Letters to the Editor
In the article on ESDI drives, the URL for the MCA page should be: http://glycerine.cetmm.uni.edu/mca/. I love your magazine. Three articles in this issue answered questions I had been having. Keep up the good work.
I have read articles saying many wonderful things about Linux and I believe most of them to be true. Unfortunately, the extent of hardware support that some of these articles claim is not a reality—or at least does not seem to be when it comes to Compaq equipment.
I have just spent the best part of a day searching the Internet by various means, including various search engines, trying to find drivers to support the embedded NCR 53C710-based SCSI controller in a Compaq ProLiant 2000 and also drivers to support a Compaq SMART SCSI RAID array controller. Result: nothing, except a lot of stress.
Please can someone help me (and the many others who I have encountered looking for these drivers). Linux claims to support quite a bit of hardware—please extend this support to include some key Compaq server items.
—Graeme Nelson firstname.lastname@example.org
I'm writing you after seeing one too many odes to the glories of the Open Source movement. I have a serious problem with the whole Open Source bandwagon due to the fact that Open Source is almost solely about making free software palatable to business—a segment of society which has played a largely non-existent role in the development of free software. Business has done nothing to make the user and programmer community at large more aware of the benefits of free software. I feel the primary benefits are individual and social freedom.
The June article by Eric S. Raymond, “Open Source Summit”, is a good example of the fundamental emptiness of the Open Source movement. The O'Reilly conference report struck me as being more about how Larry Wall, et al., can strike it rich than about how the lives of users and programmers can be enhanced through free software. I have nothing against people being financially compensated for their labor, but being financially compensated for one's labor has always been a secondary or even irrelevant consideration in the free software movement and rightfully so.
The most appalling notion implied in the rhetoric of the Open Source movement is that we, those of us who use/write/support free software, have to change our ways and adopt a more corporate mindset if we want free software to be successful in the real world. This is manifestly ridiculous. If free software hadn't already proven itself thoroughly in the real world, there wouldn't even be an Open Source movement. In fact, I think that free software and the free software movement have proven themselves to such an amazing degree that the corporate world now wants to find a way to squeeze a buck out of us. Again, there is nothing wrong with making a buck, but don't you dare do it at the expense of my freedom.
Unfortunately, free software developers are not a major source of advertising dollars for LJ, so it is not likely that LJ will be publishing alternate views to the Open Source camp anytime soon. That apparently being the case, I would suggest that if LJ readers are interested in an alternate view of the free software movement, check out, for starters, Richard Stallman's article “Why Free Software is better than Open Source” at http://www.gnu.org/philosophy/open-source-or-free.html.
—Shawn Ewald email@example.com
While I disagree with your stated beliefs, I'm always happy to publish alternate views—I have done so in the past, do so now with your letter and will do so again in the future. While it is true that LJ does not receive advertising dollars from free software, we put free software items in the “New Products” column and publish reviews and tutorials of free software.
Linux Journal strongly supports “freely available” software and the Open Source movement. This is one reason we chose the Debian distribution to use in our office.
By the way, I see no reason for you to have singled out Larry Wall as looking for a way to “strike it rich”. Perl is free and Larry is most definitely not a money-grubbing type of guy.
My thanks to the numerous people who've written in response to my article in LJ #50, “PPPui: A Friendly GUI For PPP”. To anyone interested in more features—especially anyone who relies on single-use passwords—please check http://www.teleport.com/~nmeyers/PPPui/ for features added to PPPui since the article was originally submitted.
—Nathan Meyers firstname.lastname@example.org
I just wanted to mention that we have a couple of people in our lab who provide Sybase connectivity (server running on Irix) in their Linux programs. They told me that it is fairly easy to get it working using freely available C code downloaded from the Web.
In Issue #50's article, “PPPui: A Friendly GUI for PPP”, Mr. Meyers notes that PPP does not have a good user interface: the only way you know if your connection succeeded or failed is to check the process list. Mr. Meyers offers a solution.
There's actually a simpler way than his program: direct syslog to Console 9 as described in an earlier issue of Linux Journal, and enable logging in chat using --v. Then, just hit alt—f9 to view your syslog console, and you can watch the progress of chat's attempt to connect. Failed connects show up as Alarm, exit or hangups; a sluggish connection can be observed as pppd sends EchoReq's out. Disconnects show as hangups. This live syslog is also invaluable when debugging your chat script.
So there I was, reading the review of Caldera OpenLinux (June 1998), and the reviewer, Sid Wentworth, writes:
Caldera, by default, uses the Looking Glass Desktop. Not being a desktop sort of guy, I am not particularly excited about it, but, if you want a desktop, it seems adequate.
Great! He's not a desktop sort of guy. Why is a “non-desktop” kind of guy getting paid for reviewing anything other than AWK scripts?
It's hard to take seriously reviews that completely leave out subjective comments about functionality a reviewer doesn't really have an interest in, or more to the point, his audience does have an interest in. Don't care about it, don't review it—simple! The rest of us, though, might have been interested in the state of this product's constantly evolving user environment, but the heck with us. You need better writers, which shouldn't be too difficult. At least the Windows techies take apart the toys they review.
There is a lot more to Caldera OpenLinux (or any Linux distribution) than the desktop. Also, desktop choices are available with any Linux flavor. In the case of Caldera OpenLinux, I found their proprietary desktop to be unexciting, but I also found that to be unimportant over all.
I could have easily written a 50-page review of this product—so many capabilities are there to discuss. For example, each language translator could have been reviewed. Even if I just looked at GUI capabilities, a comprehensive review of desktops would need to include comparisons with XFM and other free file managers.
Don't get me wrong. If Looking Glass had been exciting, I would have talked more about it. It wasn't and I don't think it matters. What did matter, as I said in the article, was StarOffice which, by the way, is its own desktop.
I believe that the “Best of Tech Support” column in the May 1998 LJ contains a small error. Regarding the question of Linux's behaviour when a file system is infected with an MS-DOS virus, Chad Robinson states that because Linux spreads file system meta-data more evenly throughout the physical disc, “random potshots” are more likely to cause corruption of the meta-data. This is misleading, in that if the amounts of meta-data were the same, the probability of a potshot hitting the part of an MS-DOS file system containing meta-data is equal to that of it hitting more evenly distributed meta-data on an EXT2 file system. “Concentration” does not affect the probability, only the amount of meta-data.
Simon Maurice in the June 1998 LJ criticizes Red Hat v5.0 for perceived shortcomings. Is he truly serious in saying that Red Hat's RPMS is a “Microsoft-like effort”? I think the package management system is one of the features that puts Red Hat distributions at or near the top of the pile. Where are all the bugs he alleges? I've found a few wrinkles, but nothing I'd call a serious bug.
As to the alleged problems with the actual distribution, I have run both the v4.2 and the v5.0 distributions as “official/supported” releases. I haven't applied any patches—yet my system is so stable it hasn't crashed since March 1998 when RH5.0 was installed. (Which is more than I can say for my Windows NT4 system.)
However, I do agree that Red Hat should be more polite when it comes to customer (e-mail) support. I asked several questions about the v4.2 distribution and was curtly told the questions were outside the support structure and to try the mailing lists. I wasn't happy with that answer, but it taught me to go to the documented sources first.
For $50 or so, Red Hat's distribution is by many orders of magnitude easier to use and install than the first distribution of Linux I bought back in 1993 (Trans Ameritech, v0.90 kernel). Compared to commercial operating systems (e.g., SCO, Windows NT), there is no comparison, whether for value for money, stability or user support. Where else but in the Linux community can you get bug fixes (if they are needed) so quickly?
I just received the June 1998 LJ (love> that airmail delivery folks). In “Best of Technical Support”, Chad Robinson and Pierre Ficheux both gave useful answers to “how do I back up NT and Linux”. However, you should know that the AMANDA package (Advanced Maryland Automated Network Disk Archiver, http://www.amanda.org/) works well with NT as well as with just about any flavour of UNIX. The current release version is 2.4.0 (with 2.4.1 in snapshot development form), and people all over the world are using it to back up all sorts of machines.
—James C. McPhersonjcm@kalessin.humbug.org.au