Sections: Main page Linux in the news Security Kernel Distributions Development Commerce Announcements Back page All in one big page See also: last week's Back page page. |
Linux links of the weekFolks on the move may want to check out the Dynamics HUT mobile IP project. They have put together a set of software facilitating connectivity to mobile systems via a number of media. C|Net's Download.com has set up a Linux area. They offer downloads of various distributions as well as other interesting packages. (Thanks to Benji Selano). Section Editor: Jon Corbet |
June 24, 1999 |
|
Letters to the editorLetters to the editor should be sent to letters@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. | |
To: letters@lwn.net Subject: antivirus software for linux... Date: Wed, 23 Jun 1999 09:07:48 +0100 From: kevin lyda <kevin@suberic.net> a recent poll in comp. reseller news said that 26% of vars cited a lack of antivirus software was hurting linux. ok. here's version 1.0 of antivirusd for linux: /* antivirusd.c - virus checker for linux. copyright kevin lyda * * licensed under the gpl. see some link at www.gnu.org */ #include <stdio.h> #include <unistd.h> #include <limits.h> int main(int argc, char *argv[]) { close(0); close(1); close(2); chdir("/"); if (fork()) { exit(0); } for (;;) { sleep(INT_MAX); } } kevin | ||
Date: Thu, 17 Jun 1999 20:57:22 +1000 (EST) From: Conrad Sanderson <conrad@hive.me.gu.edu.au> To: editor@lwn.net Subject: Mindcraft Times Three Microsoft The first Mindcraft report was muddled in execution, and we were lucky that there was so much negative press about it, mostly for a good reason. But it also showed that Apache and Linux both have performance weaknesses which need to be addressed. In effect we got the benefit from the benchmark without the bad publicity. Microsoft is pissed off because of this. Mindcraft wants this as well, because their reputation got hurt badly the first time around. It is fairly evident that Mindcraft (or should I say Microsoft) wants to force the Linux Community's hand into a benchmark, from where Microsoft will make Many Press Releases (tm) about NT vs Linux. Using their standard marketing and FUD tactics they will take lots of liberty in the interpretation of the results and ignore others - eg. non-SMP performance of Linux and NT, where Linux wins right now. MS smells blood and is willing to follow this Mindcraft benchmark up to the end. In version 3 of this benchmark, all the previous "publicity" and PR bugs have been fixed - we now have a involvement of Linux people (two from Red Hat and one from Penguin Computing), and the place of testing is apparently independent. We know that we will lose this benchmark, so why on earth did Red Hat get involved ??? We could have refused participation until the kernel and the web server had performance enhancements. Refusing participation is nowhere near as bad as hard benchmark data, which is going to stick around for years. Microsoft can and will use all the mileage it can get out of it, and then some. It is our right to do refuse participation until we are ready - after all, one of the main strengths of OSS is that stuff isn't released until it's ready. But instead, we are playing directly into Microsoft's hands. Related sites for performance enhancements in Linux: kernel based web server: http://www.fenrus.demon.nl/ Mindcraft Redux: http://www.kegel.com/mindcraft_redux.html -- Conrad Sanderson - Microelectronic Signal Processing Laboratory Griffith University, Queensland, Australia http://hive.me.gu.edu.au/ | ||
Date: Fri, 18 Jun 1999 11:52:03 +0200 (MDT) From: Maurizio de Cecco <Maurizio.de.Cecco@ircam.fr> To: letters@lwn.net Subject: Trademarks ? Talking about "Open Source" or "OSI Certified" trademark, i think you this time miss completely the point. The power behind the "Open Source" term never came from being a trademark (that never was); the power came from having a community reconizing the value of the concepts behind the term, and actively controlling its usage. Having a single organization, any single organization, certifing the membership of a project to the open source/free software community is a non sense, exactly because we are speacking of a community, not a closed club, with undefined borders; the only thing the community can gain from this is an infinite series of conflicts and division. Maurizio | ||
Date: Fri, 18 Jun 1999 07:23:33 +0100 (BST) From: Richard Corfield <Richard-Web@littondale.freeserve.co.uk> To: Alessandro Muzzetta <Muzzetta@geocities.com> Subject: Re: Your letter on Linux Weekly News about patents Lack of awareness, at least of the issues, does seem a problem. I read the computing section of a newspaper over here and they appear to be afraid to mention anything more complex than the latest MS Word features. In coverage of the Explore.Zip virus for example its has been stated that it erases "Word and Excel files and some others" as if mentioning of the words "source code" would be too much. Whenever the words "Operating System" are mentioned an explanation is generally given as to what an operating system is so explaining Unix, even though it still runs most of the 'net, may just go too far. Given this, coverage of Linux in the mainstream press is limited or not at all. It seemed strange to me to see an article about competition to Microsoft that only talked about the Mac. A non technical friend of mine thought that the book "Open Sources" looked "heavy going". If you want to get the issues of software patents in Europe widely known you'd be best getting into the mainstream press. The hurdle would seem to be justifying to them the relevance of open source to their audiences and the relevance of patents to open source - sufficiently to be worth having to explain the concepts to their target audiences. Otherwise it would just be like the many other EU rules that come out and are ignored. Maybe they should read "Open Sources" first. - Richard. _/_/_/ _/_/_/ _/_/_/ Richard Corfield _/ _/ _/ _/ Web Page: http://www.littondale.freeserve.co.uk _/_/ _/ _/ Dance (Ballroom, RnR), Hiking, SJA, Linux, ... [ENfP] _/ _/ _/_/ _/_/_/ PGP2.6 Key ID: 0x0FB084B1 PGP5 Key ID: 0xFA139DA7 | ||
Date: Sat, 19 Jun 1999 12:12:44 +0100 To: letters@lwn.net From: dps@io.stargate.co.uk Subject: 2.3.6 misinformation Claiming that 2.3.6 features the boredom of code that works is misinfomation, at least given some libraries. It features a nasty mmap bug (at least) that produces problems like the inability to load dynamic libraries even with arbitary amounts of free memory. I was hoping a 2.2.x and 2.3.x bug that seems to stuff processes that seg fault into permanent disc wait status (exploits include the current gcc-2.95 pre-release and at least one of the things in the testsuite). I was hoping 2.3.7 might be a version without this particular feature... and the (locking methinks) permanent disk wait status bug). I am not the only one affected by the mm/mmap.c bug either. Getting a relaibel boot from 2.3.6 still eludes me after I fixed that particular bug. I suppose 2.3.6 does feature the boredom of code that does not trash your file system, prefering merely to hit lethal problems it the boot process. I was proposing to take a look at changing the sheduling code to use a heap for faster reschedules. Duncan (-: | ||
To: jp@ncfocus.com From: sharkey@superk.physics.sunysb.edu Subject: Re: Look Before You Leap Into Linux Adoption Date: Sat, 19 Jun 1999 02:23:48 +0900 Mr. Morgenthal, This message is regarding your article at: http://www.internetwk.com/columns/logic061499.htm With a title like "Look Before You Leap Into Linux Adoption" I was expecting a well written article debating the advantages and disadvantages of adopting Linux. Linux has many pitfalls associated with it and I expected you to mention a few of them. However, I was greatly disappointed to read your article and find criticisms which ranged from misleading to just plain wrong. Let's go through some of the points you bring up. > Linux is an open-source project; therefore, all changes to the > kernel are subject to review and approval by a small team that > controls this portion of the operating system. True, but misleading. Any company may make any changes it wishes, it may distribute those changes, it may even call their modified version "Linux". Only those copies of the code which are distributed by Linus Torvalds need to have code approved by Linus Torvalds. Every user is on equal footing. If Toshiba Corporation wants to distribute a custom Toshiba Linux, they are free to do so. (Of course, I am using Toshiba as a generic example. It works equally well with any company.) > Companies that add features they need, but that are not accepted > into the core distribution, may find themselves in a redevelopment > and retesting cycle every time a new version of Linux is released. This would only happen if said company wishes to included new developments of Linus Torvald's Linux into their own version. If their code is working well without those developments, such testing is not needed. There is no need to upgrade, just for the sake of upgrading. But more to the point, how is this any better than other operating systems, where the kernels are closed? Your statement in this paragraph amounts to: "If you change the kernel, you may need to maintain those changes." Is an operating system where you are prohibited from making the choice of whether or not to modify the kernel really better? Finally, this core group of programmers you refer to (which ultimately means just Linus Torvalds himself), is generally considered to be a rather intelligent group of people. When a feature is rejected, it is done so with explicit reasoning which is publicly announced. If a company's modification is truly a good thing, then it will be accepted. Can you provide a single example of a company which has been put in the position of maintaining a Linux kernel fork? > Windows supporters still outnumber Linux supporters because > Microsoft provides a better value proposition. Perhaps, for some individuals, this is true. Linux still lags behind in multimedia and entertainment software, and for the home market, where such applications are essential, Linux may not provide the best value proposition for all users. Given this, I was very surprised to read your supporting arguments to this statement: > Windows NT Server Enterprise Edition ships with a full complement of > Internet services, including Web, proxy, index, messaging, database, > transaction and firewall services. With Linux, these services will > soon be available as a multivendor product. The first statement is true, you can get all that from Microsoft (for a *very* hefty price tag for unlimited clients), but such services are also included with any fully fledged Linux distribution. For example, Debian Linux 2.1 (released in March '99) includes: Web: Apache, Boa, Cern, dhttpd, and Roxen Challenger Proxy: socks (generic ip proxy), squid, and transproxy (www), plus other specialized applications for particular protocols index: swish++, and others messaging: Zephyr, and many others database: msql, mysql, postgresql, others firewall: support built-in to kernel Everything you need is included right there, from a single source, all in one box, with no hefty licensing fees. Now, it's true that these various pieces of software were developed by different individuals, but the same is true of the Microsoft code. There are different departments inside Microsoft corporation devoted to these pieces of software. They are not coded by a single individual. Furthermore, this is not a new development. Linux has been a strong performer in the internet server market for several years now. The assertion that these services will "soon be available" is laughable. > Linux is just beginning to be retrofitted for symmetric > multiprocessing. Now, this is just plain wrong. The first officially released SMP code I know of was included with Linux 1.3.26, released September 13, 1995, almost four years ago. At that time, Linux was "just beginning to be retrofitted for symmetric multiprocessing". In June of 1996, over three years ago, Linux 2.0.0 was released which contained full, stable support for SMP systems. The release of 2.2.0 at the beginning of this year provided significant SMP performance and reliability improvements. I purchased four dual processor Pentium Pro servers in December of 1997 and they have been running Linux quite satisfactorily since day one. ale% uptime 1:18pm up 124 days, 36 min, 13 users, load average: 1.95, 1.91, 1.84 I can live with that. :) > Without robust SMP, Linux servers can support only small companies > and single applications. Again, I can't seem to make any sense out of this statement whatsoever. Very few companies *need* the kind of SMP support provided by NT or Linux. Having two or four processors in a box is nice, but it's not going to give you the edge that will make or break your business. If your business really needs heavy parallel computing, then you'll buy some 32 or 64 processor system from Sun and run Solaris, and do the job right, but for most companies, that's overkill. Using cheap PC hardware, today's dual processor machine will be outperformed by tomorrow's uniprocessor system under any OS. This statement also seems to imply that without SMP support, an operating system is incapable of multi-tasking. That can't be more wrong. Linux has been multitasking since it's inception in 1991. > If you're managing multiple servers for increased scalability, > you're better off using multiple NT servers all participating within > the same domain. Are you not familiar with Cplant? (Currently 129 on the top 500 list): http://www.cs.sandia.gov/cplant/ Or Avalon? (160): http://cnls.lanl.gov/avalon/ These testaments to the scalability of Linux speak for themselves. I don't see the need to belabor this point any further. > Also, remember that Linux is still Unix. One of the reasons for > Windows' growth has been the complexity of configuring and > maintaining Unix operating systems. Simply because a low-cost > version of Unix is now available, it does not automatically generate > more people capable of managing and configuring these systems. This is your best point so far, but I think the ease-of-use issue has been overstated by the press. At the current time, Linux is still difficult for beginners to use, but when you think about it, this is true of all operating system. If you're used to one operating system, (or not used to any) it takes some time to learn another. But seriously, I think it's important to keep in mind that many high school students can manage to install and run Linux on their home computers just fine. How hard can it be for an experienced computer professional? If you're in the position to be worrying about databases, firewalls, proxy servers, and distributed processing, then you better have enough basic knowledge of computing to be able to flip through Linux for Dummies and have a server up and running in 24 hours. It's just not that hard. I find the continued portrayal of Unix as an impossibly difficult system to master as condescending the "Math is hard." Barbie doll. It's a self- fulfilling prophecy. > Linux is a college student's project gone astray. This is just a low jab, with no meaningful content. Lots of projects began as something small and grew to something much larger. Remember Netscape? Hell, the whole internet itself grew out of the original ARPANET project to link three universities, and SRI. > The version that will be supported by Sun Microsystems and IBM on > its hardware will fall far short of each of these company's own Unix > operating systems in features and capabilities. In general, the feature sets of any two operating systems will not overlap completely. One will almost always contain a feature the other lacks. So, I'm sure you could find a case where you could say "Solaris can do THIS, while Linux cannot.", but as a regular user of both Solaris on Sun hardware and Linux on Intel, I'll pick Linux on Intel 9 times out of 10. It suits my needs better. One striking difference between my Solaris boxes and my Linux boxes is the contents of /usr/local. My Solaris boxes all have /usr/local full to the brim with GNU and other third party utilities which make the operating system easier to use. GNU utilities tend to have more options, features, and command line switches than their commercial counterparts. Under Linux, /usr/local is almost barren. The operating system comes with almost everything I need. In summary, I feel you had a valid point to make. We should all take a look before we leap. I just have a hard time figuring out what it was you were looking at when you wrote that article. Eric Sharkey sharkey@superk.physics.sunysb.edu | ||
Date: Mon, 21 Jun 1999 11:40:00 -0400 From: Anand Srivastava <anand@nmi.stpn.soft.net> To: metcalfe@idg.net, letters@lwn.net Subject: Linux's '60s technology, open-sores ideology won't beat W2K, but what Hi, Thanks for the ethernet, my world wouldn't have been anything like this without it. With due respect I want to disagree with you on your thoughts about Open Source and Linux. Saying that Unix has turned senile due to old age, is like telling that Wine gets bad when it gets old. Its like ethernet although its no match for newer technologies like ATM etc. But still it exists because its simple. The others are complicated, and they will take a long time to become stable enough. And Ethernet can take advantage of any technical advancements, as fast or maybe faster than the more complicated technologies. Unix is similar, its simple. And I don't see anything yet that is better. May be Plan9 or BeOS would be better, but how do they compete, with MS. Linux can compete because it is free, the rules of the business doesn't apply to it. Also its Unix, ie its simple, it will be adapted faster and better with the newest technologies (BeOS and Plan9 are anyway dead because they don't have the resources, I am just talking about the rest of the OSs out there). Linux would not win because its better or anything, it will win because it will always be optimized for the most common tasks and it is free. As people using GUIs become more common on Linux, its GUI will keep on increasing. You must know how many bad technologies have won because they were open and cheaper. Linux in not that bad, its open and its free. It is bound to win. If the Unix that you see is senile, all the kudos should go to the Unix Vendors, who conceded defeat before the war had begun. They just kept on fighting for the ever smaller Unix server space, because that gave them huge profit margins. If MS is a bully, they are fools. Have you ever wondered what drives Open Source developers? It is egotism. Ever read Ayn Rand, that is the character she associated with capitalism. At least as far as she is concerned, Open Source Developers are capitalistic. The only difference is that they don't deal in money, they deal in intellectual thoughts and peer respect. Why does one want to earn lots of money? If it was only for living comfortably and with peace, I don't think you need all that money. There are many reasons you would want to earn more money, to show the world, to gain respect, to gain power. Some are plain misers, they want to amass the money. Some others don't care about the money but they would not let others get the fruits of their labour, which they do because they enjoy it. Open Source developers also want to do the same things, but they don't care about everybody. They just care about people who are like them. They want to show them, they want to gain their respect, they want to gain some power over them. There are others who just want to hog all the limelight. And their are those who want to just program they don't care whether anybody likes thier programs or not, but they would not let anybody earn money from their labour. Hence the GPL. So effectively Open Source is really Capitalism, without the money ;-). Communism also involves force. You won't find any force in Open Source. They don't use anymore force than software companies use. They just use GPL to preserve their right to give their software to anybody who would not try to benefit from them in an unethical way. Open Source would win, but that doesn't mean proprietory software would be extinct, it would just be relegated to niches, just like Open Hardware (with open interfaces) (intel PCs) relegated the proprietory hardware (Sun, SGI, HP, etc) to a niche market. MS is acually killing the goose, by closing their interfaces, with everybody. They prospered once, because they were a lot open. But now they are trying to do too many things and using their proprietory interfaces to support their software. I hope you would think about it. thanks again for the ethernet :-). -anand | ||
Date: Mon, 21 Jun 1999 15:09:55 +0000 From: Dave Finton <dfinton@leonine.com> To: metcalfe@idg.net, letters@infoworld.com, letters@lwn.net Subject: Response to Metcalfe's recent article regarding "Open Sores" I've just recently read your article entitled "Linux's '60s technology, open-sores ideology won't beat W2K, but what will?" I have a few issues with this article, and would like to point out a few items that you've brought up. > Why do I think Linux won't kill Windows? Two reasons. The Open > Source Movement's ideology is utopian balderdash. And Linux is > 30-year-old technology. > This argument has been thrown around before, and it was shown to be hogwash. Linux, as you know, is based off of Unix, which was originally developed 30 years ago. Unix was designed from the very beginning to handle things that the original inventors could not envision. Therefore, the Unix framework is very bare-bones. But it was also designed to be extensible. The fact that Unix is being used today for a wide variety of purposes only serves to prove my point. Did AT&T actually think that one day Unix would be used for desktop applications like modern-day word-processors like Word Perfect or (soon-to-be) Lotus Notes? Heck I don't think they thought Unix could handle having a GUI. But today it has those things, and LOTS more. And to your assertion about Open Source. Open Source is NOT utopian balderdash. It is the highest form of pragmatism a computer programmer can think of. No secret little API's, no hidden "features". If I want to know if something works, I look at it. If it's broken, I fix it and then get my work done. No more spending hours (or days even!) finding workarounds or wondering why my application won't work as advertised. In this light, it is the "black box" model of thinking that most software companies today subscribe to that is utopian balderdash. Why do I care that Microsoft considers its API's and the underlying code its private property? That's MS's problem, not mine. I'll stick to something that won't try to hide from me the important details of a toolkit just because some multi-billion dollar corporation thinks I don't "need to know" those vitally important details. > The Open Source Movement reminds me of communism Funny, the Open Source Movement reminds me of capitalism. No one has some sort of proprietary lock on the market, and the rules of competition decide who wins. Hey, isn't that how a free market is supposed to work? I'll make another analogy: The closed-source model reminds me of a dictatorship, with one ruling class (Microsoft) reigning with an iron fist over subjects friends, and foes. *This* is the system you like to work under? Yeesh! > Stallman's EMACS was brilliant in the 1970s, but today we demand > more, specifically Microsoft Word, which can't be written over a > weekend, no matter how much Coke you drink. > You're comparing apples with oranges. I wouldn't dare use Word for application development. Why use Emacs for writing business documents? Your analogy fails here, too. And, BTW, Emacs is *still* a fine piece of work. I use it. Daily. Beats the hell out of any other IDE I've ever used (including Borland's IDE, among others). > Unix and the Internet turn 30 this summer. Both are senile, > according to journalist Peter Salus, who like me is old enough, but > not too old, to remember. The Open Sores Movement asks us to ignore > three decades of innovation. It's just a notch above Luddism. At > least they're not bombing Redmond. Not yet anyway. > The Open Source movement asks us to use the last three decades of innovation that have been put into Unix and related technology, and extend it even further. It's Microsoft that asks us to throw away all that hard work. And for what? A cheap knock-off? Don't get me started over NT. It only *wishes* it were Unix. And this "senile" internet you speak of... um, yeah right. The greatest engineering feat in decades that allows people across the world to communicate transparently without any hassle (and without long-distance phone charges). This senile old internet has done more for this civilization than anything else since the industrial revolution. So what you ask? What about the next-generation (not-so-senile) internet? Last time I checked, Microsoft *still* doesn't have good support for IPv6, and it was damned expensive too. Linux has it. If I actually had a connection to this new network, I could connect to it using the utilities I already have on my hard drive. So much for "senile". > All Unixes make up 17 percent, and Linux is a small fraction of >that. Nope, Linux is in it's own category. That 17 percent are Unix OS's *other* than Linux. Linux market share went up 212% last year. NT and Netware share went *down*. I'll make this prediction: Sure W2K will go up in market share, temporarily. But its price tag (concerning support, hardware, and licensing costs) will keep it from destroying Linux. Quite the opposite in fact. NT is going to have a struggle ahead of it just to survive. So much for that. In a departing note, I'll ask this question. Remember IBM, or DEC, The unstoppable juggernauts of Christmas past? IBM barely made it into the 90's with the shirt on its back, and DEC is now a subsidiary of Compaq. Oops, looks like those juggernauts weren't so invincible after all. ;^) - Dave Unix Systems and Applications Consultant | ||
Date: Mon, 21 Jun 1999 14:45:21 -0700 From: Ariel Faigon <ariel@cthulhu.engr.sgi.com> To: metcalfe@idg.net Subject: Re: your column: linux reality check [Response to Bob Metcalfe's linux column, Cc'ing letters@lwn.net and jallison@sgi.com] Dear Bob, As a pretty practical person watching the computer industry who is nevertheless, a strong believer in linux, I have to say your column was so far off and detached from reality that I'm not sure where to start. To start: seems like you've misquoted the IDC report on server operating systems. All commercial Unix'es combined, *not* including linux were 17.4%, while linux is an *additional* 17.2%, up 17% from nowhere five years ago and almost tripling from 6.8% a year ago (please go and reread those IDC numbers.) In the same report, Microsoft NT remained at a stable 36%, not 60% as you write. Unfortunately, your column doesn't mention the source for the numbers, this is hardly responsible journalism. Here's one possible source for the numbers I'm quoting: http://www.news.com/News/Item/0,4,30027,00.html?st.ne.fd.gif.a Your assessment of the amount of talent applied to linux is just as wrong. At my company alone there is a large number of people paid full time to work on linux. They may not be on your radar screen yet, but as some linux companies go public in the next few months this trend will become clearer. In the meantime, you may search the linux kernel mailing list or sources, for some big company names. I recently checked the netcraft web-servers survey at www.netcraft.com. Updated monthly, it is the only fully automated, unbiased survey of web servers on the open Internet, and noted that Microsoft IIS has slipped in market share for the third month in a row (after rising for over two years), while web servers running on Linux continue their steady rise (see note on the new CnG server who took 1.45% away from Apache). While all the web servers behind firewalls aren't represented, this is the best I can rely on. Serious readers should keep checking this site for a cue on linux market share, rather than relying on anyone's opinionated columns. Linux is winning not because of the ideals (yes, some of its proponents are extreme idealists, more power to them, so what?). It is winning market share because it is a very practical, flexible, and cost efficient solution for many computing needs. One example is network file-serving to Microsoft clients which Samba does better and much more cheaply than Microsoft's own NT server. [Cc'ig Jeremy Allison of the Samba team who is a full-time employee of SGI] To anyone who have witnessed the greatest boost to capitalism (read Internet commerce and services) ushered by free software like TCP/IP, HTTP, BIND, perl, Apache, etc. the reasons should be obvious: linux commoditizes the OS layer. This is unlike your perception, capitalism at its best. And this is not an anti-MS statement. MS is an extremely clever company. Expect MS to jump on the Linux bandwagon (shifting its point of "lock and control" up to the API) just as it standardized on TCP/IP when it realized LAN-Manager cannot win. If and when this happens, it'll be great for consumers. If you have any question about Linux and free software, with regards to business models, commercial potential etc. please don't hesitate to email me. I may have the answers for you. Could this be another "I'm eating my column as promised" column in the making ? :-) -- Peace, Ariel | ||
|