Monday, 29 April, 2002

SuSE 8.0 Annoyances

The SuSE Linux 8.0 install isn't flawless (see Saturday's entry).  Even though I configured my system for DHCP, the installation program didn't install the DHCP client, so I couldn't access my network until I tinkered a bit.  I had to get to the command line for that one.  I guess I could have used the GUI installer, but dang this system is slow!  It's a 200 MHz Pentium with 64 MB of RAM.  It seemed to run KDE 1.2 just fine, but it's terribly pokey with the latest SuSE release running KDE 3.  I suspect that it'd run much nicer with more memory, that old machine is maxed out at 64 MB.  Looks like I'll be buying a new motherboard this week.

A few annoying bits to mention:

  • The YAST2 setup program has a very bad user interface.  And I'm talking bad as compared to other text-based interfaces.  Ugh!  The thing borders on unusable, although it seems to work so I'll cut it some slack.
  • Why does SuSE insist on putting some silly background on the console window and scrolling the console output there?  When I'm at the command line, I want white text on a black background.  Leave the silly themes for the GUIs.
  • Why won't the YAST2 setup program recognize the right hand Control and Alt keys on my keyboard?  This is probably some arcane thing I'll have to change in the NCurses library or something.  You'd think, whatever it is, they would have make the default configuration recognize those two keys.
The above notwithstanding, I'm still impressed by the smoothness of the new SuSE install.  I'll report again once I install it on a more modern machine.  It's hardly fair expecting the latest and greatest GUI to be real snappy on 6-year-old hardware.

Sunday, 28 April, 2002

Tom DeMarco's The Deadline

I finished Tom DeMarco's book The Deadline over the weekend.  It's subtitled "A Novel About Project Management."  The book has some keen insight into project management, all wrapped up in a silly story about a project manager who is kidnapped and "encouraged" to help the country of Morovia become a software powerhouse.  It's worth the read, although if you've read DeMarco and Lister's Peopleware, you probably won't find anything new here.  The treatment in Peopleware is more thorough, but it's not as fun to read.  I'd suggest you read both books, and then pass The Deadline on to those in your organization who aren't directly involved in project management, but who would benefit from learning a little more about if.  If they like The Deadline and want to learn more, then give them Peopleware.

Saturday, 27 April, 2002

SuSE Linux 8.0

I've been tinkering a bit with Linux at home, although I haven't mentioned it here in quite some time.  Today I bought SuSE Linux 8.0 at the local Fry's, along with a DVD-ROM drive to replace the CD-ROM that died last year.  I installed the new drive, popped the SuSE install DVD in it (SuSE 8.0 comes in a package with 7 CDs and one DVD), and rebooted.  First problem.  My 6-year-old Dell Dimension didn't know how to boot from the DVD.  The Web is a wonderful thing, though.  A quick trip to Dell's support site, and I had a new flash BIOS upgrade on diskette.  I then was able to boot the DVD, and I selected the "New Installation" option from the first install screen.

The folks at SuSE have been listening to people like me rant about the difficulty of installing Linux.  Their installation process is very much improved over the 7.0 version that I worked with about 18 months ago (seeDecember 17, 2000).  The new installation program did an excellent job of detecting my hardware, and leading me through the install process.  I set the options and went to take a nap while the install proceeded to copy files from the DVD onto my computer.  When I came back, I verified a few options (screen resolutions, sound card settings, and network configuration), and rebooted.  With very little effort, certainly no more than was required to install Windows ME, I now have a fully-functioning Linux system running KDE 3.0 and StarOffice.  I haven't tried StarOffice yet other than to start it up, but I'm going to give it a whirl.  Since my new writing project is Linux-centric, perhaps I'll do the writing on the Linux box.  The only drawback is that the computer is relatively slow:  a 200 MHz Pentium with only 64 MB of RAM.

The KDE folks have been busy, too.  At first glance, I find KDE 3.0 to be a reasonably good GUI.  I'll have to work with it for a while before I give it any stronger endorsement, but it's a huge improvement over the KDE 1.2 that I've been running for two years.

I still have reservations about Linux as a general-purpose desktop operating system, but I'm going to give it another try.  The progress that SuSE and the KDE team have made in the last two years is astounding.  If other Linux software has evolved similarly, then perhaps my next computer upgrade will find the Linux box as my primary machine.  Wouldn't that be something?

Friday, 26 April, 2002

Old Technology Saves the Day

Today I had to recover a bunch of files from a crashed laptop at work.  Somebody had installed Windows 2000 on a FAT32 partition, which is never a good idea because when a FAT system gets corrupted, the chances of actually recovering are pretty slim.  NTFS is a much better choice.  The result was, the only way to boot the machine was with a Windows 98 boot diskette—DOS mode.  So here I am faced with the task of copying about 200 megabytes from the laptop, and only a diskette to do it with.  And then I remembered the old parallel ZIP drive that I bought back in 1996 or so.  That thing's been gathering dust in a closet ever since I got my first CD burner.  I pulled it out, dusted it off, and found the installation diskette.  In less than five minutes, I was happily, albeit slowly, copying the files onto the ZIP drive.

The only real hitch was a 150 MB ZIP file that the user had created.  There was no way I was going to get that monster onto a 100 MB removable cartridge.  So I dug into my archives again and came up with a 9-year-old copy of PKZIP, which happily spanned multiple disks with the file.

Some days—rarely, but it happens—it pays to be a packrat.

Thursday, 25 April, 2002

Internet Mail Standards

Would you care to guess how many RFCs deal in some way with Internet Mail?  I submit to you this list ofInternet Mail standards from the Internet Consortium.  Count them if you like.  I lost track after 150.  No, I'm not smoking anything.

Oh, sure, you don't have to understand all of them in order to understand how mail works, and you can even implement an Internet Mail server without having to read all of the standards.  There are 60 documents, for example, that discuss message and transmission security—something that's not essential—at least not yet.  Still, 90 is way too many, and you have to understand most of those in order to build a mail server that will function in today's environment.  There are too many options, too many formats, too many obsolete commands and features that have to be supported for "compatibility with older systems ," too many contradictions, special cases, and ambiguities.  Individually, the documents are dry and poorly organized, and the writing is uninspired even for technical specifications.  Combined, the Internet Mail standards base is a big icky incoherent pile of goo.  It's a wonder that Internet Mail works as well as it does, and no wonder at all that people are finally starting to notice that there's a problem here.  What frightens me, though, is that everybody's proposing patches and extensions to the existing system rather than a totally new approach.  There's simply no way to extend or fix the current Internet Mail system (Simple Mail Transfer Protocol, or SMTP) to address the concerns that have been raised in the last 5 or 8 years.  Volume, security, and spam are proving to be very difficult problems requiring creative solutions that cannot reasonably be implemented on the current system. 

SMTP is history.  We just haven't found the replacement yet.  We best start looking soon, lest our email remain forever in 1989..

Wednesday, 24 April, 2002

Problems Installing Windows 95

Today I had to install Windows 95 at the office in order to test our product.  We don't "officially" support Windows 95 anymore (heck, even Microsoft doesn't really support it), but you'd be amazed at how many companies still have thousands of machines running Windows 95.  So we make at least a token attempt to assure that our product runs on that platform.

As it turns out, installing Windows 95 is not trivial, especially on a computer that has recently had Windows 98 or Windows 2000 installed on it.  95 will notice the newer version of Windows and refuse to install over it.  What you need is a Windows 95 boot disk, with CD-ROM driver.  Try finding one of those laying around the office.  Do you remember how to configure MSCDEX.EXE?  I thought not.

But let's say, for argument's sake, that you were able to get Windows 95 installed and now you want to hook your computer up to the network.  But the network card you installed in the computer was built in 1999 and Windows 95 doesn't recognize it.  Where to find a driver?  The Web, of course.  Just find the manufacturer and model number on the network card, find the manufacturer's web site, and download the driver.  Simple.  Clean.  Neat.  Except the @#$% manufacturer decided not to print any useful information on the card!  Now there's a good idea!  What idiot thought that up?  Fortunately I found an old 3Com card that had the model number on it and I was able to download the driver to a diskette and install it on the machine.  I still don't know who made the other card.  I guess I could drop it into a Windows 2000 machine and let the New Hardware Wizard tell me.  But I figure if the manufacturer is too ashamed of the card to print their name on it, I probably don't want it in my computer.

I've run into this problem with all types of cards over the years.  Is it really that expensive to print a few more characters on the board?  The lack of identifying characteristics on these boards makes it impossible to use them in many cases.  Perhaps it's the manufacturers' way of getting us to buy new hardware.  

Sometimes I just shake my head in wonder.

Tuesday, 23 April, 2002

Feeling Slightly Depressed

I've not been posting entries regularly for quite some time now.  I've tried to pass it off as my being too busy, but today I finally realized that I'm just having difficulty finding the motivation to do anything other than the absolutely minimum necessary.  I've been unhappy for the silliest reason:  I temporarily forgot that my happiness is my own responsibility.  Things and people aren't going to make me enjoy my life.  I'm not going to find happiness in new friends, new things, or easier work.  Happiness, like unhappiness, is a positive feedback loop.  If I wake up in the morning and think "It's going to be a good day," then I'm halfway there.  All I have to do, whenever something unexpected happens, is maintain the right attitude and things don't faze me.  They have a way of working out.  It's when I don't get enough sleep, or when I don't take time to sit back and relax from time to time, that I let small worries and little problems disrupt my attitude.  It becomes a vicious feedback loop, where each small disappointment pushes me closer and closer to becoming the type of lethargic, cynical, spiteful person I most detest.

Fortunately, Debra and my friends aren't afraid to slap me upside the head repeatedly until I get the message.  Things are looking up.

Tuesday, 16 April, 2002

The Flo Control Project

Now here's a good use of technology:  The Flo Control Project.  The guy's using a digital camera hooked to a computer to take a picture of his cat Flo on the way into the house.  If the picture compares favorably with a picture on file, then the cat's door is unlocked.  If the cat is carrying a mouse or a bird or other present, or if the cat at the door isn't really the cat, then the door remains locked.  This is very cool stuff, and the web site gives a very good explanation of how it works.  Well worth the read.

Monday, 15 April, 2002

Icons of Evolution

In his book Icons of Evolution, Jonathan Wells explores many of the teachings of current evolutionary theory, and shows how preconceived notions and dogmatism are being substituted for science.  The book explains that many of our "icons of evolution"—breakthrough discoveries—are not nearly as conclusive evidence as their proponents would have you believe.  Peppered moths, for example, don't normally rest on tree trunks, and the extra wings on four-winged fruit flies aren't really working wings at all.  Those are just two examples of evolutionary evidence that I remember from college biology, which Wells discredits in his book.

Are his accusations true?  I don't know enough about the science behind these things to say one way or another, but if Wells is even partly right, then evolutionary theory is resting on a house of sloppy science, half-truths, lies, and fabrications.  He doesn't say that there is no evidence for evolution—quite to the contrary.  There are obvious examples of evolution in action within species.  He points out, though, that we've yet to see concrete evidence of speciation—a new species emerging due to morphological change in an existing species.  It does give one pause to think.  Is it true that much of what high schools and colleges are teaching about Darwinian evolution is wrong?  Do we really know so little about the origins of different species?  Wells makes a very good case to show that we know a lot less than the Darwinists claim.  I wouldn't go quite so far, but you can bet that I'll be looking a little more critically at the evidence from now on.

Sunday, 14 April, 2002

Some Things Are Too Complicated

So many things in this world seem overly complex to me.  Today I've been wondering about measurement scales.  In elementary school I learned my weights and measures, like every little schoolboy should, and prided myself on remembering that 12 inches is a foot, and there are 1,760 yards in a mile.  When I discovered the metric system in high school, I immediately wondered why we don't use that.  It's so much easier to use.  Water freezes at 0C, and boils at 100C.  Simple.  Much simpler than remembering 32F and 212F.  And measurements?  I find it much easier to calculate the difference between 8mm and 10mm, than between 9/16 inches and 11/64 inches.

And time.  I won't even worry about the weird time scales: 60, 60, and 24.  But what's with the AM/PM thing?  Why doesn't everybody use 24-hour (military) time?  When somebody says "It's 6 o'clock", you'd better know if he means AM or PM.  But if somebody said "1800", you'd know immediately.  And the whole thing about time zones.  Forget it.  Yes, there used to be a reason for calculating local noon.  But no longer.  Why don't we all, the whole world, standardize on a time system.  UTC (what used to be called GMT) seems simple enough.  Rather than posting your hours of operation as 8AM to 5PM Central Daylight Time, just say 1400 to 2200 UTC.  Simple.  And don't even get me started on the whole Daylight Savings Time idiocy!  Arizona seems to get along just fine without it, why can't the rest of the country?

Not that I expect any of this to change in my lifetime.  Change requires thought and effort, and people in general are loathe to put forth effort to change anything that doesn't present an immediate problem—the sad state of recycling in this country being a case in point.  People say "if it ain't broke, don't fix it."  Well, it's broke—really broke.  But it seems to work okay for now, so we'll continue stupidly along until the effort required to overcome the problems of incompatible standards outweighs the effort required to change.  And then people will start pointing fingers at others, blaming everybody else for the time and productivity lost and the inconvenience of having to change immediately something that they've known about and have had the opportunity to change for years.  It's the Y2K problem all over again, but writ large.  Very large.

Friday, 12 April, 2002

Changing Jobs

Come Monday, I start a new job.  I'm still with Catapult Systems, and still working in the Inquisite product group.  But now, rather than working on custom solutions for clients, I'll be a sales engineer.  Yes, I've joined the Dark Side.  I'll be sitting in on sales calls to answer technical questions, writing technical documents for Sales and Marketing, supporting the Sales and Marketing teams' computers, and whatever else needs doing.  Although I have been performing many of these functions on a part-time basis for over a year, Monday I start a transition to doing it full-time.  In this new position, I will be writing almost no code, whish is quite a stretch for me, personally, as for almost all of the last 20 years my primary job has been programming.  The challenge here will be keeping up with technology and writing enough code to stay sharp so that I can talk intelligently to the development team and to potential clients' technical staff.

Wednesday, 10 April, 2002

A Quick Trip to Los Angeles

I flew to Los Angeles yesterday for today's meeting with a potential client.  It was an unbelievably uneventful trip—one of the best since the security changes took effect after September 11.  Last night we had dinner at an odd little Thai restaurant called Toi, on Sunset Boulevard.  The place is decorated with rock and roll posters, they play good old-fashioned rock and roll music at a reasonable volume, and the food is quite good.  Today we finished our meeting at about 11:30 and headed to the wharf in Redondo Beach, where we had lunch at a Korean restaurant.  Fish soup is ... different.  Walking around the wharf, I ran across a pearl shop where for $9 I got to pick out an oyster.  They sliced it open, got the pearl (a nice sized black pearl), and then mounted it on the setting of my choice.  The setting, of course, was an additional cost.  This probably has been going on for some time, but it's the first I'd heard of it.  It makes me wonder, though, if I can get a nice pearl like that for $9, why pearls cost so much in jewelry stores.

Monday, 08 April, 2002

Denver Fights Childhood Obesity

In a pilot program in Denver, aimed at battling childhood obesity, 200 overweight school kids will be given pedometers in hopes that they will compete among their peers to see who walks at least 10,000 steps a day. This article has the details.  Yeah, right.

According to the article, "[D]octors and pediatric psychiatrists say the problem of childhood obesity is complicated and can't be blamed just on video games or the prevalence of fast food in middle and high schools."  True enough, but certainly those are contributing factors.  The claim that increased academic loads are a culprit is pure B.S.  Kids are fat these days because their parents feed them too much (or allow them to eat too much junk), and don't force the kids to be active.  So what if they can't walk to school.  When they get home, send them outside to play.  Or enroll them in soccer, karate, marching band, or something else active.  The blame for childhood obesity  lays squarely on the parents.  Period.

Thursday, 04 April, 2002

Off to Harlingen

Debra and I are off to Harlingen, Texas for my annual school reunion at the Marine Military Academy.  This will be the seventh year in a row that I've been able to take a couple days off and go there for the reunion.  Every year they invite all former Cadets.  We converge on the Lower Rio Grande Valley—usually about 100 of us—and spend the weekend telling war stories and reliving the good old times.  Some people do reunions every 10 years or so.  We do them every year, and every year I come away with a renewed appreciation for the friends I made during those years, and for the staff and faculty there who helped shape my life and continue to shape the lives of young men every day.  I can't say enough good about MMA, and the wonderful people who make it work.

Monday, 01 April, 2002

Fighting with SMTP

I've been fighting with email again, trying to construct a message that contains hyperlinks that don't get broken by Mail Transfer Agents.  The project started because I needed to solve a problem for work, but I'm starting to take it personal.  It's an interesting exercise in frustration digging through dozens of Internet RFCs and trying many different approaches.  Just when I think I have it working, I get a message from one of my testers telling me that Mail Mangler 43.6 decided to truncate a URL, or that the URL doesn't appear active in Envelope version 5.  I've decided to really dig into the guts of Internet mail in order to figure out why it's so dang hard, and whether we can fix it.  I'm tempted to just can the whole mess and start over, but perhaps there's a way to salvage some of the knowledge we've gained over the years.