I have just received the April 1998 issue of Linux Journal. My highest regards for such an informative article as “Satellite Remote Sensing of the Oceans” by S. Keogh, E. Oikonomou, D. Ballestero and I. Robinson. My attention was held very closely by the details and explanations written in this article. I'm well aware of the problems in remote sensing systems, and the potentially enormous amounts of data which must be manipulated in order to make sense of it. This was great! Find more authors like them.

—Bill Staehle staehle@netvalue.net

First let me congratulate all of LJ's editors. I am staying up hours to digest and learn all I can possibly read. It is really wonderful to have a chance to learn so much.

Here I am referring to “Writing a Linux Driver” by Fernando Matía. I find it to be a good foundation for device drivers. An added reference for anyone who would like to look into drivers further is Writing A UNIX Device Driver by Janet Egan and Thomas Teixeira (John Wiley and Sons, 1988). Even though it is for UNIX, it proves very useful for Linux. Thanking you and Fernando Matía.

—Silvio agola@grumpy.igpp.ucla.edu

I have subscribed to your wonderful magazine since 1995 and I have found it to be useful, informative and fair. But I must tell you my disappointment about your graphical user interfaces issue, March 1998.a) XView is old and passing out of favor, even though it is still useful.b) CDE is proprietary.c) TkDesk is nothing more than an elaborate file manager.d) GTK and GNOME are by no means ready for production machines.

You failed to mention the wonderful, useful, highly advanced KDE project (see http://www.kde.org/), despite the fact that it is in the beta three stage. I have installed it on production machines in radiological clinics under both Linux (x86 architecture) and SunOS. I wonder why you did not dedicate a few lines to a project like KDE—in my opinion, the only runner capable of stopping the wave of Microsoft's products. If libQT licenses are the problem, it is inconsistent with the presence of articles on all-commercial products like CDE and X-Designer/Motif, since libQT is free for free Linux development. Please reconsider this position and take a look at KDE.

Thank you very much for your consideration and keep up the excellent LJ--I love it.

—Daniel Benenstein dbenenst@cs.com.uy

There was no conscious decision on my part to snub KDE. In fact, I had a KDE article scheduled, but it came in after the deadline and the issue was already full. I plan to run this article in the near future. The May issue does have a short comparison review of both KDE and GNOME in the Linux Gazette column —Editor

The most elegant method for dealing with cookies is simply to symbolically link the cookie file to /dev/null, like so:

ln -s ~/.netscape/cookies /dev/null

This has the effect of accepting all cookies, so you are not denied access to any web sites, but immediately funnels them into a black hole. Having /dev/null “world readable” is infinitely preferable to disclosing your browsing habits to webbots.

—M. Leo Cooper thegrendel@theriver.com

I have been reading LJ (which I find very nice) for three months. I have learned different things (as I do from any journal), but I have also found some bugs. In the article “Ghostscript” by Robert Kiesling (March 1998), there are two points I do not agree with.

The author says that it is not possible to see .eps files included in LaTeX files with xdvi. I say: wrong. I use LaTeX2e and xdvi almost every day, and I can see the .eps files through xdvi. (I cannot zoom, but I can see what is displayed, which is usually enough.) The .eps files I use are generated by xfig and transfig.

The author says that gs used with device X11 will create a window. I have tried this, and the X11 device is not a default one for gs 4.xx. I can see my .ps files with gs, but I must use Ghostview as the GUI. I do not think that the programmers have taken the X11 device away from version 3 to version 4. Has the author made a mistake in his script files?

I am using Red Hat 4.0 (with a lot of patches and a Slackware-like installation) but my LaTeX2e, xdvi and Ghost(script/view) are original and standard ones, so I should be able to do what the author mentioned, but I cannot.

—Raphael Marvie marvierx@cs.man.ac.uk

I have to blame my use or installation of xdvi for the inability to print \special commands. The distribution on which the article is based has been upgraded several times now. I've received plenty of mail from TeXperts saying they have no problem with reproducing EPS graphics on screen.

The X11 driver is standard in every recent version of Ghostscript, but it is not necessarily the default. It can, however, be specified on the command line with the parameter -sDEVICE=X11. During the process of building Ghostscript, it and all of the other supplied driver code can be specified in the Makefile.

—Robert Kiesling kiesling@terracom.net

I noticed in your April edition, you used PhotoShop to create the cover image. I understand that at SSC you use non-Linux systems for some of your graphics and page layout. Wouldn't it be nice if you could do it all in Linux? I think the GIMP is great, but nothing on Linux compares with PhotoShop.

The only thing to do is to port PhotoShop (and possibly Framemaker) to Linux. Adobe is aware of the desire for such products, and even confronts it on the following web pages: http://www.adobe.com/supportservice/custsupport/QANDA/259e.htm and http://www.adobe.com/supportservice/custsupport/QANDA/2bf6.htm I believe if enough people show an interest, Adobe will eventually come around. What a great day that will be for Linux! Adobe asks that we send all suggestions of this nature to:

Adobe Systems, Inc.PhotoShop Product Management345 Park AvenueSan Jose, CA 95110-2704

I hope that many of you will send requests to Adobe to let them know how you feel.

—Jason F.jasonf@usi.net

I'd like to ask in an open forum that computer product resellers, and Linux folks specifically, allow for U.S. forces overseas to have better web commerce access.

Although we can often order products, many web-based address databases are unable to accept an APO/FPO address. There also seems to be a small lack of understanding of how mail from the U.S. is delivered to troops stationed overseas.

When a package is sent to an APO/FPO address, it is carried by U.S. Postal Service mail to one of three military postal offices. New York serves European troops, Miami serves troops in the southern hemisphere, and San Francisco serves the Pacific installations. There is no additional cost to the shipper—the military picks up the tab as soon as it gets to the APO/FPO. If the city is added to the address, it can have terminal effects on packages, many of which never arrive or are bounced around several places. We cannot receive mail from either FedEx or UPS.

There are web commerce servers that provide for this type of addressing, but more need to be available. I appreciate the efforts of the web maintainers in keeping good commerce available and ask for this request to be considered.

—Leam Hallgers hom@spidernet.it

Lost & Found

Write us at ljeditor@linuxjournal.com or send snail mail to Linux Journal, P.O. Box 980985, Houston, TX 77098. All published letters are subject to editing.