public - discussions over implementation, scalability, bugs, and other technical issues are often a google search away. Furthermore, these projects usually offer forums (web forums, mailing lists, usenet groups, etc) where one can seek out informed answers to direct questions. Sometimes these projects are the product of a group who consult on implementing the software - engage them. Or seek out some of the other organizations involved in providing business support for Linux and associated products (such as RedHat, SuSE, or Caldera). There may be enough free information available - or a small enough fee that is easily absorbed as a "cost of business" when compared to future licensing fees (or lack thereof).
The final step is simply testing. After you've narrowed down on a few possible products based on your desired needs (you DO know what your user base NEEDS, right?), underlying architecture, etc... implement it. Set up a testbed. Stress test it. Look for odd performance kinks and usability or administration issues.
Of course, this should be old hat. After all, we know not to trust the glossy brochure. The proof is in the performance. And that holds true no matter if the product comes from a Mega-corporation or a modest listing on Sourceforge.
UL does not change any of this. Granted - it does provide another vendor listing. But this is not something new. Linux vendors, contractors, and independant contractors have been serving business interests and concerns for years.
The reason for no binaries is obviously to control the brand "UnitedLinux".
Say I wanted a preinstalled "UnitedLinux" system, or I wanted a prepackaged "UnitedLinux" distribution.
What this does is prevent "Bob Schmoe's UnitedLinux distribution", since Bob is not able to legally use the trademark on a binary release compiled from their sources. Only binaries from the participating companies can use the trademark. Bob would have to "buy in" to obtain the same rights (Ransome said as much).
This puts UnitedLinux in relative the same boat as the UNIX trademark, where use of the trademark is controlled by licensing fees and/or buy-in to the club. The main difference appears to be that the base system is, in fact, compatible -- at least until you load on the vendor "value add" or try and load software from a third party that depends on a particular vendors value add (this is the same problem all non-RedHat Linux distributions already face today, since RedHat has used the same embrace-and-extend technique).
The consumer benefit to this is that if a software package that depends only on the UnitedLinux base system is shipped by a software company, they are guaranteed that the software will run on any UnitedLinux system.
The LSB fails to provide the extended features of UnitedLinux, and therefore there is a barrier to entry for LSB compliant systems, and even systems derived from the UnitedLinux code base, in the corporate market.
Or, in one sentence: it is a marketing play.
I actually rather doubt that there is a liability issue involved for UnitedLinux vendors, unless they (some day) pool their technical support services. The only liability left is on the part of software vendors, who want to guarantee that their applications will "run on UnitedLinux". If each vendor guarantees that, then a third party hacked up version could potentially fail to run the tvendor's application. So it's also a minor amount of risk mitigation for the vendors they hope to attract to get them to port software to the combined platform.
Most likely, what will happen instead, is that each vendor will have it's own support services for their product, and each vendor will embrace-and-extend UnitedLinux -- the same way that the UnitedLinux vendors have done the same thing in regard to the LSB.
Unfortunately, "standard, plus extensions" is, by definition, non-standard.
I don't think this is really a reesclation of "the UNIX wars"; GNU already lit that fuse when they started embracing and extending command like "tar" and "gprof", etc., with options that are not present in POSIX, and with more than single character options, and with expression reevaluation in GNU Make, etc..
What this works out to is a shot across the RedHat bow. It says that RedHat is not going to be permitted to own the distribution standards for Linux.
Frankly, someone needs to shoot at FreeBSD the same way (IMO).
-- Terry