[I owe an apology to Jim and the politechnicals who replied for my delay in compiling these responses. Let me just say in my defense that there were a *lot* of them. At about 120KB, this may be the longest message I've sent to Politech in the eight years since I founded the list. Also, I should say that although Jim writes quite a bit about economics nowadays, he is also a Harvard Law School-educated lawyer. Previous message is here: http://www.politechbot.com/p-04007.html --Declan] --- Date: Thu, 19 Sep 2002 16:23:01 -0700 From: Jules Agee <julesaat_private> To: declanat_private Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use Declan, I'm sure this has already generated a flood of responses, but I would like to make a couple of comments in response to Jim Delong's statements: > If IT companies, universities, and IBM want to donate the fruits of > their labor to computer purchasers, including governments, that is > their privilege. But we have just gone through a half a decade in > which the business model was give it away, and it did not work. In the > end, software might be bundled with hardware, or vendors might give > away software tied to a services contract both are increasingly common > -- but the code writers will want pay for producing it, which means > money must ultimately come from the users somehow. When all you've got is a hammer, everything looks like a nail. Jim's experience as an economist seems to be distorting his perception of Linux. Linux isn't a company, it's not a business model, and it doesn't need to be profitable to be successful. Linux has been around a lot longer than there have been corporations pouring money into developing it. Linux is currently being incorporated as essential infrastructure in businesses across the country. If every corporation and university currently developing Linux were to disappear tomorrow, the open nature of Linux means that anyone using it can hire programmers to modify it to meet their business needs. The success or failure of any business developing Linux will only slow or hasten the increasing functionality and stability of Linux. Linux itself isn't going away anytime in the forseeable future. Jim wrote, "The Linux community is moving toward proprietary aps, but it is chancy. Writing aps without incorporating some operating system code is difficult, and those who want to engraft proprietary aps onto Linux are taking a legal risk." This is the biggest canard put forth by the anti-open-source spin doctors. The whole point of developing an Open Source platform is to let application developers get away from the absurd control games played by proprietary platform developers. Open Source developers want their platform to be useful; they have no other incentive for their work. Almost every library is wide open to be freely used and linked to from proprietary applications. The "legal risk" that Mr. Delong dramatically hints at is a myth invented by vendors of proprietary platforms who are nervously watching Linux make incredible advances in acceptance among technical professionals. -- Jules Agee System Administrator --- Date: Sat, 21 Sep 2002 13:41:10 -0500 From: Steve Langasek <vorlonat_private> To: Declan McCullagh <declanat_private> Cc: rmorrisonat_private, jdelongat_private Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use Hi Declan, As a long-time supporter of the Free Software movement and an armchair advisor to the Debian project on software licensing, I feel a response to Mr. Delong's message is needed. He is certainly correct that the GPL requires anyone using GPLed code to place their own code under the GPL; and he is also correct that "Open Source believers" think this is ok -- after all, just think what requirements would be imposed by a comparable license to not just view, but to use and modify Microsoft's source code! Where he misses the mark is when he suggests that the GPL poses a legal risk for application developers. He writes: > Writing aps without incorporating some operating system code is > difficult, and those who want to engraft proprietary aps onto Linux are > taking a legal risk. As one who has far more experience than the average economist when it comes to writing application software for Linux, I can assure you this is not the case. Although most components of Linux-based operating systems are Free Software, few software libraries (the OS components normally incorporated by application writers) are licensed under the GPL; they are typically covered by other licenses, such as the Lesser General Public License (LGPL), which do not pose this sort of legal entanglement for application writers. There are certainly some in the Free Software community that believe the GPL should be used as a wedge to free all the world's software; but most Free Software authors take a more pragmatic stance about coexisting with proprietary software, and the licensing of many key libraries reflects this. It should perhaps also be noted that those libraries which ARE licensed under the GPL are not usually Linux-specific. Many of them have been ported to other platforms, including the (non-GPL) BSD Unixes, and sometimes even Windows. Of course, any software firm that's planning to incorporate Free source code into their applications would be well-advised to consult a lawyer regarding these issues, rather than basing their decision on the word of either an economist or a programmer. Regards, Steve Langasek Debian GNU/Linux Developer --- Date: Sun, 22 Sep 2002 15:59:04 -0400 (EDT) From: mwillemsat_private To: jdelongat_private cc: rmorrisonat_private, <declanat_private> Subject: Linux etc Jim, As the CTO of a small international company (80 people, in 4 countries (UK, US, Canada, Hong Kong), may I say a few words? I may have misread, but your piece came across as somewhat anti-Linux. Since I am agnostic on the subject, and we use a mixture of Windows and Linux in our company, maybe my real life experiences (rather than religious feelings, as I fear most of the reactions will be) will be helpful. First of all the technology. I have no hesitation in saying that Linux is better than Windows at our corporate server tasks. Now. Today. "Free" has nothing to do with this. Better means years of uptime, easy changes without reboots (this is a BIG issue in a dynamic environment such as ours, where uptime is God as well!), more standards compliance, and so on. Also, service packs have been much less reliable than Linux fixes. I cannot talk for others, but for my corporation, Linux, at least on the server, has been a godsend. We switched most of our servers to Linux, and thanks to Linux and other OSS technologies such as MySQL and PHP, we have seen a real increase in technical reliability and uptime. Yeah, the "this is free" feeling helps, but as you say, nothing is really free, so this is peripheral at best. It helps me defend our plans to the board, of course ("you saved HOW much on licenses?"). :) Second, a much 'softer' point: control. Consider me funny, but I feel that having our core (we are a tech company and Internet technology is core to what we do) controlled by another company telling us what we can and cannot do (and this driven by its own interests and by marketing) is an uncomfortable feeling. If MS was open, we would not have any trouble: we would then have the option of going elsewhere. Now, we do not have that option with our MS products. Reason to at least look at alternatives (as indeed we did). Then the licensing: we write propriatery software on Linux. I do not see the problem. You can write proprietary software in Windows, without being affected by the Windows license, and similarly, you can do the same on Linux. The fact that the OS our app runs on is Linux, and the language is PHP, say, and the database underlying it all is MySQL does not in any way prohibit us from making OUR app proprietary. We are paid rather a lot of money by our customers, so it obviously works in practice. I do not understand the "viral" objection: why would anyone include GPL'ed software in their app? It runs on TOP of GPL'ed software, which is fine. Unless of course I were to re-issue a proprietary version of Pine, say, or Mozilla. And even then: at least with Linux you CAN do it, as long as you add source - try releasing a proprietary Internet Explorer! Then there is Support. MS has been so bad we do not ever try anymore: we know more than the people on the phone do. Wait on hold - go through procedures - pay $$$ - and then be told things you already know. Support has been atrocious. 10 years ago, MS support was excellent: now it is merely a moneygrab. My personal opinion. Linux support, run by volunteers, gets me replies from subsystems' authors within hours. Try that with "big corporate" software. I wax lyrical, but I am really agnostic - whatever works, works. Of course for desktops, Windows is invaluable (although personally I now use StarOffice, and am very happy with it: maybe I can try Linux on the corporate desktop soon. Who knows.) Anyway, I can see the merits of both - am merely observing that the Linux/OSS model that you say "does not work", in fact DOES work, and very well too. Michael --- Michael Willems BSc(Eng) Chief Technology Officer Digital View Ltd London/Hong Kong San Jose/Toronto --- Date: Mon, 23 Sep 2002 04:25:10 -0500 From: "Karl O . Pinc" <kopat_private> Subject: Linux editorials: What they don't get To: Declan McCullagh <declanat_private>, rmorrisonat_private, jdelongat_private, "Newsforge . com Managing Editor Grant Gross" <grantat_private> Hi, Any or all of you are welcome to publish or otherwise re-distribute the following article. I ask, but do not require, attribution, the chance to discuss editorial modifications or deletitions, and notification of publication. Web links are encoded in HTML. If you would prefer plain text I'd be happy to convert them to parenthesized addresses, or, you have my permission to do so yourselves. Regards, Karl O. Pinc <kopat_private> ----------------------------<snip>---------------------------- Making Cents with Linux? GNU/Linux has a habit of turning things on their head, with the result that those who aren't paying close attention see the world up-side down. The <a href="http://www.cei.org/gencon/016,03211.cfm">September 19, 2002 c:\spin</a> by James V. DeLong, Senior Fellow, Project on Technology and Innovation, <a href="http://www.cei.org">Competitive Enterprise Institute</a>, is a case in point. Of Linux, Mr. DeLong writes "But the movement is not the folk song army depicted in the NYT. ...money must ultimately come from the users". He is absolutely right, yet he has no idea how it is that the money _has_ come from the users. He seems to think that Linux is "supported, often handsomely, by universities and IT companies" as well as "hardware companies, notably IBM, [who] are now pouring billions into it." The money does come from the users, generally in the form of direct contribution of labor. Corporations often have a programming staff whom they pay to deliver the software the corporation needs. If the corporation uses Linux, the well run corporation will, even through it is not required to, contribute back to the community any improvements made to Linux. This relieves the corporation of <a href="http://www.cs.rug.nl/~bosch/papers/SAAModifiability.pdf">50% to 70% of the modification's total cost</a>, the burden of ongoing maintenance. If the corporation relies on a Linux program, it may be in the corporation's interest to pay a programmer to manage the development of the program -- deciding which improvements to incorporate into the program and how they should be integrated. What's true for the corporation that uses a Linux program is just as true for an individual programmer who uses a Linux program, and many smaller programs are maintained by individuals. Those who are not programmers always have the option of paying a programmer, any programmer, to make any Free Software more useful. This is a decentralized and highly efficient means of producing programs. There is no marketing department, top management, or sales department to suck up resources. No corporate books, no accountants, and no taxes. (Although of course taxes are paid on the programmer's salaries and on the Linux users' profits.) Corporations who use Linux have these overheads, but they have them regardless of whether they use Linux or not. Further, anyone who thinks they can do a better job of maintaining a program is welcome to try. If the newcomer really does produce a better program, it will replace the old program or relegate it to the backwaters. A reality check is in order. It's been estimated (<a href="http://www.dwheeler.com/sloc/redhat71-v1/redhat71sloc.html">Wheeler</a>) that it would cost over 1 billion dollars to produce a representative Linux distribution by conventional proprietary means. Clearly, this money did not come from the goodness in the hearts of "universities and IT companies". Nor did it come from IBM. To the general Linux community, at least in the U.S., the most notable result of IBM's billion dollar expenditure was the vandalism of the public sidewalks with spay painted penguins. (Did I mention Linux has a habit of turning things on their head?) Certainly Linux ran on IBM's mainframe architecture _before_ IBM made it's investment. IBM has found it advantageous to spend money to give Linux what it lacked, publicity, white papers, sales, and marketing, as well as, no doubt, improving the compatibility between Linux and IBM hardware. No, the money it took to develop Linux, likely less than 1 billion dollars due to the efficiency of the development process, came from throughout the entire Linux user base. A little labor at a time. Upside down as usual: from each according to his need. In his discussion regarding the creation of applications for Linux, Mr. DeLong is spot-on in his summary of the <a href="http://www.gnu.org/licenses/licenses.html#TOCGPL">GNU General Public License</a>, right up to his concluding sentence: "Writing aps [sic] without incorporating some operating system code is difficult, and those who want to engraft proprietary aps [sic] onto Linux are taking a legal risk." A moments thought is enough to see the fallacy. Microsoft's license for it's Windows products are _much_ more restrictive than the GNU General Public License, and _much_ more "viral" in that the mingling of any of Microsoft's source code with yours surely taints your ownership of the resulting program. Use Microsoft's code in your program and, well, you're not running a risk, you're sure to have lawyers in your future. Yet, there are (ahem) many proprietary applications sold (_without_ Microsoft's permission, mind) to run on Microsoft's operating systems. Mr. DeLong is just plain wrong. Writing apps for Linux is no riskier than writing apps for any other operating system. You can expect to see proprietary Linux applications when the software manufacturers see a large enough deployed base of Linux systems. Indeed, this is already happening in the Linux server market. Mr. DeLong writes "governments should not treat this as an arena for industrial policy." I agree that government should not mandate solutions to industry. Government must, however glacially, respond to a changing world and alter it's own operations to best act in the public good. The <a href="http://www.sincerechoice.org">Sincere Choice Initiative</a> has laid out a good set of policies which would benefit the populace should governments follow them. Included in these policies are the use of both Free and proprietary software, competition by merit, and support for interoperability and open standards. A notable aspect of the Sincere Choice policies is the rejection of a practice current among proprietary software manufacturers -- proprietary encoding of the users data in formats which deny the user any useful means of accessing his data other than through the vendor's products. Anyone who has imagined how much money their office might save if they didn't have to upgrade all their word processing software can envision how much a government might save if the government didn't have to upgrade either. The future of Linux is questioned when Mr. DeLong writes: "The incentives fueling the Linux movement are not necessarily those required for long-term production of software suited for the public...." A better probe of the future turns the question on it's head and focuses on the one rather than the many, examines the present rather than the future. The pertinent question is: "Can Linux meet my needs today?" If it can then it's future is assured, because, unlike proprietary software, Free Software is never taken off the market. Neither economic failure or marketplace manipulation makes Free Software unavailable. As anyone can pay for improvements, the worst that can happen is the absence of free improvements. Free Software does not go away, it's only replaced with something better. If today's Linux meets your needs, the future of Linux is sure to look mighty fine. Karl O. Pinc <kopat_private> --- Date: Sun, 22 Sep 2002 15:41:16 +0200 To: declanat_private From: Massimo Mauro <Massimo.Mauroat_private> Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use Declan, I like very much your list, which is a precious source of information on matters that are important to me, which you select very carefully. Why you chose then to publish this silly nonsense truly escapes me. No, I am not at all a fanatic Linux lover, but as an IT professional [and not an economist] I can safely state that DeLong's comments: - provide no new objective element whatsoever; - keep drumming up axioms that are in fact entirely US-American, like "there must be profit in everything, else it ain't gonna work"; there is rather a big world beyond the US of A, in case the hapless writer had not noticed yet; the "business model" elsewhere may differ from the prevailing US-WASP model and still be feasible, though not necessarily in GW and Bill's country [my heart bleeds...]; - the statement "Microsoft(...) is developing shared source to open up code to scrutiny while the company keeps firm hold of the pen" may qualify for "Joke of the Year", if one just bothers to read the letter MS wrote to the Peruvian Representative trying to promote Free Software in Peru, Rep. Villanueva; "shared source"? What on earth is that supposed to mean? I enjoy the info you supply, but unfortunately this does border on BSA/MS infomercials... Massimo (in Brussels, Belgium, slightly north of France...) PS To prove DeLong narrow-sightedness, one may also read his hysterical peanas on IPR, in the message previously published on Politech : "Protection of intellectual property, like protection of physical property, has vibrant moral, economic, and political justifications." I don't vibrate, nor resonate with this generic inflammatory and semantically shallow statement... IPR should be protected, but whether IPR violations justify drastic retaliatory action, verging on legalised hacking, as in the legislation now considered by Congress, is open to debate. --- Date: Thu, 19 Sep 2002 18:57:54 -0400 (EDT) From: proclus@gnu-darwin.org Reply-To: proclus realm <proclus@gnu-darwin.org> Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use To: declanat_private Cc: politechat_private, rmorrisonat_private, jdelongat_private Delong must go back to the the drawing board and study the "freeware spirit" with more care. Part of the freedom of the GPL is commercial freedom, which means that developers are free to sell the software for money. Delong mistakenly thinks that the opposite is true, and as a result, most of his analysis breaks down into a "good old greed" rant which is based on a fallacy. There are plenty of companies, which are making a commercial play with free software; RedHat, Apple, Sun, IBM, etc. It remains to be seen whether the world of proprietary software can stand up to such competition. Regards, proclus http://www.gnu-darwin.org/ --- Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use Date: Thu, 19 Sep 2002 15:53:32 -0700 From: Paul Schreiber <shrubat_private> cc: <rmorrisonat_private>, <jdelongat_private>, "Declan McCullagh" <declanat_private> This article contains numerous factual errors. I hope this is intendad as an op-ed piece and not a news article. >A second problem is the creation of applications for Linux. The General >Public License that controls the program s distribution can be paraphrased >as thou shalt not charge for this program and its source code shall be >public. This license is also viral; if you write an ap for Linux, and >incorporate any code covered by the GPL, then your ap is also subject to >the GPL, and it too becomes open source and free. You're wrong here on both counts: (1) You can charge money for applications that use GPL'd code. (2) You imply all Linux applications use GPL'd code. This is not the case. >The Linux >community is moving toward proprietary aps, but it is chancy. Writing aps >without incorporating some operating system code is difficult, and those >who want to engraft proprietary aps onto Linux are taking a legal risk. The statement "writing aps[sic] without incorporating some operating system code is difficult" is simply incorrect. This indicates to me that (a) you're not a software developer and (b) you're not familiar with the LGPL. >Finally, governments should not treat this as an arena for industrial >policy. The incentives fueling the Linux movement are not necessarily >those required for long-term production of software suited for the public >as well as the nerds. Governments, which are as naïve as editorial >writers, should keep their hands off. An aside: this paragraph appears to have no point. First, you made a vague statement about incentives without enumerating what those incentives are or how they conflict (or at least don't match up). Second, you've launched an ad hominem attack on the _Times_' editorial staff. Paul shad 96c / uw cs 2001 / mac activist / fumbler / eda / headliner / navy-souper fan of / sophie b. / steve poltz / habs / bills / 49ers / --- Date: Thu, 19 Sep 2002 15:39:35 -0700 (PDT) From: Abraham Ingersoll <abeat_private> To: Richard Morrison <rmorrisonat_private> cc: Declan McCullagh <declanat_private>, <jdelongat_private> Subject: Re: CEI's Weekly Commentary: Software Wars Hello Mr. Morrison, I'm a twenty-two year old software developer who makes his living using, supporting and developing systems that are often - but not always - based on Linux. I expect that folks carrying much more weight than myself will see your editorial and reply with a more in-depth critique. But even so, I feel compelled to respond. On your first point, that "linux isn't free", I have an as-simple retort: Linux (and open source, free software in general) carries the best value proposition of any modern body of software, especially when compared to the closed wares offered by the likes of predatory software companies such as Microsoft. Perhaps you're not tuned into this fact, but slowly and surely more and more folks are becoming 'clued in' to free software and discovering massive benefits. Standing against this free software tide is akin to the Kansas school board's 1999 decision to have the state's science teachers remove evolution from their curriculums. Your second point, on the GPL and "application creation for Linux," is also critically flawed. Application developers who create Linux applications can distribute their wares under any license they please. Companies such as Corel (WordPerfect), ActiveState (Kodomo), Macromedia (flash, coldfusion), Adobe (acrobat reader), etc do this all the time. And once and for all -- the GPL IS NOT VIRAL. The GPL is simply a set of rules that one must follow if they "borrow" GPL'd source code for their own projects. You don't like the GPL? Well then, DON'T USE GPL'd CODE! Saying that the GPL is "viral" is as about as stupid as saying that the uniform commercial code and all related case-law defining one's freedom to contract should be thrown out. Thanks for your time, - Abraham Ingersoll <abeat_private> Dajoba, LLC - http://dajoba.com --- From: Cynthia Grossen <cgrossenat_private> To: "'declanat_private'" <declanat_private> Subject: RE: CEI blasts Linux as unsuitable for government, business use Date: Thu, 19 Sep 2002 18:45:16 -0400 James V. DeLong said: A second problem is the creation of applications for Linux. The General Public License that controls the program s distribution can be paraphrased as thou shalt not charge for this program and its source code shall be public. This license is also viral; if you write an ap for Linux, and incorporate any code covered by the GPL, then your ap is also subject to the GPL, and it too becomes open source and free. Then don't incorporate the code. Or make the GPL'd piece you need a component and distribute the source for that. While your code is only distributed in compiled form. Or make a deal with the author(s) of the program; most of us are only too happy to license our work to you, for a price. Or re-invent the wheel, if its all your own code you can do whatever you want with it. .cyn. --- From: "Derek Scruggs" <derekat_private> To: <declanat_private> Subject: RE: CEI blasts Linux as unsuitable for government, business use Date: Thu, 19 Sep 2002 17:29:23 -0600 > If IT companies, universities, and IBM want to donate the fruits > of their > labor to computer purchasers, including governments, that is their > privilege. But we have just gone through a half a decade in which the > business model was give it away, and it did not work. In the > end, software > might be bundled with hardware, or vendors might give away > software tied to > a services contract both are increasingly common -- but the code writers > will want pay for producing it, which means money must ultimately > come from > the users somehow. I am by no means an expert on open source, GPL or Linux (or BSD and its license), but the "university" part of this statement seems specious. Universities aren't just giving - they're taking. CS departments often don't have budgets to buy expensive proprietary software and hardware, even at educational discounts, so naturally they turn to open source where feasible. But I suspect that the wide pentration of Linux in academia is only one-half of the equation, the full manifestation of which is a feedback loop. Not only does the university get the software for free (or very cheaply), CS students - including of course the masters and PhD candidates who usually drive innovative research - get to hack the kernel without signing cumbersome NDA and licensing agreements. And they in turn often contribute the results of their research back to the community. (Dunno how this interacts with various institutions' patent-commercialization practices.) I haven't even read the NYT piece so I can't comment on Delong's response to it per se, but in this respect the money *doesn't* come from the user, except in its most diffuse form via research grants, alumni donations etc. Contrary to Delong's assertion about "give-it-away" business models, this phenomenon is the result of rational behavior by all parties. Linux, BSD and other open-source projects have matured considerably over the last decade, to the point that their use in the corporate data center is becoming more common. And though a lot of VC money was poured into Linux-commercialization startups, I suspect that we would still have seen wide adoption of these technologies anyway, if for no other reason than a lot of programmers who came of age during that era first cut their teeth on them. -Derek ----------------------------------------------------------------- Derek Scruggs Creative-Mail Email Results by Design "Results aren’t everything. They’re the ONLY thing." http://www.creative-mail.com 303-543-1186 425-920-6124 fax 303-808-6614 mobile ----------------------------------------------------------------- --- From: Eric Lee Green <ericat_private> Organization: BadTux: Linux Penguin Gone Bad To: declanat_private, politechat_private Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use Date: Thu, 19 Sep 2002 17:05:21 -0700 On Thursday 19 September 2002 02:36 pm, Declan McCullagh wrote: > positive that Jim's views are his and his alone (he's an economist, and it > can't be helped). :) > > Jim is responding to this New York Times editorial yesterday: > http://www.nytimes.com/2002/09/18/opinion/18WED2.html Okay, and I'm responding as someone who has made his living with Linux since late 1995, when I was the architect of the port of one of the first VAR applications ported to the Linux platform. > The New York Times recently editorialized about Linux and open source > software, exuberant that an operating system written and updated by > volunteer programmers in a communitarian spirit, > and available for free > might challenge Microsoft s Windows and result in major savings in computer > costs. > > The paper also exulted that governments such as Germany and China are > pushing Linux, and it urged everyone, including U.S. agencies, to join them > so as to foster competition. My view, as a long-time participant in the Linux industry: The VAR that I worked with went to Linux in 1996 because it was free, and because we had control over the platform. We could fix problems (unlike with SCO Unix, where several long-standing problems went unfixed for years), and we could underbid other companies for our contracts with school districts. Unfortunately, we ran into a major problem. We could get into the smaller school districts. They did not have entrenched purchasing and IT departments that had standardized on one product for all duties. We could not get into larger school districts. They said "If it doesn't say Microsoft on it, we won't buy it." Never mind that we were selling a solution including hardware, software, and support, not an operating system, and that said IT department would not be supporting the product. Unlike private industry, there was no incentive for a government bureaucrat to choose the best product for the best price. In addition, purchasing in government is also extremely influenced by money -- often even outright bribes. We ran into a problem with lobbyists from rich and powerful multi-national companies buying politicians. One company flew the entire state board of education and the state superintendent of education on all-expenses-paid vacations to tourist destinations in order to "view the software in action". Coincidently, they shortly afterwards ruled that all other software vendors (including us) could no longer do business selling school administration software in Louisiana. My boss raised enough stink about this corruption to get it into the major state daily newspapers as well as get a number of state legislators raising hell about why was the department mandating this expensive out-of-state solution when we had native-grown Louisiana software for less money, with the eventual result being that the ruling was overturned, but the point is that not all small software businesses have the political savvy to manage that, and politicians are prone to bribery. And most Linux software businesses are small. I interpreted the New York Times editorial as saying that such government restrictions against Linux and other competing products should be eliminated, and that all products should be evaluated on their own merits, not based on politics. > But the NYT misses in some ways. First, none of this is free. Software > is a complicated industrial product requiring continuing re-creation and > support, and money to support it must come from somewhere. You are correct. However, you inadvertantly repeat some incorrect information which has been promulgated by Microsoft Corporation, most notoriously in their "Linux Myths" paper which was a collection of innuendos, half-truths, and outright lies. See below: > Linux > programmers are not street people who sleep on steam grates so as to > indulge their passion. They are supported, often handsomely, by > universities and IT companies. Absolutely. As I noted earlier, I have been paid since late 1995 to write Linux software. > Even this support is not sufficient to keep > Linux going, The truth, or lack thereof, of that statement is ridiculous. What do you count as "going"? My current employer is using a 3 year old version of Linux in its embedded devices. This version of Linux works fine, and will continue working fine pretty much forever, since we are a hardware company and can update the drivers as necessary using our own in-house talent (ah, the advantages of source code!). Linux could keep "going" in that fashion pretty much forever, even if, tomorrow, every current Linux kernel guru suddenly Saw The Light and went to work for Microsoft Corporation to improve their Windows product. I've been paid to work on Linux for 7 years now. If that's not "going", what is? How many 7 year old computer software companies do you see out there? Let me give you a hint: Not many. This is a cut-throat business, and today's darling is tomorrow's bankruptcy. But the deal is that, unlike Be Inc. and its BeOS, Linux can't go bankrupt. That, more than anything else, is why my employer uses Linux. If we'd gone with BeOS, we'd be out of business just like Be Inc. We have control of our destiny with Linux. > and hardware companies, notably IBM, are now pouring billions > into it. There is nothing wrong with this; IBM has good competitive > reasons in that it wants to dish Sun and Microsoft. But the movement is > not the folk song army depicted in the NYT. True enough. As I'm sure you know, nothing gets done if there are no economic reasons to do them. I once wrote the beginnings of a paper exploring the economics that led to Linux. The most fundamental cause of Linux is that not all markets get filled by normal market forces. There are markets that are too small, or too risky, or not perceived as being a market by the people in marketing departments who detirmine such things. If there is a demand, economics says that this demand must be filled. In frontier times, if there was a demand for a barn to be built, but not enough people in the area to support a professional barn builder, the person who wanted a barn would often gather together many members of the community who would then do a "barn raising". In many ways Linux took place in that fashion. Ideologues such as Richard Stallman are often associated with Linux, and in fact did write major portions of the Linux software, but the fact is that Richard Stallman did not manage to produce a full-fledged operating system. His ideological approach produced only a few stand-alone programs only of use to fellow geeks. It took thousands of people worldwide, working in their own best interests, to produce a full-fledged operating system. And that of course that describes any economic system: people working in their own best interests to produce products that fill a need for themselves. Your job as an economist is to figure out the self-interest of these people, not to say that such self-interest does not exist (which is utter nonsense, as you correctly point out -- only self-interest creates products). > If IT companies, universities, and IBM want to donate the fruits of their > labor to computer purchasers, including governments, that is their > privilege. But we have just gone through a half a decade in which the > business model was give it away, and it did not work. At my current employer, we do not give away our product. It is embedded in a piece of hardware that we sell for major dollars. In my last job, when I worked for the backup software vendor, we made contributions to Linux (and FreeBSD) primarily in order to improve it as a platform for our commercial backup software. The same deal when I was working for the VAR -- we made contributions to Linux primarily in order to improve it as a platform for our solutions. Anybody who says that Linux is primarily a product of people generously giving things away is being ridiculous. People act in their own best interests. We gave away our contributions to Linux because it was in our own best interests to have a better platform for our commercial products, not out of generousity. > In the end, software > might be bundled with hardware, or vendors might give away software tied to > a services contract both are increasingly common -- but the code writers > will want pay for producing it, which means money must ultimately come from > the users somehow. Of course. > A second problem is the creation of applications for Linux. The General > Public License that controls the program s distribution can be paraphrased > as thou shalt not charge for this program Nonsense. Our BRU-Pro product included several GPL'ed programs, such as the "mtx" changer control program and the "lzop" compression program. We sold those programs just fine. We also provided the source code to those programs of course, as provided by the GPL, but nothing in the GPL prohibits you from selling a GPL'ed program. > and its source code shall be public. Yes. > This license is also viral; if you write an ap for Linux, and > incorporate any code covered by the GPL, then your ap is also subject to > the GPL, and it too becomes open source and free. This is only partially true. BRU-Pro was not (and is not) open source and free, despite having code covered by the GPL. In most cases you can negotiate a license with the author of the code for royalties if you wish to incorporate someone else's software into your product, or you can simply include the complete application as a seperate component. For example, we approached the publishers of the MySQL database (published under GPL) for permission to incorporate it into our product.. The MySQL people gave us a full royalty schedule to use for this. This is no different from approaching Oracle or any other vendor requesting permission to incorporate Oracle into your product. Similarly, we called the "mtx" program as a seperate program, not linked into our own code. Again, no GPL problem. What I see you saying is that we should have been free to simply steal someone else's code and incorporate it into our code for free. That's nonsense. That person released his code under the GPL license, but it's still his code, and if we want to incorporate it, we must approach the person and negotiate a seperate license to incorporate it -- just as if we wanted to incorporate Microsoft code into our product. > True open source believers think that this is just fine -- all aps should > be open and free. But it is not clear that the freeware spirit, or the > IT/university willingness to subsidize, runs deep enough to provide > anything approaching the number of aps available for Windows, where good > old reliable greed creates an incentive for developers. The Linux Yet there are tens of thousands of applications for Linux. You are correct, altruism did not create those programs. But they exist. How do you explain that this exists, if not for self-interest on the part of the developers? To me, you appear to be in the position of the theoretical scientist saying "It is impossible for a bumblebee to fly" -- all while the evidence is before him that said bumblebee is indeed flying. Go to http://www.freshmeat.net some day and browse the software available for Linux. You are correct, that altruism does not produce product. But something did. Your job as an economist is not to say that "Linux can't fly" -- it is flying. Your job as an economist is to figure out WHY it is flying -- what self-interest is motivating these people to release these programs for "free"? > community is moving toward proprietary aps, but it is chancy. Writing aps > without incorporating some operating system code is difficult, and those Utter nonsense and drivel. I have been writing commercial software applications for Linux since 1995, and that is the most ridiculous, ignorant repetition of a Microsoft *LIE* that I have ever seen. My employers over the years have employed some of the best lawyers around and consulted Richard Stallman (the master ideologue) and his own lawyers, and we have *never* run into an issue here. My last employer sold commercial Linux software from 1994 up until its parent company went bankrupt in 2001, and *never* had a legal issue regarding the GPL. My employer before that sold commercial Linux software from 1996 to this day, and has *never* had a legal issue regarding the GPL. At my current employer, our product incorporates a complete Linux software distribution -- and again, there is *no* problem with doing this. The fact of the matter is that we have 8 years of experience selling commercial software in the Linux world showing that there's no problem here. > who want to engraft proprietary aps onto Linux are taking a legal risk. Complete and utter nonsense and drivel. Proprietary applications have been sold for Linux since at least 1994, when Enhanced Software Technologies ported their proprietary "BRU" backup software to the platform. There has never (NEVER) been a legal risk. We have lawyers' letters, position papers from Richard Stallman and his own lawyers, etc. that very well lay out the legal boundaries as to what is proper and what is not. Now, if someone does not want to take advantage of all that public domain legal knowledge and does something stupid like steal someone's code without negotiating a non-GPL license with that person, well, that's no different than the case of someone doing something stupid like stealing Microsoft's code without negotiating a distribution license with Microsoft. But as long as you follow the law and the rules as laid out by the GPL license and the widely published legal scholarship surrounding it, there is no (NO) legal risk, and we have 8 years of commercial Linux software experience proving it. People who spread idiotic FUD nonsense like this are simply showing they know nothing about Linux and its marketplace. > Finally, governments should not treat this as an arena for industrial > policy. I agree utterly here. I just want the government rules prohibiting Linux repealed so that government managers can choose the best product for the task. Right now most government bodies mandate Microsoft software, and make it very difficult for a manager to buy a Linux-based solution. I just want a level playing field. I don't want anybody mandating Linux. > The incentives fueling the Linux movement are not necessarily > those required for long-term production of software suited for the public > as well as the nerds. Complete and utter nonsense. I was hearing this driven in 1996, when my employer was selling Linux-based VAR solutions into specialty marketplaces. It is ridiculous that, 6 years later, we still hear this drivel "Linux is just a fad". The fact of the matter is that Linux is well suited for a number of commercial tasks, and that has been the case for at least the six years that companies have been paying me to write Linux software. Six years is definitely "long term" in the computer business. > Governments, which are as naïve as editorial > writers, should keep their hands off. Well this is something that I can agree with. I do not want government mandating or subsidizing Open Source software. I just don't want government prohibiting Open Source software in those applications where it is the best product for the best price. Linux usage in private industry is booming -- close to 1/3rd of all servers in private industry are running Linux (albeit usually in various specialty tasks rather than as the primary server operating system). Linux usage in government is abysmal -- virtually no government bodies use Linux. That's because government doesn't operate under the profit motive, and thus has no incentive to buy the best product for the best price when considering a solution for a particular problem. The ridiculous state of the Indian trust systems (for which the Department of Interior has spent over $500,000,000 over the past 7 years on a failed trust system that just got Sec. of Interior Norton whanged for contempt of court for saying "It's working fine" when it wasn't, when they could have bought an off-the-shelf trust system for a few million dollars) is a case in point. Often government procurement is about empire building and insuring a comfortable retirement, not about the best product for the best price. What I want to see is government rules adding Linux to the list of "authorized platforms". I don't want to see anybody mandating Linux. As a long-time Linux software author and member of the Linux community, I am confident that my favorite operating system is perfectly capable of standing on its own merits as the best solution for the best price for a large number of tasks. But if we can't get into the door because government bureaucrats prohibit Linux, well, I think you see the problem. The profit motive is overturning such "Thou shalt not use Linux" edicts in private industry, but because government bureaucrats have no profit motive, there's no solution other than an edict from the top saying, "Thou shalt consider Linux as well as Microsoft". I am confident that, under those terms, Linux will have no problem providing needed competition for Microsoft in government markets. Give us access to the government market -- not subsidies, not mandates. Access. That's all we want -- and all we need. Eric Lee Green Chief Penguin Wrangler, BadTux Enterprises -- Eric Lee Green ericat_private http://badtux.org --- Date: Fri, 20 Sep 2002 03:16:22 +0000 From: Jim Smilanich <jsmilanat_private> To: "declanat_private" <declanat_private> Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use Declan; I'm sure that you're going to hear from a lot of other people on this subject, but let me give you my perspective as an enterprise software architect for a good sized bank (50,000+ employees). As such, I am one of the people responsible for determining what is the best choice of software and hardware for our company to choose to use. Therefore, as a representative of a fairly large user community it has been part of my responsibilities to follow the development of Linux, GPL code, and Open Source software in general for quite some time. Jim Delong fails to recognize a couple of critical facts concerning the GPL and open source development in general. First, no one is forced to use it. Stop for a second and re-read that sentence. NO ONE is forced to use the GPL for new development of code. Second, the GPL is carefully crafted to extend rights traditionally granted only to the creator instead to the end user of a product (software in this case). Therefore, if you don't want to use the GPL you are free to write code from scratch to meet your needs. The GPL only comes into play if you choose to use code written by someone else. I defy anyone to find a codebase licensed by IBM, Sun, Microsoft, Oracle, etc. that is placed under normal copyright restrictions that allows you to reuse any part of that code. As an end user, this is a plus for the GPL. Some people point to the BSD license and say that is a 'better' license because it does allow any reuse of code developed under it, including reusing that code in a closed source application. My response is, if the software developer in question chooses to use a BSD style license, then he is free to do so. He is also free to not release his source code at all. Or, to release it under the GPL. In each case, it's the DEVELOPER'S choice, no one else's. If someone wishes to build on his work, then they must honor his wishes or be in violation of US (and other countries') copyright law. Not to mention the ethical and moral questions involved in stealing someone else's work. Second, Mr. Delong misunderstands the economics involved in software development. It has been demonstrated time and again that less than 10% of all software development is directly for sale. What does this imply for his theory that 'good old reliable greed' has no place in open source software development? I respectfully submit that Sun, IBM, etc. are not necessarily motivated by altruistic notions about the good of mankind. Instead, I think that they see something that Mr. Delong does not; the only way to compete with a monopolist who has demonstrated time and again its willingness to break laws all over the world to maintain and extend its monopoly is to change the rules of the game. This is especially true when the executive branch of the US government has failed in its duty to maintain an even playing field in the marketplace. If that requires that Microsoft's competitors open up some of their crown jewels to the rest of the world in order to gain a long term competitive advantage, so be it. If they get instead the benefit of being able to sell lots of internal development and support contracts for that other 90% of software development, why shouldn't they support open source software? He also makes the statement that "Writing apps without incorporating some operating system code is difficult". This is false on the face of it. Computer systems are built with well understood layers between hardware, OS, network, and applications. This has been true since IBM scientists first articulated what became the OSI 7 Layer Model back in the late 1960s. Even prior to that, the concept was understood. The layers are defined by what are called Application Programming Interfaces, or APIs. No program which calls an API for a layer above or below it requires any code from those layers. This specifically includes operating systems code. Finally, Mr. Delong states that "...The incentives fueling the Linux movement are not necessarily those required for long-term production of software suited for the public as well as the nerds." and that this is why "...governments should not treat this as an arena for industrial policy." Maybe Mr. Delong should instead be asking himself why at least 2 dozen countries feel so constrained by the current closed source world that they feel it necessary to set policies encouraging open source (or at least open standards) solutions? I won't go into the arguments in what has already become a much longer response than I initially intended. Instead, let me point you to Dr. Edgar David Villanueva Nunez, Congressman of Peru's response to Microsoft. It brilliantly analyzes the closed source arguments and clearly lays out why governments have a responsibility to use open source and open standards software wherever and whenever possible. Note, by the way, that the good congressman does not advocate making industrial policy. Instead, he proposes that this is what his government should do, not what they should do to regulate the larger marketplace. This is, by and large, what most of the two dozen countries are doing. The URL is: http://pimientolinux.com/peru2ms/ One of the side effects of his and other governments' efforts may be to marginalize closed source companies. I would argue that this is a trend that is already underway. Those companies that understand the new marketplace and adapt will thrive. Those that don't will fail. This is simply a free market in action, not a socialist or communist plot. Respectfully, James Smilanich --- Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use From: Ryan Marsh <meat_private> To: declanat_private Cc: politechat_private, rmorrisonat_private, jdelongat_private In-Reply-To: <5.1.1.6.0.20020919142406.01a717c0at_private> References: <5.1.1.6.0.20020919142406.01a717c0at_private> Jim misses two key points that many people outside the community commonly miss. 1) Linux development has been, and always will be, driven primarily by the intrinsic motivations and reward structures of individual contributors. No lack of profitability, or usefullness to the average American, will ever stop development. When *geeks* decide Linux is not fun to play with anymore, Linux will die. That is the risk that must be assumed, not that companies will stop subsidizing it. Look at the individuals that are contributing the most work to Linux and Linux software, not much has changed with IBM's "billions". 2) The GPL *is* very restrictive but it is not impossible to work with, and it *is* getting better. But the GPL is a red herring, it is not the issue. NVidia releases a binary (closed source) hardware driver for the Linux kernel, and it does so legally and successfully. Also, just because the Linux *kernel* is governed by the GPL, does not mean that any applications that run on top of it must be. Oracle releases their fine database in binary form for Linux, and it performs very well without any need of kernel modifications. In fact, precious little software should ever require changes to the operating system's innermost core. -ryan --- Date: Fri, 20 Sep 2002 10:16:01 -0400 (EDT) From: David Herbst <vitamincat_private> X-Sender: vitamincat_private-hosting.com To: declanat_private Subject: RE: CEI blasts Linux as unsuitable for government, business use Declan, I appreciate Mr. Delong's candor in this editorial but I believe that he is misrepresenting both the legalities and economic impact of Open Source and the GPL. In reply to: >>"The General Public License that controls the programs distribution can be paraphrased as thou shalt not charge for this program and its source code shall be public. The GPL does not prevent one from charging for software, in fact in the GPL FAQ it explicitly states this position. There are many distribution providers who charge for packaged versions of open source software including SuSe and RedHat. More importantly than the above stated technicality, I believe that he misses where the actual economic benefit of the open source model is developed. It is true that open source software is not "free" to develop. Corporations, governments, and universities support the development indirectly through their staff that makes up the community of programmers. The result of this incremental contributions by these entities, however, are software applications which in many cases are far more usable (and valuable) than those developed under the proprietary model. As a hypothetical example, let's imagine that an application for tracking web server traffic has six components. Five of these components may be universal in there application with the one remaining component being specific to a particular organization's needs. Under the proprietary model, it is highly unlikely that a piece of software is going to be able to address all six of these components without becoming extremely bloated (read: expensive, buggy, slow). Alternatively, a software application developed under the open source model can cover the basics of the first five components, and with the source code can be tailored to fit the needs of the particular company /agency. This is the true strength of open source. The fact that the end user has the power and is not locked into any proprietary code or systems. It is not as Mr Delong eludes that, "true open source believers" think that everything should be "free as in beer". Regards, Dave Herbst --- To: <declanat_private> Cc: <rmorrisonat_private> Subject: CEI Economist's Report (Please withhold my e-mail address - thanks.) Declan, I wanted to comment on Jim DeLong's c:\spin article as posted on your list. Let me refute some of his points, point-by-point: > ...universities and IT companies. Even this support is not sufficient > to keep Linux going, and hardware companies, notably IBM, are now pouring > billions... I'm not sure what he means by "this support is not sufficient to keep Linux going". All of the empirical evidence points to the fact that this support is indeed sufficient to keep Linux going. Jim needs to clarify what he means by "keep Linux going". > But we have just gone through a half a decade in which the > business model was give it away, and it did not work. Again, we need some clarification here. What does Jim mean by "it did not work"? IBM has already seen a substantial return on their investment, and even though RedHat is loosing money, they reported a second quarter revenue of $21.2 million. This money has to come from somewhere; how is it "not working"? > Writing aps without incorporating some operating system code is > difficult, and those who want to engraft proprietary aps onto Linux are > taking a legal risk. This is not true at all. Code that does not use the OS kernel code itself, or some other GPL'd library, has no legal risk. That is why many libraries and tools such as Qt, Apache, and even Borland's Kylix have dual licenses; so those who want to develop free code can, and those who want to develop proprietary code can. > Finally, governments should not treat this as an arena for industrial > policy. The incentives fueling the Linux movement are not necessarily > those required for long-term production of software suited for the > public as well as the nerds. Governments, which are as nave as > editorial writers, should keep their hands off. Again, the empirical evidence does not back up his point about the incentives required for long-term production. Why has Linux not faltered? As long as Linux maes people money, people will work on it. Let's look at it this way. Why does IBM have financial incentive to develop AIX? Because customers want it. Increasingly, customers want Linux because they can realize measureable - sometimes substantial - cost savings. There is more of an incentive to use Linux than there is AIX; but we don't hear Jim asking governments not to use AIX. In conclusion, Jim is not saying anything clearly - and what he is saying continues to be refuted by the evidence we see every day. Thanks, --Josh --- Date: Fri, 20 Sep 2002 11:22:37 -0400 (EDT) From: rjhat_private To: declanat_private cc: politechat_private, rmorrisonat_private, jdelongat_private On 19 Sep, Declan McCullagh wrote: > > From: "Richard Morrison" <rmorrisonat_private> > > CEI C:\SPIN > > If IT companies, universities, and IBM want to donate the fruits of their > labor to computer purchasers, including governments, that is their > privilege. But we have just gone through a half a decade in which the > business model was give it away, and it did not work. In the end, software > might be bundled with hardware, or vendors might give away software tied to > a services contract both are increasingly common -- but the code writers > will want pay for producing it, which means money must ultimately come from > the users somehow. There is an error of sorts in the use of the word "donate". Two things that are generally not emphasized by the GPL or open source proponents are the concepts of: "Teaming Agreement" - which is perhaps a more accurate description of the GPL, and "Re-use" The GPL as a teaming agreement is much more accurate than the pejorative "viral" that Microsoft likes to use or the term "donation". It is a pre-set agreement, open to anyone, not mandatory on anyone. It comes with a price tag that may be unacceptable to some. The GPL teaming agreement is that: 1) Anyone may use the products of the team 2) Anyone may join the team 3) Only team members may distribute products of the team 4) Any team member may distribute any product of the team at any reasonable price It is the last two that are acceptable to some and not to others. It is a near certainty that at least one team member will distribute the product at cost, so you cannot build a business model around the assumption of exclusive distributorship. GPL has become an issue because the products of the various GPL teams are highly useful and valuable, and it is very easy for low level employees to incorporate GPL components and thus make their employer a team member without proper authorization. The "re-use" aspect is unique to software. In manufacturing we routinely "donate" our metal shavings, broken down boxes, and worn out packing materials. This "donation" is viewed as a cost saving because the recyclers are much less expensive than trash haulers. In software, I can "donate" internally developed software to "recyclers" without losing my ability to use the software. Since 80% of software development remains internal corporate efforts, there is a very large amount that can be re-used at no extra cost to the developing organization. This is one reason why open source has been much more successful for corporate platform software rather than end user software. Corporate platform software is not a revenue source for these companies. It is a cost. The re-use of corporate platform software benefits corporate developers by lowering their platform cost without reducing their revenues. Lower cost without lower revenues is the motivation for most corporate teaming agreements. This view explains Microsoft's intense opposition to the GPL (Microsoft is the primary vendor for corporate platform software), the concentration of successful open software in platform areas, and the scarcity of open end user software. R Horn --- Subject: a correction to "CEI's Weekly Commentary: Software Wars" From: Ed Hill <edat_private> To: Richard Morrison <rmorrisonat_private> Cc: Declan McCullagh <declanat_private>, jdelongat_private Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-kVDgVsIqmzEYDy30Cs0Y" X-Mailer: Ximian Evolution 1.0.8 Date: 20 Sep 2002 09:40:07 -0600 X-UIDL: d99a17cfc9d02bea432da3b259be5516 Dear Richard Morrison, I'm writing in response to your recent rebuttal of the NYT Linux editorial: http://www.politechbot.com/p-04007.html While you do have some worthwhile observations, the following quote: "The Linux community is moving toward proprietary aps, but it is chancy. Writing aps without incorporating some operating system code is difficult, and those who want to engraft proprietary aps onto Linux are taking a legal risk." is false. It is, in fact, quite easy to write programs for Linux (and Windows and Mac OS and commercial Unixes) that neither "incorporate" nor "engraft" any operating system code. Thus, there is no basis to the claim that application authors are somehow endangering their code through source pollution. Rather, the applications that run on the OSes listed above "link against" (or "use calls to") the underlying operating system. Within the industry and within the operating systems licenses, there is a long standing tradition that such use of the OS by application software is not an intermingling of the OS and application code. It is merely normal use of the operating system [1]. In all other respects, I enjoyed your rebuttal and I look forward to your future musings. Regards, Ed [1] For examples of commercial code on Linux which both obey the appropriate licenses and vigorously maintain their intellectual property (that is, are not in any way "corrupted" by the GPL), please see the commercial Linux software applications of the following companies: http://www.oracle.com/ http://www.ibm.com/ http://www.borland.com/ http://www.mathworks.com/ http://www.ni.com/ http://www.sun.com/ http://www.opera.com/ http://www.macromedia.com/ which include everything from huge enterprise-scale databases and software development tools to small web browsers and tiny desktop "helper" (or "plug-in") applications. Do you honestly think that the legal departments of these companies would allow their applications to be ported to Linux if there was even a wisp of support for your "legal risk" assertion? -- Edward H. Hill III, PhD Post-Doctoral Researcher | Email: edat_private, ehillat_private Division of ESE | URLs: http://www.eh3.com Colorado School of Mines | http://cesep.mines.edu/people/edhill.php Golden, CO 80401 | Phone: 303-273-3483 Fax: 303-273-3311 --- Date: Fri, 20 Sep 2002 12:48:20 -0400 From: -dsr- <dsrat_private> To: Declan McCullagh <declanat_private> Cc: politechat_private, rmorrisonat_private, jdelongat_private Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use As reported in Politech: > But the NYT misses in some ways. First, none of this is free. Software > is a complicated industrial product requiring continuing re-creation and > support, and money to support it must come from somewhere. Linux > programmers are not street people who sleep on steam grates so as to > indulge their passion. They are supported, often handsomely, by > universities and IT companies. Up to this point, Mr. DeLong is correct. > Even this support is not sufficient to keep > Linux going, and hardware companies, notably IBM, are now pouring billions > into it. There is nothing wrong with this; IBM has good competitive > reasons in that it wants to dish Sun and Microsoft. But the movement is > not the folk song army depicted in the NYT. Linux has all the support it needs. The question is, are businesses willing to pay for the Linux support *they* need? Historically, the answer is yes. Businesses pay for customization of software all the time, from single-shop point-of-sale systems to gigantic salesforce automation packages. America's corporations now understand that the four costs of a new software system are the initial price, the customization fee, the ongoing support, and the hardware upgrades. Open Source software (including Linux) tends to eliminate the initial price, and drags down the customization and support charges. Why? Because OSS is extremely pro-competitive, while simultaneously bringing the benefits of wide-scale cooperation to the table. When an enterprise chooses Siebel, for example, the choice of customization and configuration service providers is constrained: buy support from Siebel, or suffer. When support requirements go beyond in-house techs, Siebel stands to make more money. With an Open Source solution, on the other hand, there are thousands of competent programmers who have access to the source and can do the customization your business requires. That same source availability means that your in-house techs can find bugs themselves, and fix them -- and the GPL's "share and enjoy" requirement means that the bugs have often been reported and fixed by another user of the same software. Compare this to the experience of having your bug report swallowed up by a large software house -- and being told, six months later, that the fix is in the next version of the software, please pay this upgrade fee. Or worse, told that there are no plans to fix that bug. > A second problem is the creation of applications for Linux. The General > Public License that controls the program s distribution can be paraphrased > as thou shalt not charge for this program and its source code shall be > public. This license is also viral; if you write an ap for Linux, and > incorporate any code covered by the GPL, then your ap is also subject to > the GPL, and it too becomes open source and free. Mr. DeLong perhaps believes that the fruits of others' labor should be free? No? Then why does he object to paying a fee in kind, rather than in dollars? That's what the GPL amounts to, really: the fee for this software is that if you enhance it or change it -- *and* you distribute it to others outside your company -- you must provide the source code for your changes. If you don't make changes, or if you make changes but don't distribute them, you don't pay anything. That should be eminently reasonable to anyone who believes in the right to own property. > Writing aps > without incorporating some operating system code is difficult, and those > who want to engraft proprietary aps onto Linux are taking a legal risk. I can see from this that Mr. DeLong is neither a programmer nor a lawyer, because a programmer would not agree with the first half of this statement, and a lawyer would not agree with the second half. The GPL is clear: writing software which does not incorporate GPL'd code does not bring the new software under the GPL. Even if you write all your code with a GPL'd editor, compile it with a GPL'd compiler, debug it with a GPL'd debugger, and require a GPL'd operating system to run on, the new software is all yours. One has to actually copy someone else's GPL'd work into your new program to bring it under the GPL. You are safe even if you copy your own GPL'd code into a new program, because as copyright owner, you are free to relicense your own works any way you please. > Finally, governments should not treat this as an arena for industrial > policy. The incentives fueling the Linux movement are not necessarily > those required for long-term production of software suited for the public > as well as the nerds. Governments, which are as na?ve as editorial > writers, should keep their hands off. Mr. DeLong is clearly unaware that the US government is and has been a major sponsor of Open Source software for many years. -dsr- --- Date: Fri, 20 Sep 2002 11:19:11 -0700 From: Grant Gross <grantat_private> To: declanat_private, jdelongat_private Subject: Jim Delong's criticism of Linux Hello, Declan and Jim: A couple of points on Jim's rant against Linux and The New York Times editorial this week: Yes, Linux programmers are supported. Most have "day jobs," and some of those are even jobs where they're paid to write Linux or Open Source code. And yes, companies like IBM have supported Linux (although much of the billion dollars IBM has spent on Linux has gone into its marketing efforts, as opposed to code-writing efforts). But Linux has survived and grown for several years without huge corporate subsidies. While IBM and other companies have contributed mightily in the past couple of years, Linux and most other Open Source software has a long grassroots history. Much of the Linux movement is really is a "folk song army," as Jim puts it. Corporate support has helped Linux, but it is not necessary "to keep Linux going," as Jim suggests. Linux survived for nearly 10 years without backing from many large companies. The bigger question is: So what? If IBM and other companies choose to support Linux because they see profit in doing so, what's wrong with that? Sounds like free enterprise to me, and as I understand it, CEI is all for letting the free market work. Or does CEI only support the free market for large corporations like Microsoft? Doesn't a folk song army, supported by a few companies, also have the right to participate in the free market? In addition, Jim's comparison to Linux as a movement to the dot-com bust is overly simplistic. First, Jim complains that someone has to pay for Linux by supporting it, then he complains that Linux isn't a sustainable business plan. First, he criticizes Linux for not being free -- yes, some Linux companies offer support services -- then he criticizes Linux because giving products away didn't work for dot-coms. "The code writers will want pay for producing it, which means money must ultimately come from the users somehow." There are a lot of things wrong with that statement. Many of those code writers have been volunteers -- I imagine Jim understands the concept of volunteerism. Jim is thinking of Linux purely as a business model, when Linux has only recently become a business. Interestingly enough, The New York Times' editorial didn't advocate Linux as a business model. It just advanced Linux as competition to Microsoft. Jim's second critique of Linux and the GNU GPL is off target. It's interesting, of course, that Jim's criticism of the GPL mirrors the recent noise coming from Microsoft, a past backer of CEI. Many people have responded to the claims that the GPL is "viral," but let me try to put it in terms conservatives should be able to understand: The GPL is a kind of social contract, or even a business contract. GPLed code is free to use, but if you build on it, you have an obligation to contribute your code back from the commons from which you took the code. As you free marketers like to say, "You can't get something for nothing." There is a price for using GPLed code, and if you'd don't want to pay that price, the choice is simple: Don't use GPLed code, create your own. Also, Jim paraphrases the GPL as saying: "Thou shalt not charge for this program and its source code shall be public." Jim, could you please point out where in the GPL it says you can't charge for a GPLed program? If that's the case, those copies of Red Hat and SuSE Linux I see on the shelves of my local Best Buy violate the GPL. Free Software Foundation founder Richard Stallman himself tells audience how he used to charge to send out copies of his Emacs text editor on disk. As you know, Richard wrote the GPL. Finally, Jim suggests that governments should stay away from Linux, because "incentives fueling the Linux movement are not necessarily those required for long-term production of software suited for the public as well as the nerds." Again, Jim contradicts himself. If companies like IBM are pouring billions of dollars into Linux, that feels to me like the foundation for the long-term production of software. Thanks for listening, Grant --- Date: Fri, 20 Sep 2002 13:20:53 -0500 To: declanat_private Cc: rmorrisonat_private, jdelongat_private Subject: Mr. Delong's views on Linux and other open-source software Thanks to Linux Today, I read a post made by Jim Delong concerning the NY Times editorial about the use of Linux in government and business use. While Mr. Delong may have been attempting to write a piece that is fair concerning open source and free software, it would appear that his information concerning how such things work is sorely lacking. While I could debate the statement that "Software is a complicated industrial product requiring continuing re-creation and support" (a point of view that finds it hard to explain how people can choose to run old and unsupported software such as Windows 95 or Office 2000, even though many millions of people do just that) I would rather focus on the last three paragraphs. The article in question seems to imply that all Linux software is distributed under something called the "General Public License". I presume that Mr. Delong means the "GNU General Public License" as I am unaware of any other license that might fairly be described as THE "General Public License". Mr. Delong then proceeds to "paraphrase" this license as "thou shalt not charge for this program and its source code shall be public". However, such a paraphrasing is certainly not true. Indeed, the original source of the license, the Free Software Foundation, charges for the software that the GNU General Public License was created for. What one cannot do when licensing his program under the GPL is to prevent the users of that software from distributing it to third parties under whatever terms they choose, including free of charge. Nor can one fail to make the source code available should one of the original users ask for it nor can one limit the distribution of the source to third parties. However, it's not clear that this is important to Mr. Delong's second argument. While the Linux kernel and some programs are distributed using the GNU General Public License, there is absolutely no requirement that any other programs be distributed using that or any compatable license. Indeed, the vast majority of Linux software is distributed under a license that is NOT the GNU GPL and many people (myself, included) routinely use Linux software that is available as a binary and whose licenses require that I not distribute the software to third parties. There is no conflict because the license of an application has exactly nothing to do with the license of any other software on a Linux computer. While Mr. Delong states that "writing aps without incorporating some operating system code is difficult" that statement is entirely false. As an economist, that may not be obvious to Mr. Delong, but as a programmer with decades of experience it is obvious to me. None of the applications I have written for ANY operating system include ANY of the operating system's code. Programming simply doesn't work that way and there is a clear boundary between operating system code and application code. Indeed, should Microsoft so choose, they could release any of their applications using their usual license and there would be no conflict. If that were the case, many people would actually welcome the effort because it would be seen by many as legitimizing Linux as a desktop operating system. When I look at Mr. Delong's article with that understanding, I see that his conclusion is not supported by his argument. Hardware sales alone can support the continued development of the Linux kernel and similar infrastructure software. Application software can be paid for through the proceeds from sales of that software because the distribution of that software can lawfully be restricted. While it seems grand to talk about the "number of aps available for Windows" it is not clear to me that there exists a large gap in the number of applications between Linux and Windows. In fact, if games are neglected, the interest in and work towards duplicating Windows software on Linux focuses entirely on fewer than 20 applications and it has been some time since Linux computers had no equivalent to one or more of those. Or, look at it another way: I've got some 10000 distinct applications available for my Linux system. What am I missing? -- Jonathan Guthrie (jguthrieat_private) Sto pro veritate --- Date: Fri, 20 Sep 2002 12:54:34 -0700 (PDT) From: Brandon Bidewell <brandonat_private> X-X-Sender: brandonat_private To: Richard Morrison <rmorrisonat_private> cc: Declan McCullagh <declanat_private>, <jdelongat_private> Subject: Re: CEI's Weekly Commentary: Software Wars I'd like to make a few points about Richard Morrison's article: http://www.politechbot.com/p-04007.html 1) "First, none of this is free." For most people, the fact that they have to pay for software makes it non-free and, alternatively, software that can be acquired for absolutely no monetary cost if "free." In this sense Linux and many other software programs that make up a linux distribution are free. Secondly, you may have been alluding to the fact that software usually restricts that user of that software, and in this case Linux is again free unless you consider the "freedom" of a user to further restrict other user's uses of the software (you may be aware the GPL prohibits these actions). Linux is free in this sense as well, and certainly software covered under many standard EULA and standard copyright protection is not free. Lastly, you may have been alluding to the cost of using software (i.e. acquiring, installing, understanding, training and maintainance). No software avoids these costs, and this is obvious to everyone but you. Conclusion: Linux is free in more ways than one, but don't avoid the user's cost of using software. In other words, using software is not free. 2) "But we have just gone through a half a decade in which the business model was give it away, and it did not work." This is extremely simplistic. If true, why does Microsoft continue to bundle many applications (IE, Media player) with operating software for "free" when other companies sell them? Why do many people purchase Linux / Unix software when it can be downloaded for free? Why did many dot-com companies that gave services away for free fail, and so did many other companies that sold services? This statement is a red herring. The fact that many over capitalized companies blew there money giving things away for free is no reflection on the practice of individuals or companies not in the business of selling software, giving away software for "free." 3) "This license is also viral; if you write an ap for Linux, and incorporate any code covered by the GPL, then your ap is also subject to the GPL, and it too becomes open source and free." If this makes GPL software "viral" (your word), then all software except public domain software is "viral," including proprietary software. Developers are always welcome to negotiate non-viral licences from copyright holders of any software. 4) "The Linux community is moving toward proprietary aps, but it is chancy. Writing aps without incorporating some operating system code is difficult, and those who want to engraft proprietary aps onto Linux are taking a legal risk." What is your definition of "incorporating some operating system code?" It is very easy to create "proprietary aps" without being covered by the GPL. If people who create software violate copyrights of other software, it doesn't matter if software is covered under the GPL or not. The same legal penalties apply, although practically the consequences are much different. Also, the "Linux community is" NOT "moving toward proprietary aps." Many commercial software companies are trying to sell "proprietary aps" that run on Linux. The fact that high quality, freely available and freely modifiable alternative now exist and run under Linux makes this difficult. You clearly shouldn't have opinions about something you don't understand or cannot clearly articulate. Brandon Bidewell --- From: <naumat_private> To: declanat_private Subject: Competitive Enterprise Institute Date: Fri, 20 Sep 2002 16:24:03 -0400 MIME-Version: 1.0 I was surprised to read your take on CEI on politechbot.com ... Basically, they're a "paid for PR politcal agenda" deal - prwatch.org has a rap sheet on them - http://www.prwatch.org/improp/cei.html. I would give less creedence to them than I would RIAA or MPAA ... http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&q=%22competitive+enterprise+institute%22+site%3Awww.prwatch.org --- Date: Fri, 20 Sep 2002 16:30:29 -0400 From: Decklin Foster <decklin@red-bean.com> To: rmorrisonat_private Cc: declanat_private, jdelongat_private Subject: Re: FC: CEI blasts Linux as unsuitable for government, business use Declan McCullagh writes: > Subject: CEI's Weekly Commentary: Software Wars > Date: Thu, 19 Sep 2002 17:13:37 -0400 > From: "Richard Morrison" <rmorrisonat_private> > > This week's c:\spin, is by James V. DeLong, Senior Fellow, Project on > Technology and Innovation, CEI, September 19July 19, 2002. [...] I hope my reply does not get lost among a flood of "BUT LINUX RULZ DOOD!" emails. I'm a long-time user of GNU/Linux and supporter of free software, and I think you've done a good job of furthering the debate but I have few issues and corrections I would like to point out. > none of this is free. Software is a complicated industrial product > requiring continuing re-creation and support, and money to support it > must come from somewhere. This is absolutely true and needs to be repeated. Acquisition costs really only matter if you're worried about making a $199 consumer PC. The license fee for a mail server, on the other hand, is nothing compared to paying someone to administrate it full-time. What I would like to add is this: It's not only money for support that must come from somewhere, it's the ability *to* support a given piece ofsoftware. Take the mail server for example: it may be well documented, and you may be able to learn how to configure almost every aspect of its operation, but if you need to add a feature or if it crashes, and the code is proprietary, you'd better hope the vendor has a economic incentive to listen to you. My opinion is that all software should be free, but this is more because of "greed" than any sort of "let's share everything, money is bad!" idealism. If I have the technical ability to extend or fix a mail server, why shouldn't I get paid to do so? I would be much happier getting paid for my actual *work* than expected returns on a product that will eventually be sold. If the mail server is free, I can do this. If it is proprietary, obviously, I cannot. I do not claim that free software's development methodology, or pool of talent, will produce a mail server which is necessarily better. But I would be much more cautious about the government opening public services up to vendor lock-in than I would be about software regulation. Laws are not the only things that can be dangerous. > Linux programmers are not street people who sleep on steam grates so > as to indulge their passion. They are supported, often handsomely, by > universities and IT companies. Even this support is not sufficient to > keep Linux going, and hardware companies, notably IBM, are now pouring > billions into it. This argument is flawed. What it seems to state is that because IBM is spending billions on GNU/Linux research and development, Linux would stagnate and become obsolete if IBM were not doing so. On the contrary, GNU/Linux has thrived for many years before IBM decided to make use of it, and shows no sign of becoming dependent on any kind of sponsorship. I read package changelogs every day while upgrading my Debian system and the vast majority of improvements documented there are done by ordinary hackers in their spare time. Contributors to GNU/Linux are simply that -- people who have the time, skills, and inclination to make a contribution. The pool of such people is always changing, but the software keeps on going. If a scientist loses his research grant, whatever progress he might have made for mankind is not automatically lost; someone else can always step in and use what has already been achieved to increase the knowledge available to all. If a prolific hacker who made useful contributions to GNU/Linux lost her "funding", development would not come to a standstill. This is possible precisely *because* any code she had written would still be an open book. This does happen in practice; occasionally Debian developers send out messages to the effect of "some disaster happened, I'm relocating and/or finding a new job so I'll be without net access for a month; if there are any bugs in my packages, please upload a fix." And if there are, other developers do. Sometimes developers retire from the project completely, in which case their packages are "adopted" by newer maintainers. I have done the latter part of this myself several times. > IBM has good competitive reasons in that it wants to dish Sun and > Microsoft. But the movement is not the folk song army depicted in the > NYT. This is a good thing, and I wish the NYT would have emphasized it more. There is a very strong tendency currently to think of software as a product, since, in the modern (mostly-Microsoft) computing environment it is marketed as such. I am idealistic about one thing: Software should be thought of as more like a natural resource, a public infrastructure, or a shared body of knowledge. When IBM uses GNU/Linux to give itself a competitive advantage over Sun or Microsoft, it is fulfilling this idea. You could say that the question ought not to be "What can companies do for GNU/Linux?", i.e., to improve its viability as a "product", but rather, "What can GNU/Linux do for companies?", i.e., how shared code can enable them to achieve important goals. Zealots who are out to prove GNU/Linux must necessarily be better should just be ignored. Sharing code is *not* an end in itself; it is a means to being more productive as a society. (As an aside, This is why I do not feel particularly religious about the GPL. Because the rest of the world is still mostly propietary, and working with the rest of the world is important, there is a place for the GPL, the LGPL, and simple BSD/X/MIT-style licensing. The diversity here is a good thing.) This is, after all, the purpose of software -- helping humans to use computers effectively. My personal opinion is that by turning it into a "product", we have turned this relationship on its side, if not completely upside-down. > But we have just gone through a half a decade in which the business > model was give it away, and it did not work. As I have alluded to, I think we need to be skeptical about "business models" as the driving force behind software development. I will admit freely that our economic system is simply not set up to reward programmers for doing actual work that results in actual improvements to the utility of software for the human race in general; it is much more geared toward rewarding them for helping software publishers create an attractive product that can be licensed over and over for money. > Writing aps without incorporating some operating system code is > difficult, and those who want to engraft proprietary aps onto Linux > are taking a legal risk. This is merely factually incorrect. The system interfaces provided by Linux, glibc, and most useful libraries (for example, GTK+, the most popular graphical user interface toolkit) are well-documented and available for use by applications written under any license. No code needs to be copied from a GPLed program in order to write a proprietary Linux app. I hope you managed to get something from this, and that it didn't just sound like arguing. I haven't done advocacy for a quite a while, but the rational tone of your article (i.e., "let's consider the questions here before deciding Microsoft Is Better(TM) or Linux Must Win(TM)"), which sadly is not very common with this sort of thing, encouraged me to help you refine your points. Even if you don't consider it useful, I feel more optimistic about the character of future debate on the subject. -- things change. decklin@red-bean.com --- Date: Fri, 20 Sep 2002 22:10:24 +0000 (UTC) From: Barry Fitzgerald <fitzixat_private> To: jdelongat_private cc: declanat_private, <rmorrisonat_private>, <fitzixat_private> Subject: RE: Free Software production Hello Mr. DeLong, I just read your latest CEI C:\SPIN submission "Software Wars" and felt the need to comment on it. I feel that your portrayal of the Free Software and Open Source movements is largely inaccurate. You seem to push the "Free Software volunteerism is a myth" line that has recently been spewed by numerous think tanks. You state that a significant amount of Free Software (in particular, GNU/Linux) development comes from Universities and IT companies. I find it very interesting that a person who has no evidence to back up his statements is so comfortable making them. Allow me to introduce myself. I am the person that you seem to think doesn't really exist. I am a worker in the IT field who holds a BA in Sociology. Almost all of my IT knowledge is either "home grown" or comes from prior work in the field. I also come home at night and work on Free Software. My name is Barry Fitzgerald, and I am a member of the DotGNU Steering Committee. I am not paid for my work. Having said that, I understand something about what it takes to work on Free Software. I understand the limitations that are placed on people's time when they work on Free Software in their spare time. If you were to do a code audit on DotGNU's code, you'd be hard pressed to find me. Largely, my work has been involved in shaping the idea of what DotGNU will be and where it's going. I have been involved in project design and general decision making. I have not done enough. It is time consuming, that is true. But, to take that one step further and state that it doesn't have the capability to move forward at an acceptible pace, is to simply ignore reality. In my time as a member of the Free Software community, I have seen all manner of people committing their time to Free Software for any number of different reasons. Some of them were paid for their work. Others were doing it because it was fun. Still, other people put their resources into this because they were devoted to the cause. By far, the most motivated and most productive people were those who were both devoted to the cause and had personal reasons for producing or advancing those products that they were working on. These are the Donald Beckers of the world, who work on kernel Ethernet code because doing so makes their job easier. These are the Rhys Weatherleys of the world, who work feverishly to produce something because they feel that there should be an alternative to a proprietary product and that there is a market/audience for that product as Free Software. And for all of these people, there's a reason why they do it with Free Software. The goal is not to base a production model off of a "freeware" platform. You're just missing the point if that's what you think. To think that people participate in IT simply because they are attaining a monetary goal is to discount the reason that many people actually participate in *ANY* IT activity, including helping people in IRC and creating software. People do it because they want to do it. The fact of the matter is that most people aren't paid for the act of initial creation. Dare I say that most true inventions (not just innovations - I mean real, unadulterated invention of something completely original) have not occurred at corporate mandate. Science has rarely been a pursuit based on profit and, in the cases where it has been, has yielded less than lackluster results. But even looking beyond original invention (which you'd be hardpressed to find anywhere), people produce and participate because they feel the need to. Before I was a member of the Free Softare community, I was a member of my local BBS community. In that community, we setup bulletin boards because we felt good about doing it. I know of not a single BBS sysop in my area that actually profited from their BBS. It was always an act of love and, in many ways, most of these BBS' provided a better experience than larger online services ever could have. Eventually, the BBS' gave way to the Internet because the the Internet's scope. However, that is not a testiment to commercial ISPs being better than BBS's at providing support. It is exclusively a testiment to the scope of the internet. Then there is this crown jewel: "If IT companies, universities, and IBM want to donate the fruits of their labor". This, again, misses the point. If you weren't aware, the point of producing something is to have people use that product. I know, it's a revolutionary thought. Perhaps people should not be able to tie draconian demands to something simply for their own personal gain. Again, another revolutionary thought. I'm sure many people in totalitarian nations (not to mention proprietary software companies) would find something wrong with my argument. However, being a supporter of the ideal of democracy and of the idea that people should be able to openly participate in their existance, I denounce the existence of draconian proprietary software licensing. It's not an act of donation. This is simply the act of passing on what we originally had to work with. As far as production models are concerned, you might find this interesting for your argument against open production models. It is a simple fact that almost all commercial products are based off of technology produced and developed in an environment that is open. The web browser/server, the TCP/IP stack, the command line interface, the graphical user interface, etc... etc... etc. I believe that you'd be hard pressed to think of any item that is not based off of information that existed in the public intellectual commons, much less a proprietary item. So, I ask you this: what is wrong with leaving others with the same resources that you had when you began? Or, is it an inherently capitalist idea to ensure that people are ignorant and legally incapable of participation in their own lives? Is it an inherently capitalist idea that independant entrepreneurs must be left with nothing with which to create their products, since all forms of information exchange have been locked up by a copyright and patent system spun horribly out of control and which operates at the whim and mercy of large established corporations? I would tend to say that the problem is not capitalism. But with people espousing the view that we must bleed the commons dry for the purpose of the personal gain of the exclusively wealthy - it is not hard to see why that is the world that we seem to be living in. This is the wrong form of capitalism to espouse. A functional and equitable form of capitalism is an interdependant capitalism. Proprietary production amounts to nothing short of dependant capitalism and produces a loss of equal opportunity and inherently generates monopoly. That is, by definition, the goal of proprietary licensing. I would love to address your other concerns about Free Software (which are dispelled largely due to what I have said above), but this letter is long enough as it is. fitzixat_private SDF Public Access UNIX System - http://sdf.lonestar.org --- Date: Fri, 20 Sep 2002 16:37:04 -0700 (PDT) From: "Shane R. Stixrud" <shaneat_private> To: declanat_private, <rmorrisonat_private> Subject: Re: Software Wars: Open Source and the N. Y. Times A counter view point to James Delong's post. ---------- Forwarded message ---------- Date: Fri, 20 Sep 2002 16:33:18 -0700 (PDT) From: Shane R. Stixrud <shaneat_private> To: jdelongat_private James, In regards to your response to "Open Source and the N. Y. Times" I want to first thank you for taking the time to comment on these issues, I think all points of view here are healthy and will ultimately lead to the truth. I am not an economist nor do I believe I am an expert in business or law. I am however a user of both proprietary and open source software in large scale enterprise environments. I am currently employed as the Sr Network Engineer for a large medical organization (10,000 users, 30+ locations). I have been part of/used Free/Open source software for many years. I first became involved when I discovered the Internet (1992) back then the net was small enough where not much significant happened without everyone knowing about it. The concepts behind free/open source idea/software sharing was part of everyday life If you could take the time to read my responses to your comments below, I would love to hear how my ideas fit or don't fit into your model of how things work. >The New York Times recently editorialized about Linux and open source >software, exuberant that an operating system written and updated by >volunteer programmers in a communitarian spirit, and available for free >might challenge Microsoft s Windows and result in major savings in >computer >costs. > >The paper also exulted that governments such as Germany and China are >pushing Linux, and it urged everyone, including U.S. agencies, to join >them so as to foster competition. > >The NYT view has some gold. Competition is always good. And the Linux >backers have hold of an important truth, which is that persuading a lot of >smart people each to devote a small part of their time to an effort can >produce impressive results. They are also right to think that opening up >computer code to the eyes of the whole programming community can be >extremely productive. Microsoft itself sees increasing virtue in this >idea, and is developing shared source to open up code to scrutiny while >the company keeps firm hold of the pen. I think it is important to note that Microsoft's approach of look but don't touch takes away the most powerful aspect of Open/Free Software. That is the ability to resolve painful issues that are hindering/harming a user. As a result it is my belief that few people will be interested in taking the time examine and write software which has no guarantee of being a) viewed and acknowledged and b) Helping solve their problem. >But the NYT misses in some ways. First, none of this is free. Software >is a complicated industrial product requiring continuing re-creation and >support, and money to support it must come from somewhere. You are correct and Linux re-creates itself based on user needs. The money comes from different places. The most overlooked and I believe the largest contributor is from businesses who deploy Open/Free software to solve problems. If a large organization has a problem with DNS where it The people who write Free Software are those who use it, a lot of the time it is part of their day job, I believe this is become more common not less. We are beginning to see money/resources toward Free Software development come from a place that is uncommon in our software as a product market, but very common in other labor industries. I will address this further at the bottom of this email. >Linux programmers are not street people who sleep on steam grates so as to >indulge their passion. They are supported, often handsomely, by >universities and IT companies. Even this support is not sufficient to keep >Linux going, and hardware companies, notably IBM, are now pouring billions >into it. There is nothing wrong with this; IBM has good competitive >reasons in that it wants to dish Sun and Microsoft. But the movement is >not the folk song army depicted in the NYT. Whether the money/resources provided by Universities and IT companies is sufficient depends on who you ask. I do believe you are correct in saying that companies like IBM are greatly accelerating what markets / solutions Open/Free software is a solution for. I do not agree that just because Big business like IBM are donating money to Free software that this makes the movement any less "pure". As long as the ideals/licenses require everyone to play fairly, seeking advantage by adapting and putting in hard work does not harm society/community in any way. >If IT companies, universities, and IBM want to donate the fruits of their >labor to computer purchasers, including governments, that is their >privilege. But we have just gone through a half a decade in which the >business model was give it away, and it did not work. In the end, software >might be bundled with hardware, or vendors might give away software tied to >a services contract both are increasingly common -- but the code writers >will want pay for producing it, which means money must ultimately come >from the users somehow. I do not see these companies or universities as giving away anything. What these organizations are getting far out ways what they are contributing. That's why this model is working, IBM might donate 10 billion to Linux software development, but what they are getting in return both in software and potential business revenue must exceed this number, or it would be a bad investment. Whether this is a bad investment only time will tell. >A second problem is the creation of applications for Linux. The General >Public License that controls the program s distribution can be paraphrased >as thou shalt not charge for this program and its source code shall be >public. This license is also viral; if you write an ap for Linux, and >incorporate any code covered by the GPL, then your ap is also subject to >the GPL, and it too becomes open source and free. I believe you hold a few misunderstandings about the GPL license. If I was to paraphrase I would say: You may charge whatever the market will allow for the sale of this software. If you decide to distribute this software to a third party in binary form, source code must be provided giving the users the same rights as you were given. If you wish to modify this software for your own internal use you are not required to provide outside parties anything. >True open source believers think that this is just fine -- all aps should >be open and free. But it is not clear that the freeware spirit, or the >IT/university willingness to subsidize, runs deep enough to provide >anything approaching the number of aps available for Windows, where good >old reliable greed creates an incentive for developers. The Linux >community is moving toward proprietary aps, but it is chancy. Writing aps >without incorporating some operating system code is difficult, and those >who want to engraft proprietary aps onto Linux are taking a legal risk. I believe we are moving into a new software model where we have two main choices: On one side we have: Proprietary Model of "Software as a service" In this model end users (whether that be a corporation or individual) subscribe to a software service with a recurring fee. Depending on the service they pay for, they are given a fixed set of software "functionality". They have the right to _use_ this software as long as they continue to pay the recurring fees. The proposed benefit is that software will cost less up front and supposedly over the long term (when including software upgrade fees). On the other side we have: The Open/Free Model of "Programming as a service" As I mentioned before, the beginnings of this model is already in effect, although it is not generally understood as such. Corporations paying for development of Free/Open Software, in effect provide a service to the users of their products. In other words corporations roll in the costs associated with their Open/Free Software development into the cost of their for sale products, the development of free software becomes part of their product. This is true for the IBM's of the world writing new file-systems for Linux. But it is also true every time a business using Open/Free software fixes bug or adds feature (scratching the businesses own personal itch if you will, also fixes that same itch for users of that GPL code.) We are also beginning to see professional Open/Free software services by companies like RedHat and IBM. In this model a company (lets say Boeing Corp) needs specific functionality in the Linux Platform, which currently does not exist. Boeing and IBM agree to the terms of a software development contract, IBM develops this functionality for Boeing under the GPL. This is what I call "Programming as a service". In effect its as simple as you hiring a roofer to repair the roof on your home: One party pays another party to provide services. Where this model differs from a standard building contract is that once the software is developed it becomes a reusable resource by everyone, not just those who who paid for it. The GPL allows Boeing not only to leverage existing code, but allows them to create extensions to meet their personal needs, while hopefully having their extensions enhanced by outside parties, further benefitting Boeing needs. What is so different about this model is that the cost of a feature usable by everyone is simply equal to [developer skill + time spent]. Which one ups your roofer's [labor skill + time spent + materials] The more distant future: I firmly believe that Professional Open/Free Software services will be expanded to include end users (individuals). In this model individuals (society in general) fund the Software roofers of the world to build and extend The Platform. In this future you will have your large and small software services companies (everything from Bob's Software services to IBM's Global software services) just like you have Bob's roofing and North West Roofers. In the end I don't believe software is all that unique. Only two things in my opinion created the situation we currently find ourselves in. A) The software tycoons were able to convince the masses that software/technology was spooky and a kin to a physical device (requires one license per copy) B) The cost to develop and distribute software is [Skill + Time spent] making it unprofitable to charge for materials. Thanks for your time. Shane. --- Date: Mon, 23 Sep 2002 08:50:59 -0600 From: Charles Curley <ccurleyat_private> To: Richard Morrison <rmorrisonat_private>, Declan McCullagh <declanat_private>, politechat_private, jdelongat_private Subject: Re: Fwd: FC: CEI blasts Linux as unsuitable for government, business use First, a bit about me. I have 23 years experience as a software engineer, and have worked on projects from SQL databases to embedded real-time products. I have worked with Linux for eight years, and for Microsoft for some years as well. In addition I have released software under the GPL and contributed to other GPLled software. On Thu, Sep 19, 2002 at 06:11:27PM -0700, someone wrote: > > > ---------- Forwarded Message ---------- > > Subject: FC: CEI blasts Linux as unsuitable for government, business use > Date: Thu, 19 Sep 2002 14:36:03 -0700 > From: Declan McCullagh <declanat_private> > To: politechat_private > Cc: rmorrisonat_private, jdelongat_private > > CEI is a free-market think tank in Washington. It may be best known for its > work on environmental issues, where it has pointed out government reliance > on junk science. > > Jim Delong, who wrote the below message, is an occasional contributor to > Politech. Jim wrote a piece over the summer supporting the general approach > of the Berman anti-P2P bill: > http://www.politechbot.com/p-03711.html > > CEI received a small-to-moderate amount of money from Microsoft during the > antitrust trial days, but based on my knowledge of CEI, I'd say I'm > positive that Jim's views are his and his alone (he's an economist, and it > can't be helped). :) Declan, I'd be careful of this last. Mr. Delong refers to the GPL below as "viral", a choice bit of newspeak invented (so far as I know) by Microsoft. Of course, if the metaphor refers to computer viruses rather than to the biological types, then it is appropriate that Microsoft should have invented it. > > Jim is responding to this New York Times editorial yesterday: > http://www.nytimes.com/2002/09/18/opinion/18WED2.html > > -Declan > > --- > > Subject: CEI's Weekly Commentary: Software Wars > Date: Thu, 19 Sep 2002 17:13:37 -0400 > From: "Richard Morrison" <rmorrisonat_private> > > CEI C:\SPIN "C:\spin"? Are we Microsoft-centric here or what? > > This issue: Software Wars: Open Source and the N. Y. Times > > > This week's c:\spin, is by James V. DeLong, Senior Fellow, Project on > Technology and Innovation, CEI, September 19July 19, 2002. > > > The New York Times recently editorialized about Linux and open source > software, exuberant that an operating system written and updated by > volunteer programmers in a communitarian spirit, and available for free > might challenge Microsoft s Windows and result in major savings in computer > costs. > > > The paper also exulted that governments such as Germany and China are > pushing Linux, and it urged everyone, including U.S. agencies, to join them > so as to foster competition. It should be noted that US government agencies such as Ames Research Center, Lawrence Livermore and Los Alamos National Labs have not only adopted Linux and other GPLled software, but have contributed to it. For example, the NSA has its own distribution of Linux. > > > The NYT view has some gold. Competition is always good. And the > Linux backers have hold of an important truth, which is that > persuading a lot of smart people each to devote a small part of > their time to an effort can produce impressive results. This totally misses the reason people contribute to free software. Some people write open source or free software out of ideological conviction (Richard Stallman, author of the GPL, for one). Others for the fun of it. Most write it for very selfish reasons. They want to solve a problem, and the easiest way for them to do that is to modify an existing program to do that. If the modified code is in-house code, there are no licensing issues. If the code code is open source, they can legally do this. Many of those people then contribute back to the pool of common software in the hope that it will be useful to someone else. But the reason they wrote it was to solve their own problem. Stallman wrote Emacs because he wanted a powerful editor. He got one! Torvalds wrote Linux so he could study a real Unix kernel without having to pay massive amounts of money to a Unix vendor. (And, I suspect, because he wasn't enamored of Andrew Tannenbaum.) > They are also right to think that opening up computer code to the > eyes of the whole programming community can be extremely productive. > Microsoft itself sees increasing virtue in this idea, and is > developing shared source to open up code to scrutiny while the > company keeps firm hold of the pen. > > > But the NYT misses in some ways. First, none of this is free. Software > is a complicated industrial product requiring continuing re-creation and > support, and money to support it must come from somewhere. Linux > programmers are not street people who sleep on steam grates so as to > indulge their passion. They are supported, often handsomely, by > universities and IT companies. Even this support is not sufficient to keep > Linux going, and hardware companies, notably IBM, are now pouring billions > into it. There is nothing wrong with this; IBM has good competitive > reasons in that it wants to dish Sun and Microsoft. But the movement is > not the folk song army depicted in the NYT. > > > If IT companies, universities, and IBM want to donate the fruits of their > labor to computer purchasers, including governments, that is their > privilege. But we have just gone through a half a decade in which the > business model was give it away, and it did not work. In the end, software > might be bundled with hardware, or vendors might give away software tied to > a services contract both are increasingly common -- but the code writers > will want pay for producing it, which means money must ultimately come from > the users somehow. Again, you miss the point of why people contribute to open source software. Believe me, neither IBM nor Sun is in the least altruistic. Each of them would like to be where Microsoft is, and some IBM executives are no doubt kicking themselves because they aren't. > > > A second problem is the creation of applications for Linux. The General > Public License that controls the program s distribution can be paraphrased > as thou shalt not charge for this program and its source code shall be > public. Wrong, flat out wrong. Paragraph 2 of the GPL Version 2 states: "Our General Public Licenses are designed to make sure that you have the freedom to distribute copies of free software (and charge for this service if you wish),..." Otherwise, for example, Red Hat would not be charging over $1,000 for their Advanced Server product. > This license is also viral; if you write an ap for Linux, and > incorporate any code covered by the GPL, then your ap is also > subject to the GPL, and it too becomes open source and free. And what objection do you have to this? It is the right of the author to license his/her/its software under any terms the author sees fit, or not at all. Or don't you believe in intellectual property rights? However, this too misstates the facts. First, the Free Software Foundation also has a license for libraries which does not require that programs which call those libraries be released under the GPL. Otherwise Corel, IBM, Oracle, Sun and other companies would not be releasing products for Linux. Most of the core libraries available on Linux uses this library license. Second, the GPL does not always require that the author of a derivative program GPL his/her/its own work. See the fifth paragraph of Section 2. Oh, just to set the record straight, the GPL applies (or not) regardless of the OS on which the application runs (or none at all). This includes Windows. > > > True open source believers think that this is just fine -- all aps should > be open and free. But it is not clear that the freeware spirit, or the > IT/university willingness to subsidize, runs deep enough to provide > anything approaching the number of aps available for Windows, where good > old reliable greed creates an incentive for developers. The Linux > community is moving toward proprietary aps, but it is chancy. Writing aps > without incorporating some operating system code is difficult, and those > who want to engraft proprietary aps onto Linux are taking a legal > risk. This is pure FUD. Plenty of companies with lots of lawyers have written proprietary apps for Linux, with no fear of their application all of a sudden being declared open source. The Linux community is not "moving toward proprietary aps". It is embracing them whole heartedly. Vendors would not be coming back to Linux with new or upgraded if they weren't welcome. I have personally favorably reviewed a number of proprietary applications for Linux Journal. I take it that software engineering is not taught at Harvard Law. DeLong says, "Writing aps without incorporating some operating system code is difficult,..." If this statement were true, very few people would be writing code for Microsoft Windows. In fact, it is quite easy to write applications without using operating system code, and a good layered design requires that one do so. > > > Finally, governments should not treat this as an arena for industrial > policy. The incentives fueling the Linux movement are not necessarily > those required for long-term production of software suited for the public > as well as the nerds. Governments, which are as naïve as editorial > writers, should keep their hands off. I have to compliment the author for the unintended irony of that last sentence. > > > ------------------------------------------------------------------------- > POLITECH -- Declan McCullagh's politics and technology mailing list > You may redistribute this message freely if you include this notice. > To subscribe to Politech: http://www.politechbot.com/info/subscribe.html > This message is archived at http://www.politechbot.com/ > Declan McCullagh's photographs are at http://www.mccullagh.org/ > ------------------------------------------------------------------------- > Like Politech? Make a donation here: http://www.politechbot.com/donate/ > Recent CNET News.com articles: http://news.search.com/search?q\clan > CNET Radio 9:40 am ET weekdays: http://cnet.com/broadband/0-7227152.html > ------------------------------------------------------------------------- > > ------------------------------------------------------- -- Charles Curley /"\ ASCII Ribbon Campaign Looking for fine software \ / Respect for open standards and/or writing? X No HTML/RTF in email http://w3.trib.com/~ccurley / \ No M$ Word docs in email ------------------------------------------------------------------------- POLITECH -- Declan McCullagh's politics and technology mailing list You may redistribute this message freely if you include this notice. To subscribe to Politech: http://www.politechbot.com/info/subscribe.html This message is archived at http://www.politechbot.com/ Declan McCullagh's photographs are at http://www.mccullagh.org/ ------------------------------------------------------------------------- Like Politech? Make a donation here: http://www.politechbot.com/donate/ Recent CNET News.com articles: http://news.search.com/search?q\clan CNET Radio 9:40 am ET weekdays: http://cnet.com/broadband/0-7227152.html -------------------------------------------------------------------------
This archive was generated by hypermail 2b30 : Wed Sep 25 2002 - 23:21:35 PDT