Women in Technology

Hear us Roar



Article:
  The Strange Case of the Disappearing Open Source Vendors
Subject:   Essay sidesteps the demonstrable corrosive effects of some licenses
Date:   2002-06-29 13:09:07
From:   timoreilly
Response to: Essay sidesteps the demonstrable corrosive effects of some licenses

Brett, I do share your preference for BSD-style licenses, and if GPL-oriented developers are trying to take advantage of their users with a contract that they don't fully understand, I would share your concern. (I have seen this in some cases, and in fact cited one in a past article on this subject, where the developer was exulting that his bosses didn't understand what he was doing by GPLing the code.) But I believe that most GPL software developers are in fact quite ethical, even idealistic, and most users understand the consequences fairly well. The GPL is a reasonable choice for many situations.


I'll also point that all those millions of PCs you cite as dwarfing software built for use rely quite heavily on that very software. As I've been pointing out ad nauseam since 1998, any one of them that uses the internet relies at least on Bind, and likely on Sendmail, Apache, and Perl as well, for much of its work. The Microsoft TCP/IP stack was adapted from the Berkeley stack, and much of this software, developed in and shared by the commons, has made its way into proprietary commercial software.


As to your claim that many software startups have failed as a result of the GPL, you must balance that with the numbers of proprietary software startups that have also failed. It's a case of Sturgeon's law, not of some unique failing of the GPL. The history of the high tech industry is littered with failures; they are the soil from which new startups grow.

Full Threads Oldest First

Showing messages 1 through 9 of 9.

  • Comments sidesteps the point of this article???
    2002-06-30 19:43:59  mintslice [View]

    Maybe I missed something when reading this article, but the point of this article seemed to be that while the GPL might not be good for the businesses that sell software (and particularly their bottom lines), that it's great for businesses that want solutions to their problems.

    I'm sure that there are many programmers that will bemoan the effect of open source and the GPL on their profit lines, but the reality is that the end users - businesses, home users, small companies, etc - don't want software that does something, they want software that let's them do what they need to do.

    The reality is that proprietary approaches limit the end users ability to get on with the job, and open source enable's the end user.

    The GPL has two things in it's favor:

    1. Companies can create solutions without having to code everything, while keeping costs to a minimum. As Tim said, most software developed is used in-house to address the needs of the company. It's not sold, there for the GPL allows companies to share from a rich pool of software. At the end of the day, it doesn't matter how software is developed, as long as it meets the needs of the company.

    2. Coders feel comfortable with the GPL, because the GPL doesn't leave them feeling used and abused. Imagine spending you spare time developing software that anyone can use only to find that someone has used your code, failed to acknowlege it, and returned nothing.

    For example, while Microsoft has quietly used the TCP/IP stack code from BSD to solve a problem they couldn't address themselves, they have not only snuck the code in without so much as a thankyou to the developers, but they've publicly ridiculed the source of the code, saying open source is bad. If I'd developed that code, I think I might have a sour taste in my mouth.

    The GPL is about getting things done. It's about solving problems. It's also up front. There's not hidden agenda. You can read the license and choose for yourself.

    I appreciate that some coders may be scared by the implications of the GPL, and it's affect on their ability to make money - Microsoft is certainly trembling in it's pants over consequences of free software that could replace it's cash cows - but with 95% of programmers producing software that meets a specific need with no intent to sell that code, I think concern over it's consequences is a little over stated.
    • Not at all.
      2002-07-06 13:42:51  brettglass [View]

      You write:

      > The GPL is about getting things done. It's about solving problems. It's also up front. There's not hidden agenda. You can read the license and choose for yourself.

      This is absolutely incorrect. The GPL is anything BUT "upfront." In fact, its preamble is intentionally deceptive. It states that the purpose of the license is to promote "freedom," when in fact its purpose is to deny freedom -- and, by doing so, destroy programmers of commercial software and their businesses.
      • Not at all.
        2002-07-17 21:31:56  john_betelgeuse [View]

        > This is absolutely incorrect. The GPL is
        > anything BUT "upfront."

        Wrong. The GPL is remarkably easy to read,
        and quite clear, for a legal document.

        > In fact, its preamble is intentionally
        > deceptive.

        No, it's not.

        You may believe what you want, but the preamble
        is not deceptive.

        Surprisingly large numbers of people, however,
        do not understand that the word "free" has
        more than one definition. You may be one
        of them.

        > It states that the purpose of the license
        > is to promote "freedom,"

        And, it does so admirably. Due to the GPL,
        my freedoms are very well protected.

        > when in fact its purpose is to deny
        > freedom --

        A bald, inaccurate and totally unsupported
        assertion.

        The purpose of the GPL is indeed to protect
        freedom, your belief otherwise not with standing.

        > and, by doing so, destroy programmers
        > of commercial software and their businesses.

        Wrong again. The GPL is not designed to
        destroy commercial software and the software
        business.

        It MAY eventually change the software business
        model . . . but it will not destroy it.

        A mechanic gets paid per hour of work. There
        is absolutely no reason why software producers
        cannot be paid the same way. And as such, there
        is absolutely no reason why that software cannot
        be licensed under the GPL, since after you get
        paid, you're done with it.

        The new business model would be, however, rather
        different from the existing model, where a
        software producer can work for a fixed amount
        of time, then coast for years afterwards.

        While not changing the cost and functionality
        of software much at all, a model where software
        producers get paid for their WORK, not because
        they own a piece of software, would make the
        competitive playing field much fairer, and
        much less fragmented by proprietary file
        formats and networking protocols.

        Or would you prefer a software-like model
        in your automotive braking systems, where
        everytime you hit the pedal, you gotta pay
        your mechanic again?
  • Essay sidesteps the demonstrable corrosive effects of some licenses
    2002-06-29 14:12:03  brettglass [View]

    Tim writes:

    > Brett, I do share your preference for BSD-style
    > licenses, and if GPL-oriented developers are
    > trying to take advantage of their users with
    > a contract that they don't fully understand,
    > I would share your concern.

    Most often, alas, the developers who stamp the GPL on their code do so for one of two reasons:

    (a) They do not understand the GPL's true intent and history, and rely instead on the false and deceptive assertions in its "preamble. They therefore hurt themselves, their profession, and the world at large while mistakenly believing that they have done something good. And of course, once the cat's out of the bag, it's out; they can't reconsider if they later learn the truth.

    (b)They started with GPLed code to begin with, and so do not have a choice of licenses. This runs counter to your idea, stated in your essay, that they should be able to license the code as they choose. The GPL takes away what you call "Freedom Zero" by coercing developers who make changes to the code to use this coercive and destructive license, and the cycle repeats. Hence the term "viral."

    > (I have seen this in some cases, and in fact
    > cited one in a past article on this subject,
    > where the developer was exulting that his
    > bosses didn't understand what he was doing by
    > GPLing the code.)

    Did you know that Richard Stallman has actively advocated this sort of deception? In his essay, "Why Software Should Not Have Owners," Stallman encouraged programmers to pull this trick on their employers. After I pointed this out, he edited the essay (in Orwellian fashion) so that it no longer contained a direct exhortation, but the implication is still clear: Stallman wanted programmers to use the GPL as a "monkey wrench" within the workings of their organizations.

    > But I believe that most GPL software
    > developers are in fact quite ethical,
    > even idealistic,

    Like some who are driven by idealism (e.g.
    religious fanaticism) to do horrible things,
    they believe themselves to be doing
    something ethical but are not. They have
    been deceived.

    > and most users understand the consequences
    > fairly well.

    I disagree. How many of them realize that they
    are, in fact, perpetuating Microsoft's
    monopoly and harming their colleagues? Would
    they still support the GPL if they understood
    that? I believe not.

    > The GPL is a reasonable choice for many
    > situations.

    I respectfully, but firmly, disagree. A license which is discriminatory, which is intended to destroy people's livelihoods, and which snuffs out competition is not a reasonable choice.

    > I'll also point that all those millions of
    > PCs you cite as dwarfing software built for
    > use

    I assume you mean "in-house use" here...

    > rely quite heavily on that very software.

    Actually, they do not. They rely not on GPLed software but on BSD-licensed software. There's BSD-licensed code on every one of those
    machines, but in most cases no GPLed code. Which underscores the point that open source code licensed under truly free licenses is of great benefit to society, while GPLed code ultimately does more harm than good.

    > As I've been pointing out ad nauseam since
    > 1998, any one of them that uses the
    > internet relies at least on Bind,

    Which is BSD-licensed.

    > and likely on Sendmail,

    Which used to be BSD-licensed, although it recently (alas!) went to a vendor-specific "poison pill" license whose purpose is to prevent anyone but Sendmail, Inc. from making money by enhancing the code.

    > Apache,

    The Apache license is truly free. It is almost identical to the BSD license, with the exception of a few provisions to protect the Apache trademark.

    > and Perl as well,

    Licensed under the Artistic License, which is a truly free license. (It does not restrict what you can do with the code -- only what you may do to the author. In particular, it says that you can't steal the author's thunder.) Larry Wall later threw the GPL zealots a bone by dual-licensing under the GPL, which I believe to be a mistake. Guido Van Rossum has stuck to a BSD-like license for Python, which is a more appropriate course of action.

    > for much of its work. The Microsoft TCP/IP
    > stack was adapted from the Berkeley stack,

    Which is a good thing. Had the Berkeley stack
    not been BSD licensed, Microsoft would have
    tried to force a proprietary protocol upon the
    world.

    > and much of this software, developed in
    > and shared by the commons, has made its way
    > into proprietary commercial software.

    Only if it is not GPLed! Which is, again, my
    point. Open source which is licensed under
    a truly free license is highly beneficial to
    society. It promotes competition, sharing of
    knowledge, and code reuse. The GPL does not, because its intent and effect is to hurt businesses -- especially small start-ups. (As I pointed out above, Stallman wrote the GPL in a fit of rage against startups that were formed by people with whom he once worked.... He attacked them in anger because he blamed them for destroying the "Nirvana" of the MIT AI Lab. See Steve Levy's book "Hackers" for a very well-written account.)

    > As to your claim that many software startups
    > have failed as a result of the GPL, you
    > must balance that with the numbers of
    > proprietary

    The correct word is "commercial." The word
    "proprietary" has a different meaning. It
    implies incompatibility and secrecy, not
    whether or not the code is sold for money.

    > software startups that have also failed.

    Many of the commercial software companies that have failed have done so, at least in part, due to the GPL. Be, Inc. is a good example. Wedged between Microsoft on the one hand and Linux on another, it could not gain market share.

    • Essay sidesteps the demonstrable corrosive effects of some licenses
      2002-06-30 13:17:28  yodaiken [View]



      > and likely on Sendmail,

      =Which used to be BSD-licensed, although it =recently (alas!) went to a vendor-specific ="poison pill" license whose purpose is to =prevent anyone but Sendmail, Inc. from making =money by enhancing the code.


      So what's your problem with this? Sendmail makes its code available for use and modification, but not for someone else to relabel and resell. Is this wrong either ethically or as a business proposition?

      Why should Sendmail develop software so that the copier with the best marketing can win?


      >Many of the commercial software companies that >have failed have done so, at least in part, due >to the GPL. Be, Inc. is a good example. Wedged >between Microsoft on the one hand and Linux on >another, it could not gain market share.

      Be failed because their technology was weak and their business model was to try to get Apple to buy them - and Apple saw a better opportunity.
    • Essay sidesteps the demonstrable corrosive effects of some licenses
      2002-06-29 21:40:25  demoryw [View]

      You have made a number of unsubstantiated assertions, in particular your statements regarding Stallman - but I shall not bother retorting those. At best they are foolish propaganda... (Note that I do think you made *some* interesting points).

      Regarding the following, I have a few questions:

      > Open source which is licensed under
      a truly free license is highly beneficial to
      society. It promotes competition, sharing of
      knowledge, and code reuse. The GPL does not, because its intent and effect is to hurt businesses -- especially small start-ups.

      Is is beneficial to society when Microsoft receives a 900% markup on it's products (it is well known that the incremental cost of selling software is negligible, whereas the profits are not)?

      What is to stop companies from destroying the public domain by not releasing any code, ever? Would you be happier, as I might, if the law required that all source code be released, under a BSD style license, after 5 years?
      • Essay sidesteps the demonstrable corrosive effects of some licenses
        2002-06-29 22:16:51  brettglass [View]

        You write:

        >Is [it] beneficial to society when Microsoft receives a 900% markup on it's products (it is well known that the incremental cost of selling software is negligible, whereas the profits are not)?

        If you were King, exactly what markup would you mandate?

        Clearly, it's not appropriate to dictate a number. The market should reward developers according to what it thinks products are work. If the market is open, free, and fair, and there is fair competition, then developers will get what they deserve.

        The GPL, however, is an attempt to drive the amount that the public is willing to pay to zero, destroying the market and preventing reasonable compensation of developers. Stallman says as much in "The GNU Manifesto," where he decrees that good pay for programmers should be "banned."

        > What is to stop companies from destroying the public domain by not releasing any code, ever?

        Here's an analogous question: What's to keep me from destroying public parks by not donating the land on which my house sits to be used as one?

        I hope that this shows the absurdity of your question. If a company does not release code, it is not "destroying" the public domain. It's merely not contributing to it. And that's OK. Not everyone can afford to give his or her property away.

        > Would you be happier, as I might, if the law required that all source code be released, under a BSD style license, after 5 years?

        Not at all. That would be an unconstitutional "taking." All releases of source code by an author should be voluntary, not compulsory. Many can and will do so, as I do.
        • So called: corrosive effects are not demonstratable.
          2002-07-17 21:39:47  john_betelgeuse [View]

          > If you were King, exactly what markup
          > would you mandate?

          If I were King, I wouldn't mandate markups . . .
          I would mandate that all interfaces, file formats
          and network protocols would be required to be
          released as free and royalty free standards to
          the public.

          A rather nice compromise, I think, in that it
          wouldn't force a company to give away it's
          IP, but it would eliminate the problem of
          "vendor tie-in" by giving customers the ability
          to mix and match components aquired from
          different companies and sources.

          > The market should reward developers
          > according to what it thinks products
          > are work.

          Agreed. However, that is not how the market
          currently works. The market is forced to
          continue buying and using products due to the
          fact that the vendor has "locked them in" to
          its product by encrypting their data in
          secret, proprietary file formats, by locking
          them into secret, proprietary protocols, and
          by unfairly hobbling competition by providing
          information about only an inferior sub-set
          of the interfaces available on thier components.

          > Here's an analogous question: What's to
          > keep me from destroying public parks
          > by not donating the land on which my
          > house sits to be used as one?

          Nothing. So the government may be required
          to exercise eminent domain, pay you, and
          kick you out.

          What was your point? And how does this relate
          to the topic at hand.
        • Essay sidesteps the demonstrable corrosive effects of some licenses
          2002-06-29 22:47:55  demoryw [View]

          > If you were King, exactly what markup would you mandate?

          Clearly that is not the question. The fact that MS (and others) can charge an exorbitant markup proves that competition is not occurring.

          > Clearly, it's not appropriate to dictate a number.

          Agreed.

          > What's to keep me from destroying public parks by not donating the land on which my house sits to be used as one?

          Please read Adam Smith's "On the Wealth of Nations". He has a whole chapter on the inequity of land ownership.

          > I hope that this shows the absurdity of your question.

          Sorry. It doesn't. :)

          > If a company does not release code, it is not "destroying" the public domain. It's merely not contributing to it

          > And that's OK. Not everyone can afford to give his or her property away.

          Can MS? Sure they can. Why don't they? Because they profit not by working, but by owning, by hording standards (e.g. .doc), by being a monopoly, and by generally hurting the industry and consumers.

          > Not at all. That would be an unconstitutional "taking."

          See the first amendment. "Talking" is not unconstitutional.

          > All releases of source code by an author should be voluntary, not compulsory. Many can and will do so, as I do.

          If the author has any rights at all it because society gives them to him. For example, why are patents and software licenses enforceable (ignoring for the moment the corruption of the federal government by cooperations)? Because we say so.

          But perhaps you're right and we should not make such requirements. Tell me then how do we do it? BSD style licenses will not protect the consumer. Period.