Letters to the Editor
There is a discussion on the GLUE (Groups of Linux Users Everywhere, http://www.ssc.com/glue/) e-mail list about a certification program for Linux. Most of the discussion is positive. Most writers feel that a comprehensive certification will make great strides in enabling Linux to be used in the business sector with more confidence. I would like to know how a larger audience feels about it. For any certification program to be representative, it must be a cooperative effort of all the major Linux companies. A relatively subjective organization would need to head the testing such as the LDP or SSC.
—Bryan Coleman, Triad Linux Users Group bcoleman@triadsearch.com
I thought Phil Hughes's May article on web browsers for Linux was a little negative. [“Linux and Web Browsers”, Issue 37] Here is some more positive news: Sun's HotJava 1.0 browser is available from http://java.sun.com/products/hotjava/. When you look at the distribution formats it has versions for Windows and Solaris. However, the browser is written in Java, just called from a different shell script for the two platforms. Download the Solaris version and run it under JDK 1.1.1 to get a working browser for Linux.
Sun MicroSystems, Inc. has bundled the HotJava web browser with the Java Runtime Library for SunOS on SPARC hardware. The download file is now 8.5MB.
Another choice is the Plume browser (formerly Surfit!) by Steve Ball at the Australian National University (http://tcltk.anu.edu.au/). This runs under Tcl/Tk 8.0. It is still under development, but Steve is actively working on it. What's more, you get the source code so you can do things with it too. The current version of Plume is v0.62alpha.
—Jan Newmarch jan@ise.canberra.edu.au
As an avid amateur radio operator and a Linux tinkerer for nearly a year, I'd like to say a hearty “Thank You” for the positive coverage you give my favorite hobby in Linux Journal [“Packet Radio Under Linux”, Jeff Tranter, September 1997] and Linux Gazette [Issues 10 and 11]. Of course, it's great having access to the only OS that supports the packet radio protocol. Most other big-time magazines wouldn't bother to print such articles, but it proves the editorial commitment you have to covering all relevant aspects of Linux: business, technical, hobby/recreation and more.
I'm always glad to see this type of article as it introduces ham radio to a larger audience. We're always looking for more hams willing to push the digital RF (Radio Frequency) envelope. I invite all interested parties to get their license and join in building a state-of-the-art, wireless, non-commercial TCP/IP network.
—Nate Bargmann KA0RNY nfbargma@notes.up.com
If I'm not mistaken, anyone on the Internet can execute any command on a machine with the CGI scripts you published on page 58 of LJ's August issue [“Big Brother Network Monitoring System”, Paul M. Sittler].The script executes $TRACEROUTE<\!s>$*, so a cracker can feed it with a machine name such as www.tamu.edu; then type cat/etc/passwd to see the last command being executed.
In my opinion, CGI scripts should all be written in Perl with the -T option set (-T tests that the file type is text, not binary) and should include the line use<\!s>strict. Strict compliance for symbolic references, global variables and key words—violations cause immediate program abend. The Bourne shell is especially dangerous. At least, enclose the arguments between double quotes.
—bortzmeyer@pasteur.fr
I don't know that I have ever written to a magazine editor before, but Lee Brotzman's contribution to the August 1997 Linux Journal, “Wrap a Security Blanket Around Your Computer,” was very timely and very well written.
One of my client's Linux systems came under the control of hackers (who fortunately were somewhat benign in their apparent intentions for this particular system) about the time I received the aforementioned copy of LJ. During an intense weekend of observation and examination of various system logs, I was able to determine how the system had been compromised. After considering various strategies (and reading the issue cover to cover), I used Mr. Brotzman's article as a cookbook to install a series of TCP wrappers while continuing to watch the hacker's activities.
Not really knowing the expertise of the hackers, I surmised they were also “cook-booking” and decided to slowly cut off their air supply, in order to see what alternative methods of access, or back doors, they may have established. Selectively applying TCP_wrappers enabled me to do just that, and I received quite an education in the process. Today, thanks to LJ and Lee Brotzman, my client's system is secure, and I have greatly increased my understanding of security from an administrative perspective.
Once again, thanks for publishing useful, accurate information. If you have an award for writer of the year, I would like to nominate Lee Brotzman for his clear, concise presentation of an important topic. If you don't, start one.
—Mel Lester meljr@connet80.com
I would like to commend you on such a fine publication. It is good to see a magazine published on a regular basis about Linux, that has such excellent content. I would just like to make a few comments.
My first comment is regarding a couple of letters to the Editor from the September issue (Issue 41). The two letters are “More Novice Articles” and “More Technical Articles”. I feel that LJ has a nice balance of both novice and advanced articles, which makes it useful for the beginner, the guru and everyone in between.
My second comment is about the article “Robocar: Unmanned Ground Robotics” by Kerry Kruempelstaedter (also Issue 41). I found this article very interesting, although I believe that many of the pictures were missing from the article. The first figure is Figure 5, and 1 through 4 are...nowhere in site.
And finally, I have a question regarding your article titled “Quota: Managing Your Disk Space” (and again, issue 41). Towards the end of the article January says that to give everybody on your system the same quota use edquota -p <prototype> *. If this gives everybody on your system the same quota. Wouldn't this cause some problems? It is fine that all users get the same quota, but what about root, bin, daemon and other such users?
Overall though, it was yet another excellent article.
—Philip Cox phil@yucc.yorku.ca
Yes, pictures were left out of the Robocar article. There was not room for all of them, and the ones in the magazine did not get renumbered.
It's true that you would not want to set root, etc. to the same quota as users. In the article, Mr. Rooijackers mentions setting soft and hard to 0 for those accounts that should not have a quota. In the quota HOWTO, it recommends turning quota on and off for various partitions. Thus, if all your users are on one partition, while root is on another, you can use the above command to set the quota for all users, while not setting it for root.
I don't know why you invented a non-existent paper to blame the Hewson anti-Linux article for in LJ #40 [From the Editor]. You get it right in the footnote (The Sunday Times), but there is no such paper as The London Times as in your heading. The Times, the paper some people think published the article, is a different newspaper from The Sunday Times.
Congratulations on the generally high standard of LJ.
—Keith Briggs, England kmb28@cam.ac.uk
Sorry about that—the reference to it I had said the London Times. The editor who checked the article later and added the footnote should also have changed the reference in the body of the article.
The article on Microstation for Linux, in the July issue was nicely done and well written. The only problem is in the section “Installed Base” where it states “The review kit states that there are over 200,000 users of Microstation...” I am sure they don't include academic versions in that number. If this is the case, then none of the 200,000 are running Linux. In addition, the academic version is all that is available for Linux and to get it you must prove you are a student or faculty member of a four-year-degree university.
The sad truth of the matter is that Bentley, and for that matter most other software companies, don't get enough requests for Linux ports of their product, to justify the production costs. To illustrate this point, here is a portion of an e-mail I received from Phil Chouinard, a Bentley Representative, on August 4, 1997.
... to date, a complete commercial port hasn't happened. About the only thing that is holding us back right now is a commitment from the business community. Note that we do not get involved or wrapped up in OS wars—there is no winner in such (in fact, we've heard the same arguments from MacOS loyalists, OS/2 followers, etc.) Since we have commercial products running on these OSes where others don't, and we have also committed to the Internet—recently licensing Java from Sun, convincing us to do a full -fledged port really won't convince the business community to go to Linux, but the opposite may be true.
The above, I fear, is typical of the software community as a whole. Linux is now at an awkward moment. If business will use Linux, software companies will write programs for it. If software companies will write programs for Linux, business will use Linux. Neither wants to be first. How do we get out of this catch 22?
—Dave Blond elln9jdz@ix.netcom.com
More information on Bentley Systems, Inc., publisher of MicroStation 95, can be obtained from their web site at http://www.bentley.com/academic/products/order.html.
After typing in the source code for xhello.c (July 1997, page 71), I tried to compile it using the author's suggested command line. Unfortunately for me it didn't work. I received a very vexing error from /usr/lib/libforms.so complaining of undefined functions. After fighting it a bit more, I decided to look at forms.ps and found a difference between the command line suggested in the manual and the author's. To make a long story short, this command line worked on my system:
gcc -o xhello xhello.c -lforms
-lX11 -lm
It has all the same elements as the author's, only in a different order. I am running a Slackware '96 system with kernel version 2.0.30, Xfree86 3.3 and xforms 0.86. I guess I'll go with what works.
—Nate Bargmann nfbargma@notes.up.com
First off, I would like to congratulate you on a job well done. It is a credit to the Linux community to have such a well-written and timely publication devoted to it.
I was wondering if any development has been done on integrating NDS (Novell Directory Services) into Linux? Novell has recently announced the release of the NDS source code to developers. Sun, HP and SCO have taken the hint and announced that each will provide NDS support in upcoming releases of their OSs. I have yet to find a place where Linux does not have the potential to replace a commercial operating system, and I think that support of a commonly used directory service would give a network manager even one more reason to use Linux instead of SCO ODT, Solaris, HP-UX or Netware.
—Tim McQueen mcquetm@mail.auburn.edu
Caldera is in the process of porting both NDS and NetWare File and Print services to Linux. Both these technologies currently require Streams, so Caldera is building the necessary infrastructure into Linux to port these technologies. For more information, please e-mail Caldera at info@caldera.com.
I have been unable to follow current chip discussions on the Net on the basis of these two new technologies: AMD's K6-233MHz with MMX processors and Intel Pentium II, 233&266MHz with MMX processors
Are these technologies compatible with Linux? Is MMX programming likely to be taken advantage of in Linux?
—Michael T. McGurty mmgurty@ovis.net
In answer to your first question, yes. Intel and AMD design backward compatability into all new processor chips to assure software compatability.
As to the second, Linux programming is likely to use advanced MMX technology sometime in the future, but I have no answer as to when—hopefully, this year. Of course, compatablity with the installed user base is a concern. It would have to be a kernel option. Multimedia extensions (MMX) optimize graphical video display (X Windows) and sound subsystems performance.
