return to first page linux journal archive
keywordscontents

Letters to the Editor

Japanese Word Processor

I have been a subscriber to your fine magazine for several months now, and each month I look forward to receiving the newest edition in the mail. Each edition is better than the last, and you folks always seem to cover issues I need more information about right around the same time the magazine arrives (are you psychic?). Perhaps the following might be of interest to your readers.

Would you like to work on an exciting project? There is a Windows application, called JWP--a Japanese Word Processor. This package was written by Stephen Chung, and as a GNU product it is freely distributable. I've used it extensively over the past few years, and it is a great package.

This project will never get off the ground without volunteers; therefore, I invite any interested X-Windows developer who wants to make a contribution both to the GNU and Japanese-speaking communities, to lend a hand with this exciting project.

The JWP-Port Project home page contains more information on the JWP package as well as the JWP-Port project itself. If you are interested, please visit the page at http://qlink.queensu.ca/~3srf/jwp-port/.

--
Steve Frampton
3srf@qlink.queensu.ca

ispell

I just read your LJ article on ispell. [``Take Command'', February 1998] You obviously like it. I find it a large pain in the ass, and wish I had a normal UNIX spell-checker available on my Linux box.

I have two gripes concerning ispell. First, the word list it comes with is not very complete. I've added 382 words so far and keep adding. One reason for needing to add so many is that ispell (for reasons I have argued with its creator about) insists on trimming ``'s'' from possessives. That means that sooner or later I find myself having to add the possessive form of every noun in the language to my word list. And that's my second gripe: why can't they at least provide the removal of ``'s'' as an option?

The developer of ispell insists that this bug is a feature. I think his feature is a bug. What do you think? Doesn't this bug you?

--
Andrew T. Young
aty@mintaka.sdsu.edu

In general, ispell works well for me, although I do have my own frustrations. I'd like it to have a ``change all'' feature, so if the word is consistently misspelled throughout I only have to tell it to change it once. There are words I think should be in there that aren't: hydroponic and oxymoron, for example. I've added 304 words but many have been proper names (people and products), computer jargon and abbreviations.

The ``'s'' is a feature in the spell checker I have at home too; at least for words I am adding to the dictionary. I too find it annoying, but not as much as you evidently do.

--Editor

Linux

I just read the Letter to the Editor from searoy@aol.com [February 1998] and thought I needed to share with you our experience with Linux at work. First of all, you do not have to become a programmer to run Linux and many of its applications (my two sharpest guys are operators). The Red Hat 5.0 distribution (it costs a whopping $39 US) is very easy to install on Intel computers and even runs on Sun SPARCs and DEC Alphas (you need to be familiar with how SPARCs and Alphas work through the boot process). Red Hat comes with editors, databases, compilers, scripting languages (Perl, Expect, etc.), network security tool and several types of servers (web, SAMBA, printer, Novell, etc). Even the updates are easy to get off of the web and install on the computer.

The second thing we like about Linux is that it is supported by programmers, hackers, engineers and users around the world, so problems get fixed fast. I recently read an article in BYTE magazine about the bug in the Pentium processors, and the only OS to have a fix out was Linux. (The bug is in a piece of code that will try to stuff a 64 bit word into a 32 bit register.)

The third thing we like is that we learn how the computer, operating system and applications work, which has a snowballing effect. Each time we've solved one problem or gotten a program working it has helped on the Solaris and HP-UX systems we have. Last, you can get into Linux cheaply. Find an old 386/486 with 16MB of memory and at least 300 MB of disk space and you have a computer system for experimentation. Kevin (one of my operators) built a firewall for the local college with a 386/33 that they had in the scrap pile, and we run all of our network monitoring tools on 486/66 ASTs that were headed for salvage.

--
Bernie
Morinbm@aol.com

``A Partner's Survival Guide''

Brilliant! Absolutely loved it--``The two are merely coincidental.'' This glimpse into a hacker's life is a big, unattended part of the mainstreaming of Linux. [``A Partner's Survival Guide'', Telsa Gwynne, February 1998]

--
Arnim Littek
arnim@digitech.co.nz

September 1997 Issue

The article by Andy Vaught, ``Introduction to Named Pipes'' [September 1997] contains an error. Near the bottom of the first column on page 32 is the following command:

mkfifo pipe;  ls -l pipe1;  cat < pipe
The above is worthless, as pipe should in fact be pipe1. This error caused me no trouble, but it was not intended for me, rather for someone who has paid money for LJ and expects to learn and trusts LJ to be accurate. Now, ``No finger pointing between LJ and Vaught''. I would just like to see the guilty party stand up and apologize to ALL the readers for the frustration they have caused someone trying to learn. As a publisher, you have an obligation to ensure that there are no errors. And, please, no excuses.

--
August Grammas
ursa@cris.com

You are correct: the commands should have read pipe1 in all three cases. You are also correct that LJ has an obligation to publish technically correct information. We are, however, not perfect. We would like to be, and we have at least four different people look at each article. We are continually surprised when mistakes like the one you mention get past us. The ultimate responsibility for mistakes lies with me--my policy has always been ``the buck stops here''. I am sorry for any frustration caused by these typos. Yours was the only letter I received about this error and the September issue is quite old--perhaps even the newbies were able to figure out the right way to give the commands.

--Editor

Oracle In-house Linux Port? Bah!

Your piece on databases for Linux (``Databases'', February 1998) mentions a stealth, ``in-house'' port of Oracle for Linux, which we've apparently had ``some time.'' And Oracle refuses ``to sell or support it''.

Where do these rumors get started? I've worked for Oracle for two years, and have been a Linux-head that whole time. If a version of Oracle written for Linux existed, I'd have noticed.

Oracle7 for SCO certainly does run on Linux under iBCS, and quite well. Maybe a misunderstanding of that fact somehow started the rumor.

No one here at Oracle has ever heard of an actual Linux port. If one does exist, perhaps it's being used to help reverse-engineer those captured UFOs at Area 51. That might explain the super secrecy.

--
Steve Abatangle, Oracle Corporation
sabat@us.oracle.com

At the 1994 Uniforum Conference, a man wearing an Oracle badge walked up to the Linux Journal booth and introduced himself as an Oracle developer to our publisher Phil Hughes. This man told Phil he had a working version of Oracle on Linux. Unfortunately, Phil has forgotten the man's name, though not the event.

--Editor

Bleeding Edge Articles?

I am a long-time reader of LJ and have always been very pleased by your articles. Linux is finally getting some of the respect it so richly deserves, and it is great to see such a fine publication supporting the cause.

I do have a question, however. I'm an occasional kernel developer and long-time Linux user and I would like to see more articles (maybe one per issue) on Bleeding Edge Linux projects and ports. Articles detailing such relevant topics as MacLinux (Linux for Macintosh/m68k, http://maclinux.wwaves.com/), Linux/PMac, GGI, and other works in progress would definitely be a boon to your readers and allow for more people to become involved in these experimental projects. Now, in the true Linux fashion, I'm not going to suggest you do things without volunteering myself in the process and I would like to know if you accept articles from the user world and (if so) to whom can I send them?

--
Joe Pranevich
knight@wave.lm.com

We have had articles on the Macintosh and Linux in issue 31, issue 37 and issue 45. Reuven Lerner talks about CGI each month in ``At the Forge''. I agree it would be nice to have bleeding edge articles each month, so we print them as we can. Yes, we do accept articles from the user world. Please send your ideas to ljeditor@ssc.com. Author information can be found on our web site at http://www.linuxjournal.com/wanted.html.

--Editor

BTS

I think the first ``corrupted terminal'' answer on page 72 of the February 1998 issue needs some work:

  1. cat was NOT ``designed to handle ASCII files.'' cat conCATenates files to standard output (hence the name). It concatenates one or more files from argv, or if argv is empty, simply from stdin until EOF. There is nothing in cat or its documentation to suggest that it was designed for ASCII files, or could not handle binary files.

  2. cat does not ``interpret a lot of the binary file as control sequences.'' The terminal emulation supported by the console code (or an xterm) interprets any control sequences it is sent via any program that writes to it. These can come from any program. The terminal emulator is supposed to do this. It's why things like Pine and Emacs work, for example.
Here's a question for you: how many Linux seats are there, and how many new seats a month are there? Any idea?

--
Ron Minnich
rminnich@sarnoff.com

No idea. Vendors don't like to give their sales numbers to their competition, and those numbers would probably be the most accurate count. There is the Linux Counter, but not everyone knows about it or takes the time to register so its count is way low. If you haven't registered, go to http://counter.li.org/ and do so--now!

--Editor

Red Hat CDE

I recently saw the review of Red Hat CDE by Don Kuenz in the January edition of LJ, and decided to buy the software. However, on filling in the order, I saw that the January price for CDE was 25% higher than the price advertised in LJ. I think it's unfortunate that the information in the published review was out of date before I even received the magazine. I am also disappointed that Red Hat, while obviously benefitting from the PR of this review, are not prepared to honour the advertised price.

--
Dr. Steven Bird
Steven.Bird@edinburgh.ac.uk