Letters to the Editor

In my article “Little Devil Called tr” in issue 53, a mistake was made by the editors. In my submission I wrote:

        !}tr A-Z a-z

        !jtr a-z A-Z

The editor changed this by prepending a “:” to the commands. That is definitely wrong; in that case, you would start a subshell and it would try to run }tr A-Z a-z and jtr a-z A-Z. Both of which would most likely fail. Without the “:” prepended, some lines of text (to be determined by the movement character) are piped to tr and the output is inserted back.

—Hans de Vreught hdev@kbs.twi.tudelft.nl

In “Training on a Token Ring Network” (September 1998), I referred to the wrong IBM token ring card. The article should have stated the IBM token ring ISA card. I apologize for this bug in my article.

—Charles Kitsuki kitsukic@computer.org

There has been a discussion going on about PPPui (a GUI for pppd) and the ways to check a PPP connection. It is not difficult to do. There is no need for special programs or to direct syslog to Console 9 as one reader suggested. I think the easiest way is to run pppd with the -d (debug) option and chat with the -v (verbose) option and then, during the process of establishing the PPP connection, just run tail -f /var/log/messages and all the details will be output (including the assigned IP address and so on) to this file.

—Mihai Bisca mbasca@univermed-cdgm.ro

I know it may sound a bit childish, but I wanted to show this to the community:

bernward:~$ uname -a
Linux bernward 1.2.13 #2 Mon Dec 9 10:33:11 MET 1996 i486
bernward:~$ uptime
  5:32pm up 430 days, 1:55, 3 users, load average: 0.00, 0.02, 0.00

This Linux box has performed admirably since I installed it back in 1995. It is an aging 486/33 with 32MB RAM and some 2GB of SCSI disks. It serves as a primary DNS and mail relay for our whole European WAN (a few dozen sites) and handles an average of 200+ MB of e-mail a week. It also carries out other menial tasks, such as network monitoring and some form of gateway between UNIX and Netware. It has never crashed once.

That said, I know I'm not the only one. A recent poll organized on http://slashdot.org/ showed that around 10% of the participants had a Linux box with an uptime above the one year mark.

—Philippe Andersson philippe_andersson@ste.scitex.com

Some of the articles in your excellent magazine include program listings. I know it is possible to get these listings if I know the exact location/file name; however, if I don't remember this, I am lost. I can browse the “Table of Contents” on the web site and find the article in question, but there is no link or clue as to where the program listings might be found. How about including a link to each article's program listing from the web site TOC?

—Jan Thomas Moldung janth@x2.hd.uib.no

Good suggestion-we've put links into the TOC. All listings are located at ftp://ftp.linuxjournal.com/pub/lj/listings/issue##/, where ## is the issue in question. Inside each issue directory is a README file identifying the article to which each archive file corresponds —Editor

I have been using Linux for some time now and like to show it off to my friends. One embarrassing problem for me lately has been the need to boot from a floppy, because a hard drive was too large for my BIOS and I couldn't configure LILO properly. A recent response in the “Best of Tech Support” column supplied my answer, and now it boots fine.

The technician who installed the drive under Windows 95 had to shorten it to accommodate the BIOS and Windows 95. Linux has allowed me to reclaim the rest of the drive and ditch Windows 95.

Your magazine is my favourite Linux resource. Congratulations.

—Stephen Roach stroach@altnews.com.au

I just finished another fine read—nice issue as always. One minor error I noticed though is on page 45, “Technical Considerations” by Richard Kent.

He says, “Within these toolkits are functions which have a variable number of arguments, much like the standard printf system call.”

Ah, and there's the rub—printf isn't a system call. It is a function from the standard library (section 3). System calls are, as you know, described in section 2.

I know this is minor, but students new to the C programming language have enough trouble with the documentation and how to read it without reference problems.

—Wayne Bjorken wab@courier.cb.lucent.com

I have done years of FORTRAN, C and assembly programming on CPM, NS32000 and DOS systems, developing a system for operating laboratory equipment, and processing and displaying experimental data. After having seen so many platforms disappear, each time forcing a painful migration, I have just begun moving to Linux. The first issue I confronted, and resolved by making a semi-random choice, was selecting the distribution. There seems to be no guidance for newbies in this matter. After I got it installed, I started trying to learn how to program it. Here I ran into another obstacle. I have been spoiled by the packaged and documented software from Borland. Now I have to find tools for Linux and instructions for using them by looking in a huge collection of books and Internet sites. I find there is a horrendously steep and bewildering initial portion of the Linux learning curve, which could easily be a barrier to many people. Distributions need to address this if Linux is going to compete with MS Windows.

—Bill McConnaughey mcconnau@biochem.wustl.edu

I read your interview with Charles Andres in the August, 1998 LJ with great interest.

I just want to add one point on your question about “How does Sun feel about the Open Source movement?” If Sun feels that it might be advantageous for their business to give the source code to everyone, they will do so. Proof of this statement: When Sun tried to start a “Motif vs. OpenLook” war, they freely gave away the source code of XView (which was one of the best X toolkits around). It didn't help them win that war, but in the Linux community XView can still be used (guess what I am looking at ...) for free. All this happened long before the issue hit the newsstands.

Same story with Netscape: if a company feels it can win something, it will open the source. Sun will do it again if they think it would be good for them, but they won't do it for political or philosophical reasons.

—Erwin Dieterich e.dieterich@ndh.net

I have been reading Linux Journal for quite some time. It is a very informative magazine and I like it a lot.

However, I cannot help noticing one very bad thing: the portion of LJ occupied by advertisements has reached approximately one third of the entire magazine. It seems to be growing even further at the cost of actual content.

While it is clear to me that you earn good money from the advertisements, in the end you still need your subscribers. I am afraid you might lose at least one, if you continue your transformation from a journal into an advertising bulletin. I hope you will realize this before it is too late.

—Denis Havlik havlik@lisa.exp.univie.ac.at

Actually, compared to other magazines, 30% advertising is low. We need at least this much to stay in business without raising subscription rates. Our November issue was at 35%. If advertising either stabilizes or increases, we will most likely expand the magazine by another 16 pages. It is also true that many readers find value in the ads—I even had one who said we should increase the number of ads —Editor

I am Guy Barrand from Linear Accelerator Laboratory (LAL) at Orsay (France).

In the article “Open Inventor” by Robert Hartley (September 1998), Mr. Hartley mentions the Apprentice project. I have looked at the Apprentice code, and a question has occurred to me.

How far can one go in re-implementing commercial software? In Apprentice, the API differs from that of Inventor only by the prefix “Ap” that replaces the Inventor “So”. A good use of the tr command (also documented in the same issue of LJ) could easily transform an Apprentice distribution to an Inventor one. Does the Apprentice developer have the right to use the “So” prefix?

In general, is it legal to reuse a commercial product API and to provide a free implementation of this product? I assume that the Linux community has looked at these problems for a long time. Can you enlighten me on these points?

—Guy Barrand barrand@lal.in2p3.fr

Good question—I don't have the answer. Perhaps, one of our readers will know the legalities and let us know —Editor

I track operating environments for IDC and just read Phil Hughes' article on the Open Source Developers Day (November 1998). Just wanted to take a moment to echo your thoughts at the end of the article regarding applications and Linux.

I often speak to my clients and vendors about Linux (or the “Linux Experience” as I've taken to calling it; one has to have some fun). I have been telling these clients and vendors that applications drive OS sales. It really is that simple. Although there are many barriers to Linux in enterprise, the recent application development advancements/announcements are a positive step for Linux. However, now is not the time to rest on these victories. When I think about Linux “standards” and applications, I believe that involvement from the beginning by the application vendors is crucial.

—Bill Peterson wpeterson@idc.com