Letters to the Editor
It has been pointed out to me that KHello no longer compiles correctly. This is because the KDE API has changed since the time my article “A First Look at KDE Programming” (August, 1998) and program were written. An updated version of the khello source code can now be found at http://www.chaos.umd.edu/~dsweet/KDE/KHello and in the tar file ftp://ftp.linuxjournal.com/pub/lj/listings/issue52/2653.tgz on the LJ FTP site.
—David Sweet [email protected]
I have thought about running Linux on an Intel notebook. As I began to look for notebooks which support Linux, I found a site that sells UNIX Server notebooks at http://www.tadpole.com/. They have both SPARC and Alpha notebooks. I was quite impressed. A SPARC20 model of the notebook supports Linux as well as an Alpha version.
Anyway, I thought your readers might like to know about this site. Maybe one day we will see a review of these notebooks in LJ.
—Robert Binzr [email protected]
If you get one, maybe you could do the review. —Editor
I just noticed that the article I wrote entitled “Linux Hits The Big Leagues” was printed with an incorrect e-mail address for me. The address should be altered to [email protected] from [email protected] Thanks.
—Sam Williams [email protected]
Simon Maurice's letter in the June Linux Journal leaves me speechless, so I have to type this response.
He accuses Red Hat of delivering a “truly bad release” and, as evidence, draws our attention to the errata list that includes listings of bugs for which no fix is yet available. Surely openness about what's fixed and what's not is at the heart of the Open Source movement. To go on and claim that Red Hat is behaving in a very Microsoft-like way defies both logic and the day-to-day experiences of system administrators around the world.
One could claim that Red Hat does not write the software containing these bugs, they merely package it, but in truth they do far more than that. Red Hat does a good job in maintaining their distribution and keeping it current. Their pioneering work with glibc (alongside the Debian project) is just one example of a benefit to the wider Linux community. If Mr. Maurice doesn't want to download a whole RPM file, he is welcome to download the source of any package and track patches from its maintainer. Nothing in the Red Hat distribution forces him to use RPMs—most users find them extremely convenient.
—Grant McLean [email protected]
I just finished reading the July issue, and I must say that it will most likely go on the shelf, never to be read by me again. I found no useful article in the entire issue. Don't get me wrong, I am all for Linux success stories, but most of the articles were so technically dry that I lost interest after a paragraph. I have been a subscriber since January 1998, and this is the first issue to achieve the title of “useless”. That being said, I love LJ and wish to see the ongoing improvement of both Linux and LJ. Thank you for hearing my whine.
—Griffin Caprio [email protected]
I have a couple of comments on an article in the June 1998 issue: “Introducing the Network Information Service for Linux” by Preston Brown.
At the time of writing (February 1998), the latest version of Red Hat was 5.0, not 4.2, so the remark about an older version of ypserv in 4.2 has no value. What I do find a little odd is that 5.0 (released December '97) shipped the same version of ypserv, considering 1.2.5 had been out since mid-October. Currently, though, Red Hat 5.1 ships 1.3.1, which was (presumably, considering we now have 1.3.2) the latest one when they closed the distribution (June 1998).
The /contrib/hurricane directory contains software for Red Hat 5.0 (code name Hurricane, a glibc-based distribution) that can't be used (except in a few trivial cases) on libc5-based systems, such as Red Hat 4.2.
I'd like to know which features I missed by not having ypbind, but the article doesn't give this information; I can only say that all my programs worked flawlessly.
domainname was not invoked in Red Hat 4.2, but is present in Red Hat 5.0 (see the comment above on release dates).
—Andrea Borgia [email protected]
As a new convert to Linux, I am very satisfied with my subscription to LJ. I have comments about two of the articles in the July issue, one positive, the other less so.
Thank you for the avalanche predictor model by Richard Sevenich and Rick Price. It reminded me—at an optimal time—of the power and utility of fuzzy logic to many of the real-world problems we deal with on each project. I've applied fuzzy logic to a GIS (Geographic Information System) for land use decision making support. But, in the press of too many things to do and not enough time, I had forgotten about this powerful set of tools. Thanks to this article, I have a solution to at least one environmental problem which affects the mining industry. Since we use Linux, the model will be developed and run on this platform. If there's demand, we'll make it available on Windows 95, too, using Tcl/Tk for the user interface.
On a less positive note, I need to ask why Richard Parry's article on position reporting using GPS and ham radio was published. After reading the article, I still have the unanswered question, “So what?”
We use GPS data extensively in our field work. We use it to delineate wetlands, orthorectify aerial photos, measure mines and quarries for modeling of storm water runoff and reclamation planning, locate wildlife nests/burrows and other phenomena of the world out of doors. We also know that many surface mines use GPS technology to dispatch haul trucks and maximize operations, many police and fire vehicles are equipped with GPS/GIS systems for emergency response, and at least one long-haul trucking company (Schneider, with the orange tractors and trailers) uses the technology to increase their efficiency, customer service and driver satisfaction. Given all this, what is the value of having one's vehicle location transmitted to a network of amateur radio stations? I was a licensed ham operator (many, many moons ago), so I'm certainly not belittling them.
What I'm saying is that this article appears to be more of a reporting on, “we know how to do this, so we will,” than a solution to a real problem. If I've missed something in the article, please do let me know. Each of the other five articles related to the issue theme report how Linux facilitates solving a problem. This one article doesn't appear—to me—to fit that mold.
Regardless, kudos are due to all of you for a very useful information source (including the ads). I'll be buying a SCSI adapter from one of your advertisers tomorrow, since my relatively new (but shortly out of warranty) HP/CMS tape drive has decided to die.
—Dr. Richard B. Shepard [email protected]
I thought Mr. Parry's article was fun; not every article needs to address a “real” problem. If I had a GPS article describing some of the ways it's used that you mention, I'd have been happy to publish it too —Editor
In the “Best of Technical Support” column in the July issue, Mark Bishop responded to a question titled “Editing motd and issue”. Mr. Bishop forgot to point out that the Slackware distribution, by default on startup, overwrites the /etc/motd and /etc/issue files. To change this, one must comment out (#) the commands that overwrite the /etc/motd and /etc/issue files. Check out the /etc/rd.d/rc.local (I think) file in Slackware to make the changes. After doing so, you can edit the /etc/motd and /etc/issue files without fear of them being overwritten.
—Andrew Dvorak [email protected]
You have contacted me about the Slackware distribution overwriting on bootup. Not all distributions do this, however, and I believe that hasn't always been the default behavior of Slackware. I wish I had the time to keep up with all the distributions, and I plan on installing Slackware on my new machine now that a new version has been released (too bad it's not based on libc6). Thank you for pointing out that not all distributions are created equal.
—Mark Bishop [email protected]
I am writing in response to Dave Lutz's ([email protected]) letter on removing files and security. In it he discusses security issues presented by IRC and lynx, along with “hacker flags”.
He discusses how IRC is a security risk. IRC has been around for many years, and is the main form of chat used by people on the Internet. The act of being on IRC does not become a security risk. He cites the trading of “warez” and pirated software as an example. Web pages and other forms of chat are used to spread those, and they are not security risks. He also mentions the “eggbot”. The proper name of this is “Eggdrop”, and as one of the developers of this program, I can assure you that, if properly configured, the bot poses no security threat.
His letter also mentions how lynx can be used to bring in malicious files. Any program that can transfer files, be it ftp, mail, lynx or IRC, can be used to bring in files. Disabling them is not the proper way to secure a system, as this only hides the problem—it does not solve it. Several key issues should be looked into when securing a system.
One of the main issues would be setuid root programs, that is, programs which run under root permissions. Several distributions, especially older ones, come with setuid root programs which can be used to gain root access on the system. These include SuperProbe, xterm and others. The command find / --perm +4000 can be used to list setuid root programs. If a program doesn't have a specific need to be root (try chmod a-s program and run it as a normal user), you should probably remove the setuid bit to be safe.
Several security lists are available which detail security problems as they are found, one of the main ones being Bugtraq. To subscribe, send mail to [email protected] with “subscribe bugtraq” as the subject.
Once again, security through obscurity can work for only so long. If the person wants into your system, the methods described by Dave will not prevent them from getting in.
—Jason Slagle [email protected]
There is an error in your July article entitled “Encrypted File Systems” by Bear Giles. On page 67, in the first column the following commands appear:
mount /dev/fd0 /mnt -text2,loop,encryption=idea mount /dev/fd0 /mnt -text2,loop,encryption=des
They should read:
mount /dev/fd0 /mnt -text2 -o loop,encryption=idea mount /dev/fd0 /mnt -text2 -o loop,encryption=desThat's it. You people do a great job. Keep it up.
—Steven J. Hill [email protected]