Letters

Regarding the Golden Age of Linux Journal

After having seen in the past few issues of LJ readers who are disenchanted with the magazine's content, I feel I must give counterpoint.

Much like Mr Silverton's setup in the January 2007 issue, I began using Linux in 1995, failing first to install Red Hat and settling in on the astounding Slackware distribution from untold numbers of hit-or-miss floppies. Likewise, I used Linux in a rather amateur fashion for a number of years. At that time, I subscribed to Linux Journal but found only one or two articles that rapt my attention. Later, I started a Linux-based company, designing Linux-based solutions for a number of types of businesses. Even after becoming a Linux “professional”, I was no more drawn to the overall quality of Linux Journal.

For many months now, however, I have been simply astonished by the quality of most every article. I now read the magazine cover to cover. Even the articles that are “beneath” me often have a gem of information that I may find intriguing.

Even more appealing is the uncanny timeliness of recent issue themes. LJ has never really been bleeding-edge. After all, it is a monthly publication. Recently, though, I have had several issues in a row that have matched very closely my personal contemporary interests, and I have been delighted by the breadth of your coverage. You cover amply the needs of both the amateur (which I am, with respect to more desktop-type issues) and the professional (which I am, with respect to the development and server issues).

In summary, I have never, in ten years of subscribership, been so happy with the diversity, scope, depth and accuracy of Linux Journal. Keep up the excellent work.

—
Sean C. McCord

What's a Few Zeros between Friends?

Quote from Linux for Suits, February 2007: “The 15mbps they reserve for their Internet service is less than 1% of that capacity.”

15 millibits per second really is pathetic.

—
Ian Stirling

Obviously, that should have been Mbps.—Ed.

What's a Few Letters between Friends?

In “White Box Phone” [February 2007], you said that the OpenMoko phone came from Funambol. I think you meant to say FIC.

Funambol is the company behind SyncML and push mobile e-mail.

—
Bill Weinberg

Someone Else May Have to Decipher Your Code Someday

I've already made comments about Dave Taylor's column, and you know I'm behind his doing this column and what it teaches. My comment is not a criticism of that column but what I hope is viewed as an addendum to it.

The urge to put an entire sequence of operations on “one line” (chained together via pipes) is noble and quite understandable when one is “in a hurry” or doing something “quick and dirty”. It isn't the wisest of choices for beginners or “semi-production” work.

Yes, this is arguable, but I'd like to reference something else in your magazine, page 14, “They Said It” [February 2007], the third quotation, “Coding up the simplest thing....”

Over time, I've learned that as “The next guy”, trying to read and understand code written by “the guy before me”, that breaking a sequence of steps like this down into the constituent parts makes maintainability and supportability significantly easier.

While the original state is “easy enough” to work with, especially for a demonstration, in the work place, we should encourage everyone to remember “the next guy” (and there will always be a next guy) when we write scripts.

I would, in an environment where this might be used by someone else and not run every five minutes, rewrite this script with these additions:

# Comment this line out for debugging.
trap "/bin/rm -f $Tmp1 $Tmp2 $Tmp3 $Tmp4" exit


Tmp1=/tmp/tmp.1.$$
Tmp2=/tmp/tmp.2.$$
Tmp3=/tmp/tmp.3.$$
Tmp4=/tmp/tmp.4.$$

grep 'google.com/search' $ACCESSLOG | \
 awk '{print $11}' > $Tmp1

< $Tmp1 cut -d\? -f2 | cut -d\& -f1 > $Tmp2

sed 's/+/ /g;s/%22/"/g;s/q=//' < $Tmp2 >$Tmp3

sort $Tmp3 | uniq -c > $Tmp4

sort -rn $Tmp4 |  head -5

Yes, this seems wasteful. (I mention the five minutes above to point out that if run more than, say every half hour on a moderately busy box, the methods may need to be adjusted.) Many people will point out a number of issues with this method of operation, and one of them is going to be “writing to disk so much wastes a lot of time”. Yes, it does. I've never been a member of the “save the PID” crowd or the “make it hard to troubleshoot for job security” group either. Writing to disk a lot is much cheaper than the rate I get paid. If it takes an extra 20% to run, who cares as long as it gives the right answer reliably?

Using the sequence of steps I show above, the people who have to figure out what the he** I did can quickly get to the root of the problem. Now, all they have to do is comment out the trap line and look at the temp files' output to see what is happening and how to fix it. The trap line executes the rm of the temp files when the script exits (that is, it cleans up after itself). They also can see the flow of the process better and gain quicker understanding.

Thank you for the great articles and helpful insights!

—
Michael C. Tiernan

Now You See Them, Now You Don't

In /var/opinion [February 2007], Nick Petreley suggests tactics to move further into the mass desktop market. Among other things, he says, “don't remove features”.

I heartily agree. KDE has seen a steady stream of added features, but unfortunately some old ones were removed. It is no longer possible to move a maximised window partly out of the viewport. You have to de-maximise it first, just like Microsoft. Nor can you double-click on the bar at the top and reduce it to just the bar, as used to be possible. Konqueror no longer allows you to right-click on a file and see an option to delete it, merely one to move it to trash. (Fortunately, there is a way to do this: select the file and press Shift-Delete.)

Presumably, the strategy is to converge on Microsoft—up to a point. If desktop-level Linux becomes identical, why should people buy it rather than the ready-installed Microsoft product?

When Microsoft moved from command-line (MS-DOS) to Windows, Linux expanded in the same direction, retaining the old facilities. This is part of its strength. The same logic applies to desktop Linux: if we remove features, we play to lose.

—
Chris Trayner

As far as I can tell, you're right about moving files to the trash. But, I just moved a maximized window partially out of the virtual desktop, and I also shaded (reduced a window to its title) by double-clicking on the title bar. I obtained KDE 3.5.6 from a KDE mirror and run it on Kubuntu 6.10 AMD64.—Ed.