Posts

Profundity has completely avoided touching my life for the last several weeks. It's not that I have been completely without thoughts, but they just haven't seemed to require expression.

I have been reasonably busy both at work and at home and I expect that it will continue for the next few months. See at work, we have 3 system admins, one of whom is primarily concerned with the PC class machines. We other two maintain the Unix workstations in the department. Well, my partner is going on maternity leave starting this week. So I see myself doing as much of two person's jobs as I can get through each day for the foreseeable future (i.e. several months).

I'm not really worried about it. I'm far more interested in getting my Wacom Graphire USB tablet working on my RedHat 6.1 Linux system. Perhaps that is an appropriate topic for thoughts: the development cycle under Linux.

In this case, trying to get my tablet working depends on the efforts of at least two separate development groups— those working on getting USB support built in to the Linux kernel and those working on adding support for new Wacom devices to the XFree86 Xwindows server.

Since returning from my wonderful Christmas with my tablet I have discovered that the tablet will never work with Windows 95 since the most recent driver from Wacom won't let you select USB under Windows 95, even though there is some support for it.

Also since my return, there have been at least half a dozen development kernel releases for Linux plus at least two versions of a backport of the latest USB support into the stable kernel. The latest of the backports has reached the point where when I watch the raw device associated with my tablet, it spews a ton of garbage whenever I wave the stylus over the tablet. So I know that Linux has found my tablet and is listening to it.

The next task is to see if I can figure out why the xf86wacom driver only reports "tcgetattr: Inappropriate ioctl for device" when I tell it where to find the tablet.

Now your first reaction to this might be to shudder and quake at the prospect of having to keep up with a development cycle that releases stuff every few days. "Do I really have to live on the bleeding edge just to run Linux?"

The answer is "of course not". During all that development I downloaded stuff exactly twice. The first time I grabbed A) the newest stable kernel, B) the current backport of USB from the development kernel, and C) the most recent xf86wacom driver. After I had determined that this setup was not going to work, I sat back and checked on the development cycle every few days. I read the documentation they provided on their web pages and after just two weeks I realized that the USB support had reached the point where it had fixed the problem that I was seeing. So I downloaded A) and B) again. C) had not been updated yet.

There the matter stands for the moment. My tablet isn't working yet, but I have an understanding of what needs to be done and I know who is doing it. I haven't written any drivers myself and I don't expect that I will have to. I'm not sitting on the bleeding edge of the Linux kernel development cycle, though I've been watching the National Geographic special "Survival in Wilds of the Linux Kernel" so I know who's eating whom.

Say what you want about the development of Open Source software, but understand that the information is there for the interested to dig out. And the work is going on under your eyes. If you want you can check up on the builders and make sure that they are using the right mix in the concrete and putting in enough rebar. If Linux or any other Open Source software project fails it won't be because the building collapsed.

Not many thoughts recently. That's because I've been so busy at work that I simply haven't had the will to work on my home web pages. And yes, it is a matter of will and nothing else. There existed plenty of time to do things with my home machine and in fact I did do some homework, but anything on my list of things to do that looked like a project simply evoked a mild dread that prevented me from considering working on it.

But the holiday occurred. (Which by the way increased the workload due to the fact that as of Wednesday noon I still hadn't bought my Thanksgiving turkey.) I took half a day and elminated the top of my dire chore list.

Thanksgiving tends to be a mild holiday for me. Cook a turkey, homemade stuffing, homemade gravy, a veggie or two. Pig out when it is ready or mid-afternoon, whichever comes first.

This Thanksgiving I gave myself an early Christmas present. While doing those dire chores, I made up my mind and bought a CD ReWriter. This gave me something to do while waiting for the turkey to start producing juice for the gravy and when not talking to family on the phone.

But this was supposed to be about working on the web pages. So now that I've blathered all over the map, let's get back on topic:

I am going to like EmbPerl.

Ok, I'm an old Unix user. (Circa 1983.) I've programmed with some level of proficiency in the major Unix scripting languages, sh, csh, awk, sed, etc. (No, etc is not a programming language, it just looks like it. I hope no one ever really creates a language named etc.) Then I wrote a huge backup control program in csh which, despite the fact that it was essentially doing a preset sequence of commands, grew into a multi-thousand line, unmaintainable piece of junk.

Then came perl 4. Aaaahhhh.

A week later I had a 400 line perl script that did 70% of the job and I had redefined the task to elminate another 20%. A month later I had a 1500 line program that did 73% of the job and was looking really ugly. I was begining to wonder if I had made a mistake in choosing perl.

But I liked programming in perl. There's something nice about knowing that there is a kitchen sink available when a bathroom sink just won't do. So I started completely over. I was learning that the perl motto- "There's more than one way to do it"- also meant that my way was the wrong way most of the time.

For example, while learning perl, I would frequently get stuck on some really minor thing, and no change to the code would seem to make any change in the behavior. After a few frustrating hours, I would crack open the camel book. (O'Reily & Associates, "Programming perl" by Larry Wall and Randal L. Schwarz) Therein I would almost invariably find a perl function which does exactly what I was trying to make a huge subroutine do. I got so mystical about it that if I spent more than 10 minutes on a problem, I'd crack the camel expecting to find the answer in Chapter 4.

Back to the backup program. Two weeks of rewriting left me with 500 fairly clean lines of code that did about 95% of the backup. And the rest was human intervention required anyway. Perl was going to stay in my arsenal.

(About this time I discovered the Amanda backup program and all my work was moot, since Amanda was more flexible, more easily extensible and had fewer niggling little problems. On the other hand, Amanda is considerably more than 500 lines of code.)

So awk and sed and to a large extent the shell scripting languages went to dodo-vile for me. I started churning out 10-line, 20-line, 30-line perl scripts for everything under the sun.

I've wandered away from EmbPerl again. I guess I had other things that wanted to get expressed. But I'll close with a final thought about programming in general. It is my belief that the power of a programming language is measured by the ease with which you can do stupid, pointless things in it. Thus, EmbPerl:

[- $tim = time % 10 -]
[$ if ($tim % 2) $]
<BODY BGCOLOR="lt blue">
[$ else $]
<BODY BGCOLOR="red">
[$ endif $] 

This changes that background color of a web page based on whether the time is in an odd or an even ten second block. Really pointless, very easy.

I'm sick of it. How many times have I gone in search of a nice, highly recommended piece of software on the net, only to get caught in an endless round of signup forms? We are beginning to learn the true cost of free software— loss of privacy.

It doesn't matter to me that every one of these sites has a link to their privacy policy. My privacy is gone the moment I have to answer a lot of impertinent questions. Sure I can often check the box that says that I don't want to hear from anybody about this software. But I still have to trust a wide assortment of differing people and policies. And some sites don't make any such promises.

(By the way, this is one of those moments when I really wish I had a digital camera— There was a fantastically cool looking reflection from my closet doorknob that I wanted to share with the world. It is now gone and even if it returns tomorrow, I'm not sure that the magic feeling of seeing that brassy semi-circle of light will still be there.)

And the questions are getting increasingly impertinent! One site was requiring both home and work addresses and phone numbers. I really sure that I'm going to give that information to somebody who needs to have a privacy policy link. Another site insisted that I give a fax number. Right. I don't even have a personal fax number. Ok, I could fake it up with mgetty on my Linux box, but why? This was a piece of software I was downloading. There was no conceivable reason why these people needed my fax number to enable me to test or use their software. They didn't want to fax or email me any registration information.

Sun Microsystems wanted my job title before they would let me download their Y2K scanning program. Why? Is there no one out there running Solaris on their personal machines? Sun would like us to think that Solaris X86 is a viable home desktop system, but I guess that if you are looking for Y2K solutions, you must be doing it for your work.

I'd like to propose a minor revolt against signup forms for free software.

1. If you are downloading a trial or evaluation version or free software and they start asking questions, stop immediately. Add two strikes against the software on your evaluation of that software and consider whether you want to continue.
2. If you decide to continue downloading this software, look at the questions they are asking. Are these questions appropriate to the software itself? Or are they simply a method for gathering advertising information? I.e. are these questions impertinent?
3. Assuming that you haven't given up on the download yet, decide for yourself which questions are pertinent and answer those honestly. For every other required question, LIE.
4. Make your lies obvious to anyone who reads them. Make them commentary on your opinion about impertinent questions. Don't try to deceive, simply give incorrect information which is obvious to any human, but probably acceptable to computerized checking systems.
   First Name: A. Noyed
   Last Name: User
   Address: 123 Forty-Fifth St.
   City: Townville
   State: Denial
   Country: Atlantis
   Title: Gone With the Wind
   Business: Not Ure Bus, Inc.
5. For any optional information simply don't enter anything.

Naturally, these guidelines are for personal downloads. For professional downloads you must consider how your answers affect your company. I'm not saying that there won't be questions that are just as impertinent when you are downloading in a professional capacity. It is just more important to consider the ramifications of your actions when you are acting as a representative of your company.

And you still have to consider those ramifications when you are doing this for personal downloads, too. You have to evaluate the pertinence of the questions using your best common sense. In particular, if you feel that there might be legal ramifications to entering incorrect information in response to any question, then you should obviously consider the question pertinent and answer honestly or refuse to download.

Refusing to download is my prefered advice. Obtaining free software should never make you feel uncomfortable. If the process makes you feel uncomfortable, then consider that to be a cost for the software. In my case, it is almost always far more than I want to pay.