Letters to the Editor

Various

Issue #41, September 1997

Readers sound off.

Misspelling

It's great to see two (count them) of our books reviewed in a single issue. [June, 1997] However, you spelled my name wrong in the “Programming with GNU Software” article by Randyl Britten. I take that as a lapse on my part. I should be in touch with you more often. (I tend to just drop by the booth at occasional conferences.)

—Andy Oram O'Reilly & Associates, Inc .andyo@ora.com

Thanks for being so understanding. Unfortunately—a-hem—I have to take full responsibility for the spelling. It's in my original manuscript and there was no excuse. I offer my apologies.

—Randyl Britten britten@u.washington.edu

More Novice Articles

Thanks for your seamless renewal of my subscription, even though I had decided in favour of the new UK Linux World magazine, which has now disappeared after one issue.

LJ continues as well as ever, but I'd like to add my voice to the novice lobby—more stuff for us please. After some easily understood articles on RCS, shell scripting and the like, you've become all technical and clever again. I'd like to see stuff on make and gdb—I've downloaded many applications as source archives only to find the authors assume abilities I don't have. Stuff on the startup and shutdown scripts would also be nice. Anyway, great magazine.

—Bob Smith bob@bank.demon.co.uk

For every opinion, there's an equal and opposite opinion. Read on —Editor

More Technical Articles

I would like to see Linux Journal remain a Linux magazine, and not move towards more WWWsmith articles, as you have done in the last few months.

Also, I would like to see LJ become more technical, and move away from the novice corner type stuff, which can easily be found in more current form in places such as Linux Gazette, newsgroups, etc.

When I say technical, I mean articles like Alessandro Rubini's excellent Kernel Korner columns. Or detailed articles on Perl and shell scripting, hardware ports to alpha or networking (the article on ghosting in June is a nice detailed article). [“Ghosting Onto the Net”, Scott Steadman, June 1997]

I understand that Linux is new in some sense, and you may feel justified in having the “simpler” stuff in there, so I offer my feedback as simply another data point for your consideration.

—Les Schaffer godzilla@futuris.net

We strive to be balanced, offering both a Kernel Korner and a Linux Apprentice column each month —Editor

Keyboard Typos

I very much enjoyed your keyboard article in the June Linux Journal. [“Consistent Keyboard Configuration”, John Bunch] The article as published by LJ had several typos. I was able to get most everything to work as described. Did not have much luck with the arrow keys in Emacs running in an xterm. I have not been able to determine the cause of this. Something to do with the xterm translations?

—W. Paul Mills wpmills@sound.net

Note that on page 54, six lines are broken, forcing “Arrow” onto the following line. These lines should be joined so that “Arrow” is inside the comment. Also note that the escape sequences are incorrect. The double backslashes should all be single backslashes, so for example, the line for F117 should read:

string F117 = "\033\033[A" # Alt-Up Arrow

This type of error is present throughout the article. On page 57, the key translation lines have two problems. First, all of the double backslashes should be changed to single backslashes. Second, the lines were broken improperly. The first seven lines are shown in Listing 1.

Listing 1. Keyboard Corrections

Copy the rest of the lines from the man page for xterm(1). Every line of the translations, except for the last line, should end with either \n\ or \. Typographical errors here are very serious, because they cause problems without generating any error messages.

Let me know if this helps.

—John F. Bunch bunch@ro.com

Benchmark Table

I have a little big complaint about LJ Issue 37, in particular the native PowerPC article [“Native Linux on the PowerPC”, Cort Dougan, May 1997]. There was a performance listing which compared MKLinux, native Linux-PPC and OSF and some Sun operating systems, but the table was typeset all wrong. Being so cryptic it is almost, if not completely, useless since one cannot tell which digits belong to which columns. If you have so many problems getting your magazine printed correctly, you should probably hire better people, like me for instance.

—Ville Voutilainen vjv@stekt.oulu.fi

Think Linux

Hello. My name is Giacomo Maestranzi. I am from Italy and I am a frequent reader of Linux Journal because I find it very interesting—especially the last issue (June 1997) which covered Linux use in my region (TRENTINO ALTO ADIGE and the city is BOLZANO). [“Traveling Linux”, Maurizio Cachia] What I just have to say is that I love Linux and everything related to it, and I have created a slogan for it that I would like to see in a future issue. The slogan is:

THINK FREE...... THINK BIG.......THINK LINUX

Thank you again for LJ. I apologize for my English, but this is my best.

—Giacomo Maestranzi hoteuropeo@well.it

Push Media

Kudos to Doc Searls for his insight into how the WWW has the “mainstream media/advertisers” scrambling to remain a monopoly. [“Shoveling Push Media”, June 1997]

I have been an active WWW user for about 2 years—started on AOL as a newbie and graduated to a direct ISP. I am not interested in push technology. My senses are offended often enough when I turn on the TV or radio. Push technology will appeal only to the “couch potatoes” of the world who purchase WebTVs to impress their technophobic friends. I say leave TV on the TV.

The beauty of the WWW is that I can control what I see. If I find a site uninteresting, I leave. The other compelling aspect of using the Web is that I can remove the classic “middle man” from my business transactions. When I want to buy something using the Web, I go straight to the person(s) who offers it. I do not have to be offended, goaded or otherwise angered by traditional advertising. This fact scares technologically savvy advertisers. I would be scared too, but that is the reason I write software for a living.

—Jeffery C. Cann jc_cann@ix.netcom.com

Perfect Box

Eric Raymond's article, “Building the Perfect Box” (April, 1997) was quite instructive, but he did not mention one essential component—the keyboard. In my experience, this is one place you should not skimp. A bad keyboard is frustrating to use and may contribute to carpal tunnel syndrome. I like the top-of-the-line IBM keyboards, but you should always try one out before you buy it. When you experiment, you should sit in the same position you use when typing.

By the way, if there is anyone out there who does not have Raymond's book, The New Hacker's Dictionary, go out and buy it right now.

—James R. Miller jimxc@jimxc.seanet.com

Wireless Solution

I read the “Linux Means Business” column with interest this month [“Connecting SSC via Wireless Modem”, Liem Bahneman, May 1997], as I have recently switched from an analog modem to a wireless solution for net access to the office from home. I recommend Metricom's Ricochet service (http://www.ricochet.net/) very highly. I consistently receive transfer rates of 2.5 to 3.5KB/s, and it is child's play to use it under Linux; it supports the standard Hayes AT command set. I've never received a busy signal and establishing a connection is lightning fast compared to an analog modem's handshake.

—Nick Silberstein nick@fusion.com

Wabi Article

I was surprised to see in the Wabi product review by Dwight Johnson (LJ, June 1997) a suggestion to chmod<\!s>666 /dev/fd0. Giving random users permission to write to your floppy drive is not exactly a good thing to do.

Also, near the end of the article, it was mentioned that one should wait for Wabi 3.0 to drive 24-bit displays. Last time I checked, Caldera's Wabi 2.4c (not yet released) should fix the 24-bit display problem.

As for the “seamless integration of Microsoft Windows with Linux,” I personally find Wabi's handling of the focus most annoying. For Windows tasks that take time to complete, you can easily create havoc by focusing on a Linux window to do something, just to be interrupted in the middle by a “regain of focus” by a Wabi task. For machines with a lot of memory (and therefore the ability to run two X servers), I find running Wabi on a separate X server to be the safest.

(The article also did not cover keyboard remapping; the information on keyboard remapping found on Caldera's web site is not exactly helpful.)

—Ambrose Liac li@acli%interlog.com

Security and Networking

I just received your recent issue of Linux Journal, and it was very helpful and informative. [“The SYN Denial of Service”, Douglas Stewart, et al., June 1997] I do have some problems with the SYN denial of service prevention. The source was published in Phrack magazine, and also includes the methods of prevention (which were the same as you had discussed in LJ). Unfortunately, TCP SYN flooding is only one of many attacks; there is also Project Hades which deals with TCP exploitation, and Project Loki which presents the theory of ICMP_ECHO tunneling. These are just the articles I have read. If you want to stay ahead in the security field, read these articles as they also contain methods of prevention. Phrack can be found at http://www.fc.net/phrack. I hope this is of some help.

—Tom McHannes pnmtofte@imperium.net

Progress on Linux

Just read with interest your 1996 article about your company's use of Progress on Linux. [“Sticking with Progress”, Peter Struijk and Lydia Kinata, September 1996]

As an international non-profit organization, we have used SCO Unix for 8 years and have been very pleased with it, except for the growing resources needed to run it, as well as the cost to buy and upgrade it.

At our international meetings in London last month, we decided to do some testing with Linux to see if it works in our environment and if it could be used as a replacement for our office networking systems. Our German office has been running nicely on Linux for several months, but has yet to get the Progress kit for testing.

  1. Did you have a shared library from SCO or did you have to buy the license to make running Progress legal?

  2. Are you running Progress 7 or 8 now?

We are looking at linking Progress with “static binding” to eliminate the need for the shared libraries, since we have the full development kit.

Progress says it will probably never support Linux directly since there is no standards body to refer to as there is with its commercial counterparts.

—Ron Tenny ron@omusa.om.org

We actually use the free libraries distributed with iBCS: that is, they are available in a separate archive. The only two we need are libc_s and libnsl_s. I believe we still have an (old) SCO license, but we never had a need to use the original libs (although they work fine, too).

We are still running V6, but I've heard reports from the east coast and Canada that V7 and V8 can be run on Linux without (major) problems.

To join our mailing list for Linux Progress users, send e-mail to info@linuxjournal.com with one line in the body containing the word “info”.

—Peter Struijk info@linuxjournal.com