return to first page linux journal archive
keywordscontents

Letters to the Editor

All of the following letters are in response to the questionnaire we sent out to Usenet in December, 1993. The comments below have been used with permission of the authors. Some have been edited because of space considerations.

In the area of molecular biology, it appears many of the preferred programs work optimally under a Unix environment. Since most scientists lack a Sun SPARCstation at home (some may), or even at work, it seems that Linux offers the ability to easily bring that sort of computer power to the home. I've seen that some of the molecular biology utilities have been ported to Linux. Some of the examples are:

ACaDB (A Caenorhabditis elegans Data Base) -nematode database for this evolved:

AAtDB (An Arabidopsis thaliana Data Base) -Plant database and other similar databases are evelving to serve as databases to accumulate information on these model biological systems.

GDE (Genetic Design Environment) An X-windows based working environment to run molecular-biology based applications. It allows for the easy entry and visualization of nucleic acid (DNA, RNA) or amino acids (protein) sequences. It further allows analysis of these sequences by programs such as:

If a local database (like GenBank) is not available on a Linux system, GDE is designed to format an e-mail submission in the proper format to one of the national databanks available on the Internet.

PHYLIP - phylogeny analysis programs and others wihch may easily be linked to GDE.

I'm not sure if many others have been ported to Linux, but it seems quite feasible. I see new Unix versions of useful programs being offered and often wonder if anyone has ported them to Linux. The Linux Journal should be able to facilitate learning of this information. I feel Linux brings the most in easily obtainable computing power to the scientific workbench.
Gerard R. Lazo,
Southern Crops Research Laboratory, College Station, TX

Asked about other topics that should be covered in LJ, Kenneth listed

Kenneth Osterberg,
Helsinki, FINLAND

We should have had you design our magazine. See our Debian column, comparision article on Linux, DOS/Windows and OS/2, our programming tips column, our FSF column and even our interview with Linus. Guess we are on the right track.
Editor

Good day!

I'm highly enthusiastic about the existence of Linux Journal. Since I started using Linux in February of 1993, I find my current magazine subscription is becoming totally irrelevant, since the editors are violent MS-DOS/Windows bigots.

About the software wanted column

This would be useful. What do people want for linux, in the way of software, additional drivers, and so on. It should be possible for a person's e-mail or actual address to be printed with their suggestions so people can either point out possible software, or ask about details so they can implement it.

About the New Linux-related products column

Definitely. For me, the major point of subscribing to a magazine is for the reviews, and for tutorials. I also love opinion pieces.

Andrew suggests the following additions to LJ

Alpha and beta software: Not reviews, but at least mentions of it. "Andrew Kuchling's FOOGOL compiler is currently in early beta. Word-of-mouth says it's not bad, although some of FOOGOL's more complex constructs produce buggy code at this point." This is so people know what's on the way, and so corporate types can see that Linux is alive and well.

Also, keep us up to date on what's going on in other software areas, like the DOS/Windows and commercial Unix worlds. We don't want a 6-page review of Windows 4.0, but simply mention new programs that we should know about, and the reaction to them. (Was it buggy? Is Windows finally usable? Can Wine be rewritten to run 4.0's programs?) The same applies to new Unix versions; have they got any new ideas we can apply to Linux? Did they encounter any pitfalls?

Another feature might be "The Cutting Edge", on using Linux in new and experimental ways. For example, to simulate a parallel processing system in software; as a server for a database; running it on a Pentium; using Linux as a cheap firewall system; object-oriented databases for Linux; using a Linux network to test networking hardware (according to Paul Tomblin, this is what Gandalf does). This is to show that Linux makes inexpensive PC-compatibles competitive with pricy workstations for demanding Unix applications. LJ should also show how Linux/X is better than DOS/Windows, capable of running everyday applications like word processors & spreadsheets.

LJ should not exclusively discuss things from the viewpoint of a beginning user. We also need more advanced articles on system administration, and programming topics (not necessarily Linux-specific ones). Some possibilities I'd personally like to see are:

Optimizing: What does gcc do when optimization is turned on? How can I write code that will optimize well?

Using PC-specific features in Unix: How can I play a digitized sound in my own programs? Display VGA graphics? Play an MPEG? Read a mouse?

Tweaking XF86 for better performance: How can I make it use less memory? What graphics cards are really fast?

Backing up your system: How to do it onto floppies? What tape devices are good?

Kernel patches: What's available? How well does it work?

SCSI: How do I go about moving from IDE to SCSI? Which cards and peripherals are good?

Adding your own system call to the kernel: What do you do?

Texture mapping in graphics programming: This is not a Linux-specific topic, but could be given with Linux-specific examples.

Most of these topics are too big for one article, and some partially duplicate FAQs. That's to be expected; articles should never (well, almost never) say "See the XYZ FAQ for the details". If they're not giving the details, what's the point of the article? (On the other hand, articles should be written for various levels of expertise; some obscure details should be referred to FAQs, then.)

A longer-term idea: It should be possible for writers to include a demo program, possibly a complex one. For example, the texture mapping author might provide an svgalib program that lets the users change colours, and experiment with different noise functions on the fly. The software could be available via mail server or FTP; people without Internet access could order a diskette with the software from LJ. (This would be a large effort, though, and so should be postponed; maybe you could politely tell non-Internet readers that they're out of luck. :) ) This software may not be significant enough for it to be put on tsx-11 or sunsite, so LJ should handle this itself. (Or convince tsx-11 to add a LJ/software directory.)

Also, it's _very_ important to have book reviews. A possible yearly feature might be "The Linux Bookshelf", a listing and discussion of the best books for using and programming Linux (and Unix, more generally); both reference books and tutorial books should be covered, and both Linux Documentation Project efforts and third-party publishers. It would also be valuable to review Linux Doc. Project works as they're released (with allowances, when they're in alpha).

It's very important to also print negative reviews (possibly unsigned, to avoid chilling friendships). Some magazines (such as Compute!) didn't print negative reviews, saying they'd rather use the space to discuss a good program or book, but I think that's wrong. Otherwise, if product X isn't reviewed, is it because the reviewers hated it, or because the magazine has never seen it? There's no need to be cruel, especially in reviewing free products. (Our expectations are higher for commercial software, and people often write free software in their spare time.) But we should know that foo0.99 compiles out of the box and works wonderfully, while bar0.99 requires lots of config hacking and seems a bit unstable once it's running. Reviews are really the most important factor to me; if LJ's reviews section is perfunctory, then the magazine's usefulness is greatly impaired.

Another point: There's so much free software around, and it's updated so quickly that reviewing it all would be an impossible task. But a half page of 1-5 line mentions of especially notable or dangerously unstable software would be nice...

admscr0.01.tgz: A system administration script with menus. Be careful; the "Clean Man Pages" option may destroy man pages for which you have no source. admscr0.50.tgz: A system administration script with menus. Small, but it works nicely.

In short, be negative about things from time to time. Don't hesitate to say "This software is bad.", or "386BSD is better than Linux in the following areas..." If we don't talk about what's wrong, things will never get fixed.
Andrew Kuchling Hemmingford,
Quebec CANADA

About additional topics that LJ should cover, Steve suggests

  1. How to load Linux (SLS vs Slackware)
  2. A tutorial on lilo and all the things you can do with it.
  3. an analysis of how Linux stacks up to the POSIX standards (.1 and .2), BSD interface and SVID (2 or 3, or whatever the latest version is)
  4. Updates on WIN3 emulation and threads support

Steve Zanoni,
BrookField, WI