Letters to the Editor

I enjoyed [Matt Welsh's] article Thou Shalt Not Use MS-DOS, in the latest issue of LJ. I think what [he] said is true, but misses a point Phil Hughes has alluded to in an earlier editorial.

I think the reason that Linux; and free software in general; is so great is that it's an incredible consumer value. According to the dictum, “what do you put in to it, what do you get out of it, and what about support?”, most commercial software falls right down flat: it's expensive to buy, you often can't try it out in advance, and if it's broke you're at the mercy of the vendor, who is the only one with the source code.

If free software was free but you didn't get the source, it would be terrible; you could try it in advance, but when it finally did break, no doubt in the middle of an important project, you would be stuck. On the other hand, if payware came with source code, it would be a much better deal: when it broke, you could ask the manufacturer to fix it, or you could call your local hacker and negotiate a price based on how soon you need it fixed. Of course even payware with sources still doesn't let you test it all in advance of buying it, but it's a much better consumer value.

Consider the auto industry for a moment. It will cost you about as much to service a car over its life as it cost to buy in the first place. Would you pay double the price up front in order to avoid paying for service down the road? Yet this is just what software houses do, since the only way they can pay for customer support is by charging more for the product. It's no wonder software support is so lousy! This is why being able to hire your own hacker is so important.

And thus a common point of confusion about free software: it is free in the liberty sense, but it isn't free in the money sense, since programmers have to eat just like everybody else. What free software is is a much better deal. Even if you had to pay a little bit for it, it would be a great deal. Imagine that you had to pay Linus $1 every time you got a new major revision of Linux. Linux would still be the best consumer bargain going. And, indeed, many Linux users pay $40 and more for a convenient distribution.

And that, in my opinion, is what makes Linux great. It's cheap to test drive, it is high quality, it's cheap to buy, and when something breaks you can decide how much it's worth to you to get it fixed, how fast, and you can get it fixed. What do you put in to it? A few dollars. What do you get out of it? Professional-quality software. What about support? The best.- -David Keppelpardo@cs.washington.edu

1. Thanks for the Introduction to the GNU C Library and Programming the VT Interface columns in LJ. The notes on signal() were welcomed.

There have been a number of columns connected with the FSF+GNU. LJ should add a few notes to these columns on ways to support FSF and GNU. In your GNU C Lib article, you wrote “I do not know if [the reference manual] is being published on paper”. My copy of the GNU's Bulletin is at school, but I think it is listed. In any case, I'd like to see LJ more explicitly mention support of FSF (although you did mention FSF books in the sidebar).

Just to be completely clear: I'm not suggesting that LJ “beg” for money for FSF; only that LJ help new users become familiar with the project. (My only “official” connection with FSF is in support of GNU awk. I receive no money from FSF.)

2. There is a common thread in LJ and on newsgroups about the “free-ness” of Linux (Deutsch's letter in #2, Kempen interview in #3, Welsh's columns). Users forget (or are not aware of) the importance of various projects in the success of Linux. LJ could educate users in future columns on GNU software, C-Kermit, ghostscript, etc.—Darrel Hankersonhankedr@mail.auburn.edu

LJ Responds:

1. Read What's GNU for this month. While we won't run an editorial like that every month, we definitely agree that the FSF has played a key role in the development of Linux, and is worth supporting.

2. We have published several tutorials. It is true that quite a few have not been for FSF software, but that is probably because FSF is better known than some of the smaller new utilities that people express an interest in writing tutorials for. We encourage people who are excited about some package they work with to send e-mail to ljeditor@linuxjournal.com and discuss writing a tutorial for the package.