Why Open Source Isn’t Succeeding…part II

*Editors note: The following sentence was removed from the first paragraph of #3:

“Even as recently as November 2004, there has been talk from the kernel developers about a fork in the kernel”

It was removed along with the link to an article from November 2004 due to it being an erroneous reference and based on comments taken out of context. However, I neglected to post that I had removed this sentence because I felt that it did little to support or not support the paragraph. I apologize to the readers of the blog for calling into issue my integrity.

Devnet



Some of you may have read my previous entry that sparked quite a bit of debate. Looking back on the content, I realize that the title of the entry could be misinterpreted as FUD or even trolling. Please understand that this wasn’t the intention. The entry simply addressed issues that I see inhibiting open source, specifically Linux, from fully succeeding (i.e. dominating both the server and desktop market for computers). I should have titled the article, “Why enterprise applications may kill open source”. But hindsight is 20/20 right? On request, I will clarify a few points for those that have asked it.

1. Enterprise companies and applications that take from open source CAN KILL open source.

There is a crossroads in today’s enterprise OS. Micro$oft has pushed back Longhorn and the next greatest server application. Companies have begun to wonder what they are going to do for server/desktop OS in the near future. Many are seeing their support for enterprise server applications such as NT 4.0 and Win2k dry up and blow away. Enter Linux. Affordable, stable, and now certified as a ready alternative to M$. So Enterprise Linux begins to flourish this year. Novell, Red Hat, TurboLinux, and others start to churn out a profit and go into the huge server market with actual products that can offer benefits to all. The problem is this…what happens if those companies pull the plug on their open source support? Would they? Wouldn’t they? Why should we count on them? Didn’t Lotus 1-2-3 and OS/2 count on Micro$oft to keep them in the loop as well? Is it really something the community should bet on? Will the community bet on it? I hope not. Will it crush the community entirely? No…but it could fork open source or even set it back quite a bit. Of course, no one can see into the future, but these are valid questions to consider when you bring enterprise applications and business into the open source mix. Most of this will probably mean nothing for the common desktop Linux user or even someone who uses one or two open source applications on their M$ desktop. So why discuss it? Hindsight is always 20/20 right? Why not make foresight 20/20?

2. What do you mean that these companies don’t give back to open source. After all, IBM gave X Million dollars of support back to the community and Sun released X # lines of code…

Yes, that is true. Money being given back to the community and code being released is a good thing. I hope it continues. If business enters the fray, can you really count on it? What if companies decide it isn’t such a good business idea anymore to give back to open source? Will we cry foul and expect them to listen? Instead of investing our support for these companies…I say we should invest our support for those distros and software that aren’t available for enterprise applications. The free as in beer stuff. You know, those that charge 10 bucks to help the author who’s taxed beyond his means and has taken out a mortgage on his home just to put out the last release. Those are the ones we should cheer…not corporations. If I were rallying behind a business when I began with open source in 1995 I would have gotten shunned out of every single BBS and channel I was on. When did it become cool to rally behind business?

A corporation that sells enterprise open source will try to advance it’s own means first and then that of the open source community that supports it. If the open source community gets trampled or the short end of the stick…so be it. If the corporation sees an opportunity to take more than it gives to open source…it will happen (and most likely has happened). Remember that all they have to do to get accepted back into the community is release another few lines of code or donate a fraction of their billion dollar profits…it’s all smiles and “welcome backs” after that.

The problem with business is that business as a whole is incompatible with the spirit of open source…which doesn’t mean they can’t help each other or coexist…it just makes for an unknown future. Right now, companies have found a comfortable balance with open source. This is proving a very rich environment that open source is flourishing in. If FUD isn’t just something spoken of but something that becomes a reality, then where will we be? How do we prevent it? By being aware that it is an issue and NOT putting all of our ducks into the proverbial enterprise application row.

3. What’s this about Linux forking?

Sometimes forks in major projects can be a blessing. Sometimes though, they can kill a project. So, it’s uncertain what would happen if Linux forked. If you’re thinking…hey, nothing in open source will fork…read this and reconsider things…it’s not an impossibility. The good part about a Linux kernel fork is that open source wouldn’t die. Linux might suffer quite a bit, or it might not…but open source Linux would survive. However, if Linux forked it would be used as a “I told you so” by so many FUD brewers (like your favorite and mine Redmond micro-brewery) and with this happening, overall support would most likely suffer. Of course, this is all speculation. It’s not something we should be afraid of. It’s something we should be INFORMED of. It’s not something that should be uncertain…it should be understood. It’s not something that should provide doubt for us…it should provide knowledge of the possibilities.

There is a possibility with businesses supporting Linux that Linux will fork because of decisions that the business makes. If something the community wants conflicts with what the business wants…what is the business going to go with? Will they remove their support when they decide that they’re going with what they want instead of the community? Will they put undue pressure on individual developers in order to sway the development in their direction? Who’s to say they won’t?

Rightly so, we can’t see the future and we can speculate all day long. But we can change our awareness now and we can adapt ourselves back into the original intention of FOSS instead of nipping at the coat tails of businesses and having misplaced alliances.

4. LSB 2.0 (Linux Standard Base)

With the creation, perhaps evolution, of Linux Standard Base to version 2.0 and the creation of the Linux Core Consortium…we find an interesting way for Linux to become a viable corporate component. Not just companies either…anyone can create a Linux distro that is LSB 2.0 certified. The thing that bothers me about this is if businesses make the idea of LSB 2.0 mainstream will it lock Linux in like Micro$oft’s API did? Some people believe that now is the time for Linux to stop emulating Micro$oft and to start forging ahead. This article states that with the next release of Micro$ofts operating system…the API will change rendering previous M$ versions of programs and apps completely useless. Why is this important to Linux? It would be important because Linux would be able to pick up users left in the void. But does Linux need LSB 2.0 to do this?

According to the LSB 2.0 specification, “The LSB contains both a set of Application Program Interfaces (APIs) and Application Binary Interfaces (ABIs). APIs may appear in the source code of portable applications, while the compiled binary of that application may use the larger set of ABIs. A conforming implementation shall provide all of the ABIs listed here. The compilation system may replace (e.g. by macro definition) certain APIs with calls to one or more of the underlying binary interfaces, and may insert calls to binary interfaces as needed1

Now let’s take a look at what an API actually is: “One of the primary purposes of an API is to provide a set of commonly-used functions?for example, to draw windows or icons on the screen. Programmers can then take advantage of the API by making use of its functionality, saving them the task of programming everything from scratch. APIs themselves are abstract: software that provides a certain API is often called the implementation of that API2“. Anyone can tell you that the first way Microsoft ever locked in vendors was with the release of Win16 followed by Win32 and soon to follow WinFX (Longhorn). Windows API’s help proprietary lock-in. Yet we are applauding it and supporting it in open source? Is there a reason why the only package format mentioned in LSB 2.0 and 2.0.1 is the rpm? It really makes one wonder if LSB 2.0 is going to ‘revolutionize’ Linux for everyone or for just a certain select group.

Some of you might think that I’m exaggerating. Some might think I’m only here to cause strife and to go against FOSS. Quite the contrary…I feel that I am helping to preserve FOSS by writing this. I would like people to be aware that LSB 2.0 might not be the fantastic ride that everyone is saying it will be. While I don’t think it will completely wipe out FOSS, it could really hinder development if used against the community…if you disbelieve this, allow me to give an example.

Let’s set up a hypothetical situation. Say that Tom is a software developer who writes a stellar program for Linux that starts to garner him some attention. His software is quickly adopted in both KDE and Gnome desktop environments and begins to be included in default installs for Linux distros. However, Tom does not have the resources to port his application over to LSB 2.0 compliance. So, company X comes in and takes the code that Tom developed and re-releases it for their own OS. Will they give that code back to Tom? Will they employ Tom to develop for them? Will they donate to HIS original cause (i.e. developing HIS software)? Or will they just take this and say that “well, we give enough back to the community because we donated X lines of code or Y applications as open source.” In this instance, the fact that Tom wasn’t able to get up to LSB 2.0 standards inhibited his program from fully realizing its potential. Will these situations make developers look into adopting a different Licensing scheme such as the creative commons or others? Only time will tell. One thing is certain; the danger of things like this happening because of LSB 2.0 is there.

5. The real threat to FOSS is YOU!

Definition:

Main Entry: 1threat
Pronunciation: 'thret
Function: noun
Etymology: Middle English thret coercion, threat, from Old English thrEat coercion; akin to Middle High German drOz annoyance, Latin trudere to push, thrust
1 : an expression of intention to inflict evil, injury, or damage
2 : one that threatens

3 : an indication of something impending <the sky held a threat of rain>

As an avid open source contributor and user…I do not threaten nor intend to cause injury, evil, or damage to open source. I do however, try to inflict awareness, knowledge, and informed opinion toward other users of FOSS.

6. But you IDIOT! Open Source IS succeeding!

Agreed. It is currently enjoying a surge in success with applications such Firefox, OpenOffice, and various Linux distributions. However, open source is still used on a minority level by the common end user. It may be succeeding in the opinions and perspectives of those that use it on a daily basis…but overall, it’s still not successful. Take your common computer user for example. They go down to Wal-Mart or CompUSA and pick up a 500 dollar desktop. They come home and hook it up to their dialup connection and begin to surf the web. Of course, they don’t install any huge programs because that would require them to be online all day downloading it. Not only that but your average person doesn’t even know what open source is! I took a quick poll of the 40 people I work with (I work Electronics for the Military) and only 14 people even knew what open source was!

The bottom line is…open source hasn’t had the amount of exposure that it deserves…and probably never will because it isn’t a business (thank God). The common computer user still isn’t savvy enough to understand what open source is and how it can benefit them. Open source is winning small battles currently through programs gaining popularity…but isn’t successful overall. When common users a) know what it is and b) use it as opposed to proprietary alternatives…then and only then will open source be a complete success.

6. Proprietary vs. GPL

I’ve stated previously that if DeveloperX develops a program under the GPL and CompanyX snags it, stamps 1.0 on it, makes small changes, closes the source, and releases it under their own license…that they could possibly bring developerX to court later. Most of you told me I was crazy because the GPL/LGPL/SISSL or whatever license FOSS was licensed under would protect the developer. Would it? When was the last time a company brought a small developer to court? Would the small developer even last out a large legal campaign (Remember that a legal campaign for patent defense costs over 2 million USD)? Before you answer, consider the following:

  1. We have a patent system that is corrupt. Anyone can sue anyone for patent violation even if they do not own the rights to what they are suing over. Think SCO on this one. Those guys have no case and still they are allowed to sue draggin people into a multi-million dollar lawsuit.
  2. Even industry standards can be patented…look at TCP/IP, small revisions, and CISCO.
  3. With a lack of central ownership of Linux…businesses might see a gap in which they can get an ‘inside advantage’ through picking off small developers one at a time. With the lack of corporate representation…small developers are ripe for the shooting.
  4. Small open source developers can be sued for just about anything at anytime…think mambo (now I know it wasn’t brought to court…but what was stopping the idiot that brought it up from taking it to court? Nothing. He could have had he wanted to…what a crappy patent system we have).
It’s also important to take note that open source and the GPL have never been tested in the courts. We can’t honestly say we know what will happen when and if the GPL is brought up as copyright infringement on a small scale. I’m not talking SCO here…they’re the big dogs…I’m talking…what if Micro$oft brought open source developerX to court? How long would that last? I don’t know of any developer that has ever had to defend the GPL in court. Who would want to be that guinea pig?

In Closing

You may not like what I have to say. You may not agree with what I say. However, there are those out there that share the same opinions and concern that I have outlined in this entry. Most of the people that express this concern have been involved with open source since the inception of it into mainstream technology. Most that disagree use only a few open source applications or run one version of desktop Linux and have been involved with open source for a few years. You can state that concern like this is misplaced and not worth the read. Remember though that I write this not to turn people away from open source…but to inform those that use open source that there is a change in the tide for Linux. I hope that everyone takes this entry as such.


“To be prepared against surprise is to be trained. To be prepared for surprise is to be educated.” James Carse.

This content is published under the Attribution-Noncommercial-Share Alike 3.0 Unported license.

About

devnet has been a project manager for a Fortune 500 company, a Unix administrator, a Technical Writer, and a System Analyst during his 10 years working with Technology.

  • rm6990


    Sometimes forks in major projects can be a blessing. Sometimes though, they can kill a project. So, it’s uncertain what would happen if Linux forked. If you’re thinking…hey, nothing in open source will fork…read this and reconsider things. Even as recently as November 2004, there has been talk from the kernel developers about a fork in the kernel. The good part about a Linux kernel fork is that open source wouldn’t die.

    And to rebut you continued fud about forking, read this.

    http://www.groklaw.net/article.php?story=20041121124609671

    People posting fud without even contacting the maintainers/developers of the project hurts Open Source just as much. Please do your research. A simple email to Andrew Morton would have clarified this. I like the part from my link

    “We did this for the 2.0 kernel series, the 2.2 series and the 2.4 series. One day we’ll do it for the 2.6 series.”

    Again, please do more research.

  • http://n/a dd

    The article you site regarding Linux developers talking about forking the kernel involved a major misunderstanding on the part of the author, which you obviously share. They were talking about a fork to the 2.7 developement branch, a normal cycle rather than a fork in the sense you are thinking.

  • http://Thisone devnet

    I don’t really need to do more research. I went back in and underlined for you the MAIN POINT of each paragraph under the Forking section. Read those and you’ll understand that I not saying that Linux will fork from inside…but I am worried that it will fork BECAUSE OF THE INFLUENCE OF BUSINESSES.

    Please remember that when you read a paragraph that the first sentence is usually the main idea and that you should take it as the main point of the paragraph with subsequent sentences adding support.

  • http://Thisone devnet

    Damn, I really thought people who spoke English knew that the first sentence in a paragraph was the main idea.

    Tell me again where I say the Kernel is going to fork?

  • fmcgowan

    1. The GPL _specifically_ _forbids_ taking pre-existing GPL code and closing it, unlike the BSD license.

    2. The GPL _specifically_ _requires_ providing the source of a GPL application _that_ _you_ _distribute_ whether you have contributed code to it or not.

    3. Forking is not necessarily bad, but stealing someone else’s code (as in #1 above) or failing to fulfill the contractual obligations that _you_ _agreed_ _to_ by distributing GPL code, as in #2 are.

    That about covers it.

    While I see paranoia as a survival trait, I see _unreasonable_ paranoia as a handicap. Forgive me, but you seem to be concerned about companies doing things that Eben Moglen is retained by the FSF (or is it GNU.org ?) to deal with. He seems to be a _very_ talented guy and I am certain he will handle it if necessary.

    Relax.

    Drink something without caffiene.

    Relax some more.

    Repeat as necessary…

  • John Jacob J. Schmidt

    “Say that Tom is a software developer who writes a stellar program for Linux that starts to garner him some attention. His software is quickly adopted in both KDE and Gnome desktop environments and begins to be included in default installs for Linux distros.”

    Go Tom! :-)

    “However, Tom does not have the resources to port his application over to LSB 2.0 compliance.”

    Actually, there are three main possibilities: Tom enables LSB 2.0 compliance, an open source group (led by Tom, or led by one of the distros including his app, or random Fans of Tom) enables LSB 2.0 compliance, or some firm does.

    “So, company X comes in and takes the code that Tom developed and re-releases it for their own OS. Will they give that code back to Tom?”

    Does Tom care?

    If Tom does, Tom darned well better have chosen a license that increases the odds of this happening, such as the GPL. While the GPL isn’t a magic elixir, it does provide some basis for Tom expecting the modified code to be released. Company X can be prosecuted twice: in court (by Tom, by FSF if Tom assigns joint copyright or otherwise gives FSF legal standing, etc.) and in the court of public opinion (various firms have been induced to release code under the GPL that they had kept hidden when enough noise was made). It’s entirely possible neither will help (can’t afford to bring suit, and not enough noise is raised), but, then again, life ain’t perfect.

    If Tom doesn’t care, and releases his code under BSD or MIT/X or some other non-copyleft license, then it’s unclear why you or anyone else should care what Company X does with the code.

    “In this instance, the fact that Tom wasn’t able to get up to LSB 2.0 standards inhibited his program from fully realizing its potential.”

    That’s an artificial line in the sand.

    If his program doesn’t support alternative user interfaces, it will miss out on blind users, and thereby fail to realize its potential. If his program doesn’t support internationalization and localization, it will miss out on users who are not native speakers of the language Tom speaks, and thereby fail to realize its potential. If Tom’s a kick-ass coder, and can’t document his way out of a paper bag, his program will miss out on less-skilled users and thereby fail to realize its potential. If Tom doesn’t port his program to run on the TRS-80, his program will miss out on some users and thereby fail to realize its potential. Company X, or an open source team, can address any or all of these problems, should they decide to. LSB 2.0 isn’t distinctive as an issue…other than it fits better with your anti-enterprise pitch.

    “Will these situations make developers look into adopting a different Licensing scheme such as the creative commons or others?”

    It’s unclear why the GPL doesn’t fit your needs. It’s not tested in court, but 99.44% of licenses haven’t been tested in court (remember, pretty much every proprietary software EULA is different from the next), so the GPL isn’t any riskier than any other at this juncture.

    “One thing is certain; the danger of things like this happening because of LSB 2.0 is there.”

    Just as the danger of this happening because there are blind people. Or people from other countries. Or people who can’t grok his program. Or people who think the TRS-80 is da bomb.

  • Brad Spangler

    I honestly don’t want to sound overly critical here, but it genuinely sounds like you really don’t understand the GPL. I’ll leave aside talk of other licenses for now.

    The GPL creates a framework for voluntary cooperation, among individuals and/or companies, based on very clear obligations being voluntarily incurred in return for very precise benefits.

    In order to redistribute modified versions of the code already licensed under the GPL, the modified version must be similiarly released.

    That statement basically summarizes all of the benfits and costs of the arrangement. Someone or some company is a member in good standing of the community if they are honoring the terms of the license.

    Copyright law and the GPL are quite clear. All a company or individual has to do is obey the law (and, by extension, the GPL). Anything else given to the community is strictly voluntary — a gift, not an entitlement.

    If you want to posit an *additional* set of obligations that should somehow also be incurred by corporations (and presumably individuals?) under other terms, then you have a lot more in terms of ground to cover, details to specify and rationale to argue for.

    That said, of course GPL code will fork at least occasionally. The GPL specifically *enables* forking, by granting a conditional right to modify the software. That’s freedom. (shrug)

    If you want to come up with a license that allows source code to be viewed but not modified, you can do that — but it wouldn’t be open source.

    With regard to the LSB, it would be interesting if you chose to argue that the *process* for setting the standards of the LSB is to open to manipulation favoring major players. If true, it would be a legitimate concern.

    I don’t know if that would hold water or not, but it would be preferable to what you’ve done on that point — which semantically boils down to “all standards are bad”. Speaking only in terms of voluntary standards, I’d have to disagree.

    Just because the Windows API is a standard and the LSB is also a standard doesn’t mean the two are both Bad Things(tm).

    The Windows API has been and will continue to be defined by Microsoft to suit their purposes. The problem isn’t that it’s a standard, but rather, how it is set.

    I don’t know enough of the ins and outs of how the LSB came about to say that it is definitely a more open and fair process — but considering that we’re talking about the Linux community, that would be my natural expectation, anyway.

  • Anonymous

    “3. What’s this about Linux forking?”

  • Patric Conant

    So, some large company with deep pockets is going to count on being able to win a cour case against “Tom” and invest money in developing on top of a stolen code base, where they will sell something and provide concrete numbers for thier benefits from “Tom”‘s code. And this is a viable business plan? Nothing anywhere gives anyone the right to take someone elses creative work and claim it as thier own, and if it isn’t thier’s, then what is thier agreement with the author? Proprietary code is in much greater danger of being open-sourced once its proven to be a derivitive work, then open source is of suddenly being the bread and butter of a large company. Microsoft (they aren’t the bad guys by the way, it isn’t thier fault that people are purchasing pretty boxes instead of quality code, they just provide them) would be in worlds of hurt if GPL code was found in thier products, as the penalty would be pretty clearly be them opening up the source of the product. So if you are Microsoft, a leader among amoral software companies, you don’t spend your time perusing GPL code to see if thier is something you can steal, you spend your time writing code, and seeing how you manipulate the legal systems of the world into protecting your business. This sort of thing has already happened though, and was dealt with promtly and easily. A large manufacturer of network hardware apparently commissioned a net appliance from a country where labor is cheap, and they recieved these appliances, stickered with thier name and distributed them. A few whily individuals noted that the binary bore the signature of a linux kernel, and requested in writing the full source, from the hardware manufacturer, who had no idea that thier product contained GPL’ed code. They contacted the company they commissioned to create the thing, inquired as to the nature of the source code, the company in question played dumb for a while, realized they were caught and promtly released the source, along with thier changes. I would think it would be a less than lucrative business plan to steal GPL’ed code and hope noone noticed. Forking won’t create any killer products that will cause the world, to abandon its open source root, paticularly since the closed fork, by definition would lose the value of its open source counterpart. Big business shouldn’t be pushed out of the community simply on the basis that they are large, and they are significant part of the community. As for open sources success, how many more people have to use open source than proprietary code before its successful. Every windows machine in the world has open source software on it, every internet attatched device in the world has open source code in it. Your definition of success seems to be centered the amount in which it is a household name, I am not sure why this is significant, but lets suppose it is. I would guess that per marketting dollar put behind it, linux and other open source products have about 10=100 times more consumer awareness penetration than thier proprietary counterparts. Lets see, how many more people know about what Internet Explorer is and don’t know what Netscape.

  • http://Thisone devnet

    Oh yeah, that’s the point where I respond to OTHERS commenting on my first entry about the kernel forking. I see where you can get confused. Try reading the underlined sentences aka the MAIN IDEAS of the paragraphs contained underneath that point. Then you may understand..

  • http://Thisone devnet

    Somehow, companies have always found ways around license issues, patents, and anti-trust laws. You call it paranoia and I call it being informed. One can never have enough perspective on an issue…unless you wanted to limit yourself to only one point of view.

  • Quinn

    Remember these things…

    When was the GPL defended in court? Has it held up to anyone yet? Has a major business come in and gave anyone using the GPL a microsoftian ‘nudge’? Individuals don’t stand up to companies, patents don’t stand up to companies, why would I think a license that is untested in the court system could stand up to a company?

    Devnet, you have a point of view that many don’t take. It’s refreshing that not everyone is brainwashed into thinking everything is secure and ok. Being prepared for anything before it hits is always a good thing. Hats off to you man!

    Quinn

  • Brian Vincent

    *yawn*

    I’m struggling to find any original thinking in this editorial. I think I’ve seen all of those arguments a million times going back to about 1995.

    Move along.. nothing to see here.

  • http://english.webworks.se Uno Engborg

    Sorry, but you seam to missurderstand GPL somewhat. If GPLed code somehow is merged with a closed source code, this doesn’t mean that the closed source has to be released as GPL. It means that the the GPLed parts needs to be removed and replaced with other non GPL code, or else it will be illegal to distribute it.

    The other option is to GPL all of it. But this is just an option, not something that automagically follows being caught red handed mixing closed source and GPL

  • http://english.webworks.se Uno Engborg

    The only thing the auther has a clue about is that software patents is a problem. But he really misses the point.

    Software patens is a problem to the whole software industry, not just FOSS. In fact I would guess that propriatory software would be in even more trouble. Just consideer this. If you were to sue sombody over patents would you rader sue sombody with very little money, or would you go for sombody that actually was likely to cough up some money before goint to chapter 11?

    You could expect some companies making lawsuits for anticompetitive reasons, but that could happen to closed source as well.

  • oskar

    i think u dont see the whole picture of open source vs big business and thats why u have so much apprehension of big business…open source for many if not for most people is more of a pastime or hobby in the same way as we have car enthusiasts or radio amateurs…the number of people who enjoy software development for its own sake is increasing all the time which finds its expression in the growth of open source communities…its precisely the possibility to tap into this source of energy and creativity that motivates such comapnies as IBM or novell to switch to linux…

    big business is not interested in destroying open source at least not such companies as apple IBM and others…their reliance on open source is what allows them to reduce development costs …any company which learns to work with open source communities gets a competetive advantage over others due to the fact that it has hundreds of extra developers on its side for free…

    i would go as far as to say that in the future this industry will be dominated exclusively by the companies who learned to work with open source and have created big and friendly communities around them…its not eigher communism or capitalism …it s some fusion of the both…

  • Larzi

    ABI and LSB 2.0

    ABIs and standards for building Linux systems are by no means a lock in. In general, ABIs are more a blessing than a curse. In reality its more a question of which ABIs you are going to use, and not if you are going to use them.

    It is of course important to discuss whether only implementing RPM as package format in the LSB 2.0 is particular sane idea. But thats an entirely different question.

    ABIs only become lock-ins when they are used for anticompetitive reasons, ie:
    a) they are not documented properly to 3rd parties
    b) are changed at will to break compatibility

    I believe this is what Microsoft has done in the past. I don?t see this happening with the LSB 2.0.

    The only thing I really agree with you is that software patents suck.

  • Quinn

    “ABIs and standards for building Linux systems are by no means a lock in.”

    Not if used right….something the article is trying to state…ABI’s and APIs _can_ be made a very locked in thing as evident by microsoft and their standards. Will Linux do it? Who knows…but devnet is right at least on this aspect.

  • Anonymous

    You have? Wow…you are the most informed individual on the face of the planet to have heard all of this by 1995! You’re my hero! Will you tell me _everything?_

    Why don’t you pull your head out?

  • http://Thisone devnet

    Actually, I do see the ‘whole picture’

    Working for a company with over 500K employees makes me see it.

  • http://snoop.alphanetch mussi

    In reply to:

    4)

    In case you didn’t notice, the GPL is enforcable in European courts, and no software manufacturer will want to fight a few developers as money over there doesn’t buy the courts (yet), and European courts tend to be much more consumer-friendly than the US judicial system when it comes to obvious things like the GPL.

    And since no company wants to forfeit the EU market, they’re going to comply.

    Linksys had to feel the pain. Else, try to get at Russians – over there, suing for patents, copyright and DMCA-style stuff is pretty futile as the number of P2P websites clearly shows.

    3)

    A fork can be useful for some specialized appliances, but I doubt it would be of any use for general-purpose kernels. And since the Linux kernel is GPLed, I doubt that anybody would produce a closed-source fork – the reasons for this being outlined above.

    What I’m fearing much more is that BSDish operating systems get abused for things like routers.

  • Patric Conant

    Please quote to me the portion of the GPL that says you can have your derivitve works back, after you have taken code published under this license and illegally distributed them as your own. I don’t think I am the one with a misunderstanding of the GPL. I think you misunderstand the underlying point. For a company to do something of this nature there would have to be a payoff worth the risk, at least in someone’s mind. The GPL makes it very unlikely that anyone would do that rish analisys and conclude they would take thier chances. If there is money to be made by it, then there is money to be lost in court over it. Not sure how to further reply, as the reply was an attack on my understanding and not on the facts involved, is there some president of someone stealing GPL code modifying it and getting caught, where they got to keep thier modifications?

  • will cope

    your premise for open-source’s failure is that it isn’t “dominant”. that’s a very MSish attitude, no?

    for someone not trying to do the FUD thing you certainly are quite good at it. if my first exposure to open-source/Linux had been your postings I’d likely ran like hell was after me back to the MS empire and begged for BG’s forgiveness.

    as for the possibility of any or all of the things you’ve mentioned happening, well, anything is possible. should the open-source community spend the time it appears you have in fear of their coming to pass it would no longer be able to produce anything usable and so become irrevelant anyway. you either fear the monster under the bed, or you look under the bed to see if its real.

  • Anonymous

    I thought your article was interesting,
    but the only place where I disagreed and
    agree with you was your section of
    criticism to the LSB. I don’t like the
    fact that the LSB standardizes RPM as the
    packaging format. I think DEB should
    be included(I’m a debian person), because
    RPM is rather RedHat/Suse-centric(enterprise).
    But also I think there are tools that do
    such a conversion anways. The rest of the api
    issues I’m fine with, as long as it is updated
    in a timely fashion. I don’t want to be stuck
    writing in some old api that is never updated.

  • fmcgowan

    Over 500K employees? The GM website says GM has 325,000; how many companies have more (approximately 50% more, actually) employees than GM?

    I _think_ you _may_ have overstated the size of your employer by just a bit…

  • fmcgowan

    The LSB 2.0 API and ABI set are _open_ by definition. Hence, “lock in” involving them would only lock you into an _open_ platform.

    Not much of a lock at all, really…

  • Zippy the Penguin Head

    I disagree with your logic. Who, precisely, is “taking” from Linux when company X pays $$$ for RHEL instead of paying nothing for the Linux source? How would Linux be better off? Do you really think Company X would then use those $$$ to fund internal Linux development? Isn’t it better if Company X pays $$$ for RHEL and RH pays a few developers? I mean, even if FOSS developers only saw a single dollar, wouldn’t that be a dollar that they wouldn’t otherwise see? The correct answer AFAIK is “Yes, it is better that companies like RH charge for Linux and reinvest some amount.” In fact, I think you’ll admit that RH reinvests a lot more than $1.

    In your previous post, you called companies that take without returning “parasites” but that’s incorrect. The term is FREE RIDERS and this problem has been studied by philosophers, psychologists, economists, and sociologists for at least a couple hundred years… go look up David Hume and Adam Smith or other literature on the problem before you go posting wild speculation.

    IIRC, I’ve actually seen a paper taht specifically analyzes the issue of free riding in FOSS and found that because of the low transaction cost of copying the software, products like Linux can stand very high levels of free riding. All you need is a core of people contribute…. which is what you had before there was money in Linux… and you have it moreso now.

    I think you ignore a counter argument… do you *really* think that the Kernel developers would tolerate a flood of people who showed up sullen and without the needed skills to participate? Of course not. I think the vast majority of the kernel developers participate because they love doing it and it’s therefore it’s own reward.

    Further, I think most FOSS developers are gratified when their work is used. I don’t think they see it as free riding at all.

    So, I don’t get your point about companies “taking” from Linux without giving back. I’m actually skeptical that you have a point.

    So, going back to your first point… what is your point? I think your are concerned about companies like RH offering support and then folding; leaving Linux customers in a lurch? Geez… so the solution is to never offer paid support? That is *SO* ass-backwards and shows that you have _no_ _clue_ why/how companies choose IT solutions.

    Anyway, it is patently clear that the *FAR* greater risk to FOSS is that the RH/Novell/IBM/Sun/SGI/etc gray train were to derail (or even slow down) and the Linux developers had to get jobs coding proprietary software. I can *assure* *you* that if RH folds, the people left in a lurch will blame RH, not Linux. And I’m sure they will find switching to Novell or Debian (or any of the “white box” RHEL clones… this is OPEN SOURCE dude) faster, cheaper, and easier than downgrading to Windows.

  • Zippy the Penguin Head

    Not a lawyer but I agree with Patric Conant… if you mingle your code with GPL code than I think you could be forced to GPL your code. But here’s the thing, I bet you’re *A* *LOT* less likely to get hurt by the GPL copyright holder than if, say, you cross SCO or Microsoft or some other bandit company. Would the FSF or OSDL or Torvalds have the $$$ and incentive to sue you into oblivion? I think they would only resort to lawyers to protect their own since they’re not competing with anyone. Of course, if you take IBM’s GPL’d code…

    BUT, read the original point 6 again. The author is saying that Company X can steal Tom’s GPL code, call it their own and then *Company* *X* sues *Tom*. WTF!? For what? I’m *REALLY* interested to hear how this could be possible. I think that’s a far-fetched scenario…

  • Zippy the Penguin Head

    One more thing. In your first point 6 (you have two), you make some mistakes. First, the average user in the US now uses broadband. So downloading a huge program is not so silly. Second, Walmart.com is one of the few places where you can buy a pre-installed Linux computer… so not sure why you brought them up but they’re a major conduit to spreading Linux desktops to the masses.

    Finally, FOSS is catching on. My mother–who, I swear, did not know how to operate the *BACKSPACE* *KEY* –asked me to install Firefox (she called it “The Mozilla” ) on her new computer because she’d heard that it blocks pop-ups. So there’s the real recipe for FOSS adoption: (1) make a better mousetrap, (2) make sure people can easily get it, and (3) market the trap.

    FWIW, I think desktop Linux isn’t spreading like wildfire because it fails all three of these but especially the first… Windows works (fitfully) for most people most of the time. But people will adopt Linux if companies like Red Hat craft a distro that fills their needs (locked down, remotely upgradable, stable, etc.) and it gets marketted.

    So, I think you have your facts backwards… it’s companies like RH that will *allow* Linux to be adopted.

  • Anonymous

    Broadband…nope…try again. Cite figures and you’ll see…it’s still dialup.

  • Anonymous

    Think DoD

  • fmcgowan

    OK; DOD qualifies. According to their site:
    1.4 M active
    654 K civliian employees
    1.2 M Guard & Reserve
    2.0 M retirees

    OK. The original use of “company” made me think more of “commercial enterprise” than “arm of government.”

    In the end, though – so what?

    The original author and several posters seem to see the situation as “big business vs OSS”, much akin to “business vs force of nature,” and that is simply _not_ the case.

    “Business” – large or small – is neither for nor against what is essentially a tool development paradigm. In general, businesses either _use_ OSS or they do not; they either _build_ (contribute to) OSS or they do not. A specific business, like MS or Oracle, will either see OSS as a way to build their business (Oracle) or a way to lose business (MS). Each sees the situation clearly, they just see it in terms of their own interest.

    In terms of _my_ interest, I see OSS as a way to free myself from the MS upgrade treadmill while retaining the ability to perform basic business functions. I _also_ see the situation clearly.

    The author of the orignal article does _not_ appeat to see the situation clearly. It is not a case of whether the banana tree (OSS) is a healthy development for the mango tree (closed software) but whether the banana is a viable mango-alternative for the monkey (end user). I say it is.

    Because the source is open, any OSS product will remain available for me to use as long as either someone else continues to maintain it or as long as I can hire someone to maintain it for me using the source code I downloaded. Further, if the source was released under the GPL, a vendor _cannot legally_ take the source and close it again, like can be done under the BSD license as I believe Sun did with one of the BSD Unix kernels. (Nothing against Sun, they did nothing prohibited by the original license and they have done rather well with the resulting products.)

    Not so for closed products.

    Purchased commercial packages become unavailable all the time. How about Multiplan? That is a Microsoft product that is _so_ old, a representative of my employer seeking to know its Y2k compliance (5 or 6 years ago) was told to contact the original vendor because MS could _not_ be responsible for someone else’s products. Oddly enough, Multiplan has capabilities and properties that make it more useful for our staff that Excel. Too bad; when it no longer runs in a Windows console session, we won’t be able to use it.

    How about P2P networking with Lantastic? Is it still available? There must be tens of thousands of commercial products that people used in their businesses that the vendor no longer supports.

    _That doesn’t ever need to happen to you with OSS_, though it might on some particular product.

    That is one of the main virtues of OSS for most businesses, whether they realize it or not.

  • Zippy the Penguin Head
  • http://www.linuxmagazine.com.br/ Rafael Peregrino da Silva

    On Dec 28 2004, 00:10, Quin wrote:

    > When was the GPL defended in court?
    > Has it held up to anyone yet? Has a
    > major business come in and gave anyone
    > using the GPL a microsoftian ‘nudge’?
    > Individuals don’t stand up to
    > companies, patents don’t stand up to
    > companies, why would I think a license
    > that is untested in the court system
    > could stand up to a company?

    Take a look at those links

    http://news.com.com/2100-7344-5198117.html
    http://www.groklaw.net/article.php?story=2004072315554313

    and then tell me whether the GPL is not standing up to companies and to companies’ misappropriation of GPL’ed code.

    Kind regards,

    Rafael
    Editor in Chief
    Linux Magazine Brasil

  • Anonymous

    Exactly what people in businesses actually DO…they REMOVE the GPL code out of their software so they are not forced to GPL the rest of it.

    http://www.iht.com/articles/2004/12/28/business/code.html

  • Anonymous

    yep…they’d be locked into an open platform…which means they’d be locked in..thanks for playing drive through

  • fmcgowan

    Have you ever seen “Support Your Local Sheriff”? The town ran out of money to build the jail _before_ they had _any_ of the bars. The “prisoners” were told to consider themselves locked-in. The fact that they did is a main source of the humor in the movie.

    That is _exactly_ the amount of lockin you get from _open_ API and ABI sets; you can come and go as you please because _there is nothing to stop you_. If you _are_ locked into an _open_ implementation, it’s because _you see the bars_ regardless of the fact _they don’t exist_.

    MS gets lockin from their API sets precisely because the source is closed; you can call any given function (presuming they bothered to document it at all) but you cannot see how it transforms its input(s) into its output(s) or any of the intermediate product(s). Depending on this function working its magic locks you in because you cannot replace the function.

    I don’t know how to say it more clearly: API/ABI sets provide lockin precisely to the extent they are closed. Therefor, completely open API/ABI sets (the LSB sets are completely open GPL-ware) _have no lockin potential_.

  • http://Thisone devnet

    Not really. An API/ABI CAN be used to lock in anyone. They lock you into a certain way of programming. _This can be good_…but it can also be bad when used wrong. No matter what you say, you cannot refute THAT fact.

    MS API’s are open…not closed. The functions they call are built into the operating system. The API is a framework to call those functions. We’re not speaking of the closed source functions…we’re talking about the open calls (APIs).

    That’s why everyone can program an installer or program for MS windows using Win32 or WinFX or whatever API the OS requires. They open them so third party programmers can have programs that CONFORM to STANDARDS. This is what API’s do. It is in their nature to lock down options aka simplify programming that the programmer has so they have a standard to program around with as little work as possible. While this is good for development in some instances…if left in the hands of money hungry idiots…it can be a bad thing.

  • Quinn

    Yay! We won one!

    Now if we could only win somewhere that would set a precedent. You have to consider that the source of this win was Germany. Afterall, the German court system gave us (used collectively here) the German Multimedia Law that says ISP’s are liable for what their users have on their computers (that’s laughable).

  • fmcgowan

    _Nonsense!_ It can only _really_ lock you in if it is a black box, the inner workings of which are hidden from you. The Windows API set are that sort of thing: you can see the name, parameter list , return type and error indicators but not the logic that connects the parameters to the return value. The way the API/ABI set is constructed will _encourage_ certain types of programming behavior and _discourage_ others. If the interior logic is available for examination as it is in LSB implementations, you are still not locked in because you can port the code to another platform and provide any missing functionality your applicaiton requires. The hidden nature of the Win API set makes that rather difficult, judging by relative scarcity of apps ported from Windows to other platforms.

    On the other hand, if you fall into certain habitual practices that are not forced upon you even if those practices are encouraged by the API/ABI set, you locked _youself_ into those practices, the API/ABI set did not.

    Yes, the _interface_ specifications are open – some of them. I agree with you that the whole point of an API is to provide a means to call the underlying functions. But, yes, we _are_ speaking of those underlying functions when we speak of “lockin” precisely because that is the layer that imposes lockin.

    And,… yes, existence of documented API calls _is_ why 3rd party software can be written for Windows, any other OS and several applications designed to accept plug-ins or extensions, but the closed nature of the underlying functionality is _why_ Windows lockin exists. It is the same reason those Mozilla plug-ins won’t work with IE, etc. The problem with MS “standards” is that they are of the “de facto” variety rather than the “specified” variety; _MS gets to make any changes to these “standards” they want at any time they choose for any reason or no reason at all_. That is why VB programmers, for instance, have to keep up with the released version of the language.

    The point of LSB is to provide multiple programmers and multiple vendors with a _specified standard_ application framework for Linux so programs written to that standard will be binary compatible across distributions. Then end users can stop trying to figure out how to use MAKE.

    That’s the point. You can _still_ ignore the LSB specs when you write your application and still make it work on Linux. You just won’t be able to offer the same level of assurance that any non-technical user can install and run your app without using MAKE (or some analog).

    No lockin; as the developer, _you have the choice_ to make your app LSB compliant or not. Should you choose _not_, users will still have the choice to MAKE your app from the source code or not, assuming you have released an OSS application; if you released a _closed_ application, people who can’t get the binaries to work are just out of luck.

    As the developer, you also have the choice to port your app to another platform because you have access to information detailing _exactly_ what underlying functionality the OS provides to your app _because the specs and the code are open_.

    Windows denies you that information because it is _closed_. Linux grants you that information because it is _open_; LSB does not matter in this context.

    That’s why LSB will _not_ create any lockin.

  • Pingback: Self Musings | Yet Another Linux Blog()