Letters to the Editor

I just read your PageSat article. We've been waiting for our system for three months. Unfortunately if you dial the main PageSat number you get another number to call, 415-493-9592—which says PageSat is history, gone, no longer in service. (It was a good article anyway.)

It's too bad. I was eagerly waiting for this, since our Newsfeed is very poor. Do you have any other suggestions for an alternate newsfeed? Ed Longstrom edl@rock.spectra.net

Rich Myers, who wrote the article in question, recently sent us e-mail on this very question:

NCIT (aka Pagesat) has ceased operations effective 12/4/96. Norman Gillaspie is starting a new business, PC-Sat, to continue the satellite newsfeed for both current and future customers.

We should have a web page on-line by the time you read this. The location of the web page will be: http://www.pc-sat.com.

Would you please pass on this info to the interested parties?

Thanks in advance for your assistance! Rich Myers rich@webworks.net

First let me say that I enjoy LJ very much and get a lot out of every issue.

I'm considering using Linux in some commercial installations and have held off because of a glaring omission in the networking subsystem. I can find no evidence of Linux support for the NFS lockd and statd network lock managers which are required to guarantee exclusive file access over NFS. Our application support software uses a shared NFS mounted file system and must have this to work. I believe the lack of this capability will impede Linux's acceptance in enterprise computing environments that depend on robust networking support. Other than that, it seems Linux offers the most bang for the buck for Unix computing and networking. Toby L. Kraft tkraft@krafte.com

I read with interest your editorial (“The Politics of Freedom”, Phil Hughes) in the October 1996 Linux Journal, #30, and thanks for webbing Richard M. Stallman's letter for background. Will you entertain a comment from an outsider?

It's too bad the word “political” has such a bad flavor. Virtually everyone wants to make things better in his community (the polis) and tries to persuade others to cooperate in doing so, i.e., acts politically... the catch, of course, being how to decide what is “better”.

I don't use Linux. I make a living writing commercial software for commercial platforms. I have also written, published, and placed in the public domain a few utility programs. I did so for at least three reasons I'm consciously aware of (not necessarily in priority order):

At least one of these motivations—the desire to benefit my fellows—RMS also feels. Perhaps he feels the others as well. Nothing wrong with any of 'em.

You suggest RMS “feels that everyone has to believe exactly the same things he does.” I read his letter somewhat differently.

First, let's dispose of the myth of “free” software. They ain't no sech animal. Even if you obtain it without exchange of valuta, you must still invest time and skullsweat to use it—both scarce and valuable commodities, perhaps more so than money. So, any system competes with all others for your investment, whether cash changes hands or not.

Perception is reality. If Linux and GNU do not share a name, they are more likely to be perceived as distinct rather than variants of the same thing (one of RMS's points, and I agree). If they are distinct, they compete.

And “free” software already competes with the notion “if it really was valuable, you'd be charging us for it”—as Heinlein pointed out, why do you think they take collection in church? Added competition between Linux and GNU can't help but drain both against the “common enemy”, which I think is what he is saying.

I'm sorry to have to disagree with Linus Torvalds, but it does matter what people call Linux, and this is why. Since Adam's day, to name a thing is to have power over it... and to be trapped by its paradigm.

(Insert clever pun about “Paradigms Lost” here. Oh, never mind.) That's my two pfennigs, anyway. Hope you enjoyed it.Davidson CorryDAVIDSCO@Attachmate.com

Phil Hughes points out RMS asked for and was offered the opportunity to respond, but so far, we've not heard from him.

The November 1996 issue (#31) contains a very nice article on the OpenGL libraries by Jörg-Rüdiger Hill [“Linux Goes 3D: An Introduction To Mesa/OpenGL”].

In that article it is stated, “Mesa can be called from C and Fortran routines”, which is true, but I felt obligated to let your readers know OpenGL and MesaGL are also callable from Perl 5. Stan Melax has an OpenGL module which is described at http://www.arc.ab.ca/vr/opengl/.

I have just installed it and the requisite MesaGL library and headers on a Solaris box, and I must say it is plenty impressive (the clip script is only 116 lines long—if you don't count shebang and comments, it is only 70 lines of Perl!). The examples/ directory of the OpenGL module for Perl also includes a tk_demo script that uses the Mainloop() event handler of the Tk extension to Perl (whereas the other demos do their animation with built-in Perl loops). I know Perl's Tk400.200 extension compiles on most major Linux distributions, and I am fairly certain the OpenGL module does as well.

Thanks for a great issue on graphics. Peter Prymmer pvhp@lns62.lns.cornell.edu

A timely note for our focus on Perl this issue. Thanks!

Last week I was int eh field, working on a piece of equipment controlled witha computer runnin iRMx, a “unix-like” Intel product from about 1980. I needed logfiles from the machine. Using all of my clues, I was unable to write on the 3.5" diskette without first using RMX to format it. I turned the equipment back over to the customer and went away witha n unreadable diskette containing data the factory needed right away. Certainly you cannot read it on a Windoze box!

Next I tried it with my lInux system (Debian 2.0.24). While Gates's garbage reads only one kind of file system, maybe RMX is readable as one of the many file systems Linux will support...Load up the modules and hack away: Sorry, no luck!

Then I remembered my LJ issue 32, which I picked up on the way to the airport. Out it came and straight I went to Sam Chessman's excellent article, “What is dd?” I ran Sam's “Example 1a” to make a disk image. Voila! I have a file? Mov e the image file to the FAT partition, kill Linux and run Windoze to e-mail it home, along with instructinos for the Unix guys at home to run Sam's “Example qb” to recreate the diskette.

They tried it but did not succeed (you know, th edd supplies with SunOS may not be as capable as ours). So I resolved to hack on the image file. As it turned out, using CTRL-K a few timmes in Emacs turned it into radable plain text log files! —D.W. Wieboldt dwiebold@Aus.etn.com

Thanks for letting us know you found the article useful (and thanks to Same Chessman for writing it in the first place). We love to hear about Linux being useful (even as frequently as it is).

Regarding “Stop the Presses” in Linux Journal #32, December, 1996, two items:The first concerns SCO UNIX “Give-away”. I bit, paid, the 20-some bucks and got the CD. It was primarily for work; I wanted to do a demonstration of what Unix could do. The IBMers are quite leery of Linux—software with no huge corporation after thier money. They just don't know how to relate.

Anyhow, I got a PC loaner from network services and was going to load SCO up and I couldn't get to first base! Turns out, SCO doesn't seem to believe IDE hard-drives are used. The installation couldn't see I had 1.6 gigs of space—just said I didn't have a hard drive and so it couldn't proceed. Yeah, I'd be worried about that company if I wanted to see Linux thrive (giggle...).

Afterwards, I brought in Red Hat 4.0. It installed without a hitch. The only thing holding Linux back at my work site is the poor token-ring support. I just haven't been brave enough to even consider trying to develop new drivers.

The second item is a question: Have you heard of anyone running a UNIX Oracle client under LInux? Or getting access to Oracle data through ODBC or other means? Just curious.

I do appreciate your magazine. It's fun and provides a lot of useful information. Keep up the good work. —Dean Siewertdsiewert@execpc.com

Being fair, we plan to subject Free SCO OpenServer to a review in an upcoming issue of Linux Journal. It's real “UNIX” so it'll be lots better than Linux, right?

As I've done earlier in this month's letters. I refer your question regarding Oracle clients on Linux to the LJ readers.

I have taken what I consider to be a very critical look at the SCO offering and I would like to convey the following observations:

In this MS-day and MS-age, it is imperative alternative operating sytems be encourages to grow and flourish. One of the most effective means of giving a boost to alternative operating systems—and Unix specifically—is to make them truly accessible to the common user. To this end, SCO is making a significant contribution.

And almost as quickly the contribution ends. SCO's offering is sexy, slick and almost inaccessible. It provides all of the Unix tools and almost none of the more necessary understanding. Having started life as a DOS user, I view SCO's contribution as cold and disenfranchizing as Apples' Systme 7 or Windows 95. It has all of the functionality while offering no real understanding of the contents of the operating system's “black box”.

Linux, my first exposure to Unix, provided the all-important tool for true inderstanding : accessibility. I am able to lift the hood, kick the tires and generally test drive the OS functionality in a hands-on fashion. In my naive and simplistic view, this is the best enfironment for learning. Let SCO offer but know that Linux provides! —William B. Meloney VII bmeloney@mindspring.com