Sections: Main page Linux in the news Security Kernel Distributions Ports Development Commerce Announcements Back page All in one big page See also: last week's Back page page. |
Linux links of the weekOpen source and open data. This page advocates standardizing the file formats used for many popular applications, such as word processors, spreadsheets, and even databases. Your editor has seen many efforts at standardizing data formats over the years; it is not an easy task. But the rewards can be great if a format is successful. See, for example, the NetCDF format, which has brought sanity to some types of scientific data. It's a prize worth trying for. The Silicon Valley Linux Users Group gets to have an unfair amount of fun, and their web page reflects it. See, for example, the reports from their recent demonstration at the opening of the Microsoft Silicon Valley campus. "The Microsoft people thanked us afterward and bought a round of beers and sodas for everyone keeping themselves under control." Those wanting a free software news site with more of a Microsoft-centric orientation may want to check out the Why you shouldn't use Microsoft products site. |
November 19, 1998 |
|
Letters to the editorLetters to the editor should be sent to editor@lwn.net. Preference will be given to letters which are short, to the point, and well written. If you want your email address "anti-spammed" in some way please be sure to let us know. We do not have a policy against anonymous letters, but we will be reluctant to include them. |
From: "Will Tsui" <will@netpedia.com> To: <editor@lwn.net> Subject: Web Hosting for Open Source Projects Date: Sat, 14 Nov 1998 23:39:20 -0500 Greetings, My name is Will Tsui and I am in charge of a site called Netpedia (http://www.netpedia.com/). We would like to invite anyone developing open source software to get a complementary subdomain on netpedia.net. This is a "no catch" deal. We don't require links, advertisements, etc. Here developers can host their websites and keep the web community informed about their contributions. If you are interested in this offer to all open source developers, e-mail me for further info. Thanks for your time, Will | ||
From: "Ricardo Galli" <gallir@atlas-iap.es> To: <lwn@lwn.net> Subject: Linux and databases Date: Thu, 12 Nov 1998 16:04:32 +0100 Dear LWN editor, I send you this letter because I see a lot of press articles about = Oracle and Informix coming to Linux community, but nobody mentions that = there are other companies supporting and providing very good comercial = application for Linux OS. One of these companies is Empress Inc. (www.empress.com) that developed = a very fast and reliable RDBMS altoghether with a complete development = toolkit. We were using it from middle 1996, when I bought our first Empress suite = for Linux. Since then we are very happy with the product and it = overpasses in performance other RTDBMS servers. Our applications span from dynamic web pages (many, many, = www.atlas-iap.es/kuhn/ just as one example), digital newspapers = (www.diaridebalears.com) to an AAA server (programmed with embedded SQL = in C programs) where our Radius and Tacacs+ servers act as clients (we = provide dial-up Internet access services) of the AAA server.=20 Should be note that both, Linux boxes and Empress servers run very = stable with almost no maintenance at all. Our first Linux/Empress server = (an old Pentium 133) is still working and need 0 (yes, ZERO, we just = have to delete the long apache logs twice a year) maintenance and is = serving more than 100.000 queries a day.=20 Furthermore, our own AAA server (see some stats at the end), which was = developed in C using Empress development toolkit never had a problem at = all, it works on a database with 15 tables, with the bigger one (access = logs) having more then 2.000.000 records. Please note that we do not have any commercial relationship with Empress = Inc. (we pay all of our licences at standard prices). I just thought = this a good story and that it's very worth to take a look to this = company that is supporting Linux providing a very good RDMS and a = complete development toolkit from the very beginning. I feel also that = they are ignored in Linux press/media. STATS from the AAA server (running on a Pentium II 300 MHz). gallir@star:/home/people/gallir > dbstats=20 DBSERVER(862): CPU load (0-1): 0.001434 Sec/ops: 0.0304 PID: 10451 Elapsed: 85 days 22:58:08 Ops: 350789 Childr: 1 Max: 5 Overlds: 0 ** Parent Times User: 00:00:08.69 Sys: 00:03:44.83 Total: 00:03:53.52 ** Total Times User: 01:39:26.67 Sys: 01:18:00.23 Total: 02:57:26.90 Congratulation for LWN, specially for daily updates section. It's the = best source of up-to-minute Linux information. Best regards, --ricardo galli University of Balearic Islands Atlas IAP S.L., Internet Service Provider | ||
To: editor@lwn.net Subject: reply to Kastrup From: Nathan Myers <ncm@cantrip.org> Date: Fri, 13 Nov 1998 20:08:51 -0800 [I don't want to start a flame war, but David Kastrup's long, prominently placed, but badly misinformed letter demands a response.] David Kastrup predicts doom ("Microsoft Linux") and denigrates GNU g++. While his opinions are easily ignored, the false history he has promoted needs to be countered. He wrote: C++ has hobbled free compiler development: the incredibly complicated language definition has caused gcc development to freeze. The FSF's non-commercial development infrastructure of the gcc compiler for the comparatively simple C language could not keep up with the complications of the C++ language. This has resulted in the splitoff of egcs, mostly managed by Cygnus, a commercial entity and large-time contributor. In fact, C++ has nothing to do with the development of the Gcc C compiler. (They share only a back-end code generator.) The slowdown in development of Gcc had nothing to do with C++ -- blame administrative and personal problems at the FSF -- but was anyway solved in exemplary fashion by the formation of the Egcs group. Egcs is not "mostly managed by Cygnus"; on its steering committee of 13 members, I count four Cygnus employees. While the Cygnus employees are supremely competent, they are far from alone in their contributions, as may easily be seen by a glance at the Egcs web site, even though many large contributions have not been acknowledged by name there. Gcc, as delivered in the Egcs package, is as advanced as any C++ compiler available. Its incredibly rapid development since the Egcs group provided a new rallying point stands as a shining example of what free software can accomplish. Make no mistake: this was not a case of a corporation taking over development from a failed free software project. It was a classic case of users taking the source, and the development process, into their own hands when the previous maintainer failed to keep up. The Egcs group has more than equal to the "incredibly complicated" C++ language. Nathan Myers ncm@cantrip.org | ||
To: esr@thyrsus.com From: "Khimenko Victor" <khim@sch57.msk.ru> Date: Wed, 18 Nov 1998 20:49:14 +0300 (MSK) Subject: How it's possible ? RMS forgot OSS definition or what ? "After reviewing the QPL, I find that it meets all the criteria to be considered Open Source," said Eric Raymond, author of The Cathedral and the Bazaar paper and joint copyright holder of the Open Source trademark. "This license should allow Open Source software to compete on features and polish rather than ideological position, which will be better for the Open Source community." How it's possible ???? QPL is NOT compatible with Open Source Definition ! "Caldera Systems is very excited about Troll Tech's decision to publish Qt under an Open Source license," said Ransom Love, President and CEO of Caldera Systems. "Customer response and feedback to the KDE desktop environment included in our OpenLinux 1.3 release has been overwhelming. Offering Qt under an Open Source license will allow KDE to stand on its technical merits without causing undue concerns to the development community over licensing terms." "SuSE congratulates Troll Tech for their wise decision to put Qt under an Open Source License. SuSE thinks that this step will (hopefully) unite the Linux community again," said Burchard Steinbild, Managing Director of S.u.S.E. GmbH. "SuSE wishes Troll Tech much success with their products and hopes their reputation in the Linux community rises after this move, they have deserved it." What's this ??? Are you are all blind or what ? 1) QPL is not OSD-compliant 2) Even if QPL will be fixed to be OSD-compliant this will not change even jot for KDE ! To resolve KDE copyright problem QPL should be GPL-compliant, not OSD-compliant ! The Open Source Definition (version 1.0) http://www.opensource.org/osd.html : -- cut -- 3. Derived Works The license must allow modifications and derived works, and must allow them to be distributed under the same terms as the license of the original software. -- cut -- THE ONLY WAY TO DISTRIBUTE "DERIVED WORK" IN QPL IS FOLLOWING (from http://www.troll.no/qpl/): -- cut -- 3. You may make modifications to the Software. In order to preserve the integrity of the unmodified version of the Software, modifications must be distributed in the form of patches, and the following restrictions apply to each patch: a. Application of the patch must not modify copyright notices in the Software. b. The patch must be explicitly licensed by the following clauses without additional restriction: Permission is hereby granted, free of charge, to any person obtaining a copy of this patch, to deal in the patch without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the patch, subject to the following conditions: Any copyright notice and this permission notice must be included in all copies or substantial portions of the patch. c. The patch must include an accurate description of the modification, the date of the modification and the author of the modification. -- cut -- Clearly not OSD-compatible ! For KDE is more important that QPL is not compatible with GPL (and GPL compatibility is MUCH more restrictive then Open Source Definition; NPL & MPL are OSD-compatible but not GPL-compatible !!!): http://www.gnu.org/copyleft/gpl.html -- cut -- 3. You may copy and distribute the Program (or a work based on it, under Section 2) in object code or executable form under the terms of Sections 1 and 2 above provided that you also do one of the following: a) Accompany it with the complete corresponding machine-readable source code, which must be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange; or, b) Accompany it with a written offer, valid for at least three years, to give any third party, for a charge no more than your cost of physically performing source distribution, a complete machine-readable copy of the corresponding source code, to be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange; or, c) Accompany it with the information you received as to the offer to distribute corresponding source code. (This alternative is allowed only for noncommercial distribution and only if you received the program in object code or executable form with such an offer, in accord with Subsection b above.) The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable. However, as a special exception, the source code distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable. -- cut -- What's this ? You all want KDE problem to be resolved so much that you could not understood clear English ? Or may be I'm misunderstood something (Enlish is not my native language after all)... | ||
Date: Thu, 12 Nov 1998 17:04:00 -0500 From: "Zygo Blaxell" <Zygo.Blaxell.zblaxell@nt.com> To: editor@lwn.net Subject: Balkanization of Linux HOWTO [Unlimited unmodified attributed distribution permitted] Now that Linux is starting to turn the heads of major business entities, I keep seeing other Linux people asking the question "What should we do next?" or "What should we do to stop the new Microsoft threat?" My answer is simple, direct, and to the point: "Do exactly the same thing we've always been doing." For those who joined the party late, this is what we've been doing: 1. Producing the best software. 2. Supporting and defining open standards. 3. Creating replacements for existing closed-source software without infringing on protected IP rights. 4. Sharing information as widely as possible, without hesitation, in hope that it will find the right person to act upon it. In http://www.techweb.com/se/directlink.cgi?INW19981109S0020 Michael Dell explicitly states that Microsoft pays Dell to pre-install Internet Explorer and Netscape doesn't, so Dell ships IE by default. "Free" software can't compete with that kind of business logic, nor should it try. We can expect the last of the large closed-source vendors to make desparate attempts to gain market share, including such desparate tactics as paying customers to introduce dependencies on closed-source products into their critical business systems. Our response to that should be to consistently produce better products until the closed-source vendors run out of new ideas or money. In the recent past, open standards have had published specifications but typically only proprietary implementations have been available. Even specifications that include sample source code are still not entirely open, because you can't use the code without a closed-source OS, libraries, or compiler tools. Now that we have complete open-source operating systems, and therefore pure open-source all the way from the API layer to the bare hardware, it is no longer necessary to have standards "tainted" by the lack of at least one fully open-source implementation. We should advocate that the term "open standard" means that not only is the specificiation published, but at least one completely open-source implementation (right down to the bare hardware for software, or right up to the API layer for hardware) is available. Given the past explosive growth and continuing persistence of legacy closed-source software, it follows that many good ideas have previously been implemented in closed source. Linux is a re-implementation of Unix, WINE is a re-implementation of Win32, and so on. This will be necessary as long as there exists closed-source software that does not have an open-source equivalent. We should educate vendors of closed-source software about their golden opportunity right now to re-release their products as open source before competitive open-source implementations of those products exist. We should follow that up by aggressively implementing legally-clean, high-quality replacements for significant closed-source software from vendors that do not co-operate. The message is, was, and always will be "release source or die," but today the message is perhaps a little louder than it was yesterday. It is good that people with solid marketing, sales, and PR expertise are entering the Linux community. Linux doesn't just need software developers: it needs technical writers, venture capitalists, lawyers, advertisers, pure researchers, teachers, guinea pigs, testers, journalists, managers, and all the "other people" that a big software company would have access to. Some of these people are following the open-source tradition in fields very distant from software: for example, last week I read a good introductory document with some practical tips on doing one-on-one marketing of Linux and another document that explains how to write a press release. We still need to find good ways that people with work to be done can be connected with skilled people with time to spare, and we cannot possibly be too good at bringing lots of "newbies" up to speed quickly and painlessly. We should NOT be: 1. Trying to control or manage open-source development. 2. Going out of our way to support legacy closed-source systems. 3. Trying to rigidly standardize everything, or produce a "one true Linux" in any form. Open source software in general and Linux in particular work as well as they do because every developer has a vested interest--whether financial, practical or purely social--in the end product. Enforcing specific external goals will destroy a productive open-source development group, or at least limit its capabilities to those of a closed-source development group. People should always do their best work first. If some particular job needs to be done, then those who benefit most from that job should put together the resources to have it done, and those who are good at a particular job should be allowed to do it. Widespread support of open standards means that support for closed-source systems before they are replaced is merely an unnecessary distraction. Many closed-source systems do not offer any opportunity for continued development once the original system has been duplicated, as the existing installations cannot afford any modifications or extensions to existing functionality. In the extreme, closed-source systems are supported via emulation of the hardware that they run on, creating a burden for user and developer alike while providing no benefits to either at all. The best way to support a legacy closed-source system is to produce an open-source implementation of that system or a convenient migration path to a replacement open-source system. This liberates the closed-source vendor's customers by breaking the vendor's monopoly on licensing, support and distribution services, and allows customers to enjoy reduced costs, better product, and better support, all at the same time. Note that Linux has already successfully replaced closed-source Unix implementations for many purposes and is now far ahead in some areas. The computing industry in general--open and closed-source vendors included--is still decades or even centuries away from achieving a "one true" anything. After 40 years we still don't have "one true character set" (ASCII, ISO-8859-*, Unicode...), so why should we expect a large API like POSIX, Java, or Win32 to become a permanent standard overnight? The best software and hardware available today is still very limited, and if history is any kind of guide to the future, it will all be obsolete tomorrow. Trying to evangelize "one true"ness just imposes restrictions on people with good ideas, destroys the flexibility that prevents obsolescence, and pre-empts the benefits of research and development before they even start. Even Microsoft, arguably the most vocal proponent of "one true"ness, is beginning to feel trapped by this goal as their costs skyrocket while growth declines. Some Linux users will need rigid, supportable stability, and some Linux vendors should provide it to them; however, at the same time nothing should get in the way of those who want early access to the latest in research and development. | ||
|