compaq

Optimizing a Linux box in-place

Dave Farquhar General , , , , , , , ,

Here’s the Linux bit I promised yesterday. I wrote it much earlier, so I might as well throw it out there.
Our test firewall at work is an old Pentium-200 running Red Hat Linux and a commercial firewall app. (No, I won’t disclose which one. Security, you know.) It’s a bit slow. A P200 is severe overkill for the firewall built into the Linux kernel (Steve DeLassus and I made a firewall out of the first PC he ever bought, a 486SX/20 of 1992 vintage, which, save the loss of the original power supply in an electrical storm, has never required any service), but this commercial package does a lot more than the simple firewalls built into Unixish kernels do.

It had 72 megs of RAM in it and swapped mercilessly. Its speed seemed to be OK once it was booted, but seeing as this is a testbed, it tends to get rebooted an awful lot. I needed to do something for it.

So I trekked into the PC graveyard to see what I could dig up. I found a Compaq 386DX/20. I left that alone. That’ll be useful if I ever need to pillage a pair of Compaq drive rails, which has happened before. Unfortunately those rails are worth more than the rest of the computer. I also spotted a Mac SE. That’ll be handy if I ever need a doorstop. Then I found a Pentium-75 and another Pentium of unknown speed. I opened them up. The 75 had a pair of 16-meg sticks. I opened up the unknown Pentium and looked inside. Ugh. Socket 4. That meant it was a Pentium-60, or, at best, a Pentium-66. It had a pair of 8-meg sticks.

I pulled the memory sticks out of the 75. The 60 didn’t have anything usable in it, save a pair of hard drives, both 540 megs, one a Quantum and the other a Seagate. I took the Seagate because it was easier to unbolt. I don’t have any way of knowing at this late date which of those drives was the better performer, and it probably doesn’t make much difference anymore.

The idea was to add some memory, and put in a second hard drive dedicated to virtual memory. Since the likelihood of the machine needing to read data from a drive and simultaneously hit virtual memory was fairly high, I wanted the virtual memory on its own drive. Furthermore, Linux’s partition-read
mechanism isn’t terribly efficient. This doesn’t matter for SCSI drives, which re-order I/O events, but for IDE drives it matters a lot. So getting the swap partition onto a dedicated drive was likely to improve performance a fair bit. (If this were a production system, it would probably have a SCSI
drive in it.)

So I swapped in the 16s for the 4s and found an empty bay to hold the 540, which I put on the second IDE channel as master (another performance trick), and booted Linux. The next trick is to use your favorite disk partitioning tool (I like cfdisk, but I can navigate plain old fdisk) to blow away whatever partition is on the new drive (this one was /dev/hdc) and create a single partition. I just made it the size of the drive, since 2.4 can deal with large swap partitions and Linux is smart enough to use whatever virtual memory it needs, not just automatically use all it has available. Then I set
it to type 82. Linux can do swapfiles, but a filesystemless dedicated swap partition gives better performance.

Next, I edited /etc/fstab. I found an entry for the swap partition pointing at /dev/hda2. I changed that to /dev/hdc1. That means I now have a small swap partition just sitting on the first drive unused, but that’s not a big deal to me. The system’s not using the disk space it has. While I was there, I noticed the CD-ROM drive was pointing at /dev/cdrom. I asked Charlie, our Unix/Linux guru, if Red Hat had some intelligence I didn’t know about. He said /dev/cdrom was just a symlink. I changed the entry to read /dev/hdd, which is where the CD-ROM drive ended up after my shuffle. Better to just code things directly than try to track symlinks, in my estimation.

Next, I issued the command mkswap /dev/hdc1 to initialize the swap partition. Then I rebooted and listened.

Indeed, during boot, the second drive was getting activity. I logged in and ran top, then hit shift-M to have a look at memory usage. The firewalling software was eating up a lot. But swap usage was down.

I decided to try cutting memory usage down a little more. I loaded /etc/inittab into vi. Red Hat by default gives you six virtual consoles. This machine has little need for more than two. Pulling the extras saves you a couple of megs. Near the end of the file you’ll see several lines that look something like this:

1:2345:respawn:/sbin/mingetty 38400 tty1

I commented out the last four of those. Hit the i key to put vi in insert mode, scroll down to those lines, add a # to the beginning of them, then hit ESC, then hit ZZ (shift-Z twice) to rapidly save the file, no questions asked. (I know, vi ain’t friendly, but it’s there.)

Then I had a look at /etc/rc3.d to see what daemons were running. I found apmd, sendmail, and gpm running. That was a waste of a couple megs, not to mention a possible security risk. I vaguely remember all three of them having had security issues in the past, and sendmail is one of those programs that should never be running unless you need it. Yes, this machine’s just practice, but Hall of Fame catcher Johnny Bench found that if he got sloppy and just let wild pitches go while he was warming up pitchers, he wasn’t as sharp at blocking potential wild pitches during the game when
it counted. So he worked just as hard during practice as he did during the game. Now he’s considered the greatest catcher of all time.

So I applied the Johnny Bench principle and disabled them with the following command sequence:

mv /etc/rc3.d/S26apmd /etc/rc3.d/K26apmd
mv /etc/rc3.d/S80sendmail /etc/rc3.d/K80sendmail
mv /etc/rc3.d/S85gpm /etc/rc3.d/K85gpm

I rebooted to find memory usage down by about 4 megs and the system booted a little faster. It was also more secure.

Total downtime: About 45 minutes.

That was time well spent. I may end up having to just bite the bullet and get some memory, but the system will perform better with these changes no matter how much memory is in it. And, more importantly, performing this exercise made me notice something I hadn’t noticed before. It let me tighten up security.

Had I blindly just ordered some memory to put in the system, or a new PC, like some people unfortunately advocate, I wouldn’t have necessarily noticed that as quickly.

—
Speaking of Linux, I did finally get Apache, PHP, and MySQL all talking together on my church’s 486. I used phpWeblog, which is an awfully nice package. Pages load in an acceptable two seconds. I notice the machine is paging, so a little more memory will probably help that. It’s amazing that people are throwing away Pentium-class machines when even a 486 has enough power to be a decent intranet server.

Not everyone’s so fortunate as you and me. Give ’em to someone who can use them if you don’t want them.

How Linux could own the education market

Dave Farquhar Linux , , , , , , , , , , , , ,

How Linux could own the education market. I spent some time yesterday evening working on computers. They were contrasts to the extreme: One, a brand-spankin’ new 1 GHz AMD Duron system with 512MB of RAM and 80 GB of 7200-rpm storage (IDE, unfortunately–but for $800, what do you want?). The other was an elderly AST 486SX/25 running Windows 3.1 belonging to a local teacher who goes to my church.
She teaches kindergarten, and the AST used to be her home computer. When she bought a Compaq Presario a couple of years ago, she took the AST to school. It’s more useful there than in her basement, and there’d be no computer in her classroom if it weren’t for that.

I don’t understand why that is. As much as my sister jokes about it, we don’t exactly live in the ghetto. The school district has money, but it isn’t spending it on computers. Whether that’s a good or bad thing depends on your point of view. The majority of people living in Oakville probably own home computers, so this probably isn’t contributing to the technology gap. But I wonder sometimes how things might have been if I’d been exposed to computers a few years earlier.

I was shocked how much I remembered about Windows 3.1. And I was able to figure out how to get her CD-ROM drive to play music CDs. Don’t ask me how; this was the first I’d messed with Windows 3.1 since 1994 and I’d prefer it stay that way–I was so impressed by Windows 3.1 that I’m one of the 12 people who actually went out and paid money for OS/2. I own actual, retail-box copies of OS/2 2.1, 3.0, and 4.0. And I remember distinctly thinking that her computer has enough memory to run OS/2 at least as well as it runs Windows 3.1…

I also remember distinctly thinking that my employer pays someone $15 a pound to haul better computers than hers away several times a year. We regard 486s as junk; low-end Pentiums may also go out, depending on whether the right person finds out about them beforehand. Usually they work just fine–the problem isn’t the computers, it’s people trying to run Internet Exploiter 6 and Office 2000 on them. They’d run Windows 95 and Office 95 perfectly fine.

But a lot of times we can’t give these old computers away because the licenses for the software that originally came with them are long gone. Old computers are useless without software, so no one would want them anyway.

Now, let me tell you something about kids. Kids don’t care much about the computers they use. As long as there’s software on them, they’ll use them. When I was a kid 20 years ago, I used Radio Shack TRS-80 computers at school. The next year, my family moved, and my new school had Commodore 64s. I couldn’t tell much difference. My next-door neighbor had a Radio Shack Color Computer. They were computers. The Commodores had better graphics, but from a usability standpoint, the biggest difference was where the cartridge slot was so you could change programs. Later on I took a summer class at the local junior college, learning about Apple IIs and IBM PCs. I adjusted smoothly. So did all the other kids in the class. Software was software.

Kids don’t care if the computer they’re using runs Windows or Mac OS or Linux. All they care about is whether there are cool programs to run.

So, businesses throw useless computers away, or they give useless computers to schools so they don’t have to pay someone to haul them away. And schools don’t generally know what to do with obsolete computers that lack software.

Linux won’t run fabulously on old 486s, but Debian with a lightweight window manager like IceWM will run OK. (Let’s face it, Windows 3.1 doesn’t run fabulously on them either–it crashes if you breathe wrong.) I know of a project to clone Oregon Trail on Linux. Great start. How about Sea Route to India? I remember playing that on C-64s at school. It may have been a type-in out of a magazine–I don’t remember where exactly it came from. In these violent times, Artillery might be too controversial, but it taught us early on about angles and forces. Artillery was an ancestor to games like Scorched Earth, but without the heavy-duty nukes. Close wasn’t good enough to win in Artillery. You had to be exact. And no blowing up the mountains between you and your opponents either. You had to figure out how to get over them.

But what about doing homework? By the time I was in the sixth grade, they were teaching us how to use word processors and databases and spreadsheets. AbiWord is a fabulous lightweight word processor. It gives you fonts and spell-checking and good page formatting. (I learned word processing on Bank Street Writer. AbiWord is a far, far cry from that. Frankly, I’d rather write a paper with vi than with Bank Street Writer.) Besides being feature-rich, AbiWord’s been lightning fast on every computer I’ve tried it on. Gnumeric is a nice, fast, capable spreadsheet. I don’t know of a free-form database, but I haven’t looked for one lately either. (I don’t think we need to be trying to teach our 6th graders SQL.)

But what about for younger kids? I remember a program called The Factory. The object was you combined chemicals to make monsters. Different chemicals made different monsters. I seem to remember you played around to see what chemicals would make which heads and torsos and arms. Then the computer started showing you monsters and you had to figure out what chemicals to give it to match them. I also remember a program called Snooper Troops. I don’t remember much else about it, other than it was a mystery and you went around looking for clues, and one of my classmates accidentally formatted the disk one day before any of us had managed to solve it. We couldn’t get the disk replaced, because it was out of print.

And Spinnaker had all sorts of simple titles for younger kids that let them tell stories and other stuff. It seemed cool at the time. But that was almost 20 years ago, so about all I remember was that sailboat logo and some corny theme music.

The other thing about those old days was that the majority of these programs were written in Basic. An ambitious teacher could modify them, to make them easier or harder, or improve the graphics a little. As we got older and learned to program, some of us would try our hand at making changes. You can’t do that anymore with Windows or Macintosh educational titles. Open source can bring all that back too, provided the programs are written in languages like Perl or Python. And it can give cash-strapped schools a way to get computers where kids can use them.

Now I’m wondering what it would take to write something like The Factory in Python…

What on earth is going on?

Dave Farquhar General , , , , , , , ,

AOL-Time Warner in talks to buy Red Hat? I found this this morning. It’s intriguing, but I can’t decide if a buyout would be a good thing or a bad thing. After all, Netscape was in decline when AOL bought it. It nosedived afterward. Obviously, the problem was twofold. When AOL acquired Netscape, they didn’t acquire all of its mindshare. Some of the most talented people got fed up and left. You can take Jim Barksdale or you can leave him. The loss of Marc Andreesen and Jamie Zawinski, though, was substantial.
The second problem was that AOL wasn’t serious about competing. They bought a browser technology and basically sat on it. Netscape 4.x was fundamentally flawed, as even Zawinski acknowledges, although I would argue it was no more fundamentally flawed than IE 4.x. The Gecko engine, on which Netscape 6.x is based, is solid technology, even though it took longer to get to market than anyone had hoped. Although Netscape 6.x won’t bowl anyone over, other browsers based on the technology, such as Galeon, are absolutely fantastic. But AOL chose to release a half-hearted browser with the Netscape name on it and continued to use the IE engine in its flagship product even after the favorable agreement with Microsoft that prompted AOL to do so in the first place expired.

That begs the question of what AOL would do with Red Hat if it owned it. Red Hat is still the big-name player in the Linux field, but Red Hat is concentrating on the server market. You can still buy Red Hat at retail, but on the desktop, Red Hat is arguably #3 in popularity now behind France’s Mandrake and Germany’s SuSE. Red Hat is the only Linux company that’s making money, but that’s largely by selling consulting. That’s not AOL’s core business. At this point, AOL is more of a media company than a technology company. Software just gives AOL more outlets to sell its media content. Consulting doesn’t do that.

The best possible scenario for a Red Hat buyout would be for AOL to, as Microsoft puts it, “eat its own dog food,” that is, rip out the infrastructure it bought from other companies and replace it with the technology it just developed or acquired. Since AOL is largely powered by Sun servers, it wouldn’t be terribly difficult to migrate the infrastructure to Red Hat running on Intel. Then AOL could give a big boost to its newly-acquired services division by saying, “We did it and we can help you do it too.” They can also cite Amazon’s recent successes in moving its infrastructure to Red Hat Linux. There is precedence for that; after AOL bought Time Warner, the entire company started using AOL for e-mail, a move widely questioned by anyone who’s used anything other than AOL for mail.

Of course, it would be expected that AOL would port its online service to Linux, which would create the truly odd couple of the computing field. AOL, meet sed and awk. Red Hat would certainly lose its purity and much of its credibility among the Linux die-hards. AOL would bank on making up the loss by gaining users closer to the mainstream. AOL could potentially put some Linux on its corporate desktops, but being a media company, an all-out migration to Linux everywhere within is very far-fetched.

To really make this work, AOL would either have to enter the hardware business and sell PCs at retail using its newly acquired Red Hat distribution and newly ported AOL for Linux and possibly an AOL-branded office suite based on OpenOffice, or it would have to partner with a hardware company. Partnering with a big name seems unlikely–a Compaq or an HP or an IBM wouldn’t do it for fear of retaliation from Microsoft. Sun has never expressed any interest in entering the retail computer business, and even though Sun loves to take opportunities to harm Microsoft, Sun probably wouldn’t cooperate with AOL if AOL replaced its Sun infrastructure with Red Hat Linux. Struggling eMachines might be the best bet, since it’s strictly a consumer brand, has a large presence, but hasn’t consistently turned a profit. But AOL could just as easily follow eMachines’ example, buying and re-branding low-end Far East clones and selling them at retail as loss-leaders, taking advantage of its lack of need for Windows (which accounts for roughly $75 of the cost of a retail PC) and making its profit off new subscribers to its dialup and broadband customers. A $349 PC sold at retail with a flashy GUI, decent productivity software and AOL is all the computer many consumers need.

The advantage to this scenario for everyone else is that AOL would probably dump more development into either the KDE or GNOME projects in order to give itself more and higher-quality software to offer. The official trees can either take these changes or leave them. Undoubtedly, some of the changes would be awful, and the official trees would opt to leave them. But with its 18 years’ worth of experience developing GUIs, some of the changes would likely be a good thing as well.

The more likely scenario: AOL will buy out Red Hat, not have a clue what to do with it, and Red Hat Linux will languish just like Netscape.

The even more likely scenario: AOL will come to its senses, realize that Red Hat Linux has nothing to do with its core business, and the two companies will go their separate ways.

A nice Sunday surprise

Dave Farquhar General , , , , , , , ,

I had a big surprise Sunday night. A couple of months ago, I was up at Bethlehem Lutheran Church in north St. Louis, and they asked me if I’d come to their Christmas banquet this year. I said I’d love to come to their Christmas banquet. They said they’d send me tickets. “Some” ended up meaning five. They’re generous people. I ended up using one–I didn’t feel like looking too hard for a date, and I felt weird asking a bunch of my friends who’ve never been up there to go with me on a rainy Sunday night.
My relationship with Bethlehem goes back several years. I moved to St. Louis in Nov. 1998, and immediately started going to a church in Oakville, a semi-ritzy, very white-middle class suburb in south St. Louis County. I was driving 30 minutes to go to church every Sunday because I had connections there, and I’d never seen a church that was so on fire. I liked it there. It was a church that made me better, and, as I would quickly learn, it was more than willing to let me make it better too. Mark my words: When you find a church like that, keep it. They’re harder to find than you might think.

In Faith Lutheran in Oakville and Bethlehem Lutheran in St. Louis, I’ve found two. And I’m much the better for it.

The north St. Louis neighborhood around Bethlehem is about as opposite of Oakville as you can get. It’s not ghetto, but the buildings are well past their prime. A number of them are condemned. Many others are boarded up. It’s lower-middle class at best. But there are people there who are trying to make a difference.

I’d been going to Faith Lutheran in Oakville for a couple of weeks when I started receiving its newsletter. And in that first newsletter was a blurb from The Rev. John Schmidtke, the pastor at Bethlehem. Faith is one of five suburban churches that has partnered with Bethlehem to reach out to its community. Pastor Schmidtke’s letter was a wish list of sorts, but he wasn’t wanting money or objects. He wanted people. “Who can help us build a computer lab so we can teach elementary computer skills to the people of our community?” he asked. “Who can help us give our children a safe, welcome place where they can sit down at a computer and do their homework?” At the end of the letter, he gave his phone number.

The next day, I called him.

He said he already had some beat-up PCs that had been donated to him. I asked when I could come look at them. I don’t really remember many specifics anymore, other than driving into north St. Louis in a snowstorm one night to come look at a pair of beat-up Compaq Proliant servers. They were DX2-66s, decked out with external SCSI CD-ROM towers. One of them had three SCSI drives. The other had five. They were pretty snazzy servers… in 1993.

It was a humble beginning. Pastor solicited some obsolete computers from other businesses, and since this was the midst of the Y2K crunch, he was able to find plenty of people willing to give up some 386s and 486s they’d just retired. The best catch was a pair of non-compliant Pentium-75s. One of them even had a hard drive–a 40-megger. No, not a 40-gig drive. A 40-meg drive, like most of us had in our first AT clone.

Basically, we had a whole lot of nothing, and I did a whole lot of nothing with it. Sure, I was able to impress a few people by taking hard drives out of 486s and putting them in those Pentiums and booting up DOS, but as far as doing anything useful, we didn’t have much. So the project pretty much sat there, a pile of beat-up PCs in the corner of a storage room.

Then one day in the summer of 2000, I got a voice mail message. It was Pastor Schmidtke. He sounded excited, but there was a certain plea in his voice. He had a grant for several thousand dollars, and it was pretty much there for the asking, assuming he knew what to ask for. He didn’t know what to ask for. So he asked me if he could have five minutes of my time to tell him the wisest way to spend a few thousand dollars to build a computer lab.

I hopped on the ‘Net and checked it out, then faxed him a shopping list. For the budget he gave me, I figured I’d be able to get several name-brand PCs and a laser printer. The grant needed three competitive bids, so I priced systems from IBM, Compaq, and Dell to give him ballpark figures, plus phone numbers to call to get hard quotes if that was what he needed.

A few months later he had the money. A couple more months after that, we’d turned that money into eight new Compaq Deskpro PCs. I wasn’t going to leave him high and dry at that point–what good is a room full of computers when no one there knew how to make them go? A couple more months after that, some volunteers had turned that storage room into a nicely laid-out computer room. So then I set about taking those PCs, installing network cards, cabling and hubs, configuring them identically, and connecting a printer. We had a usable network. An Internet connection was the tough part. I took one of those Pentium-75s, installed a 56K modem and an Intel 10/100 NIC, and configured Freesco. We were live. While 56K dialup split among 9 PCs isn’t fabulous, it’s better than it sounds–while people are reading pages, after all, their computers aren’t loading stuff. I tried setting up a Squid server to help ease congestion a little, but Squid seemed to hurt as much as it helped, so I scrapped that idea.

So now, three years after we initially met, they have a working, useful computer lab. Neighborhood kids come in and research and type. Pastor’s family comes in, and with that many computers at their disposal, the kids can play around all they want for hours and his wife can get work done. It’s not the best, but it’s worlds beyond a pair of Pentium-75s. And in a neighborhood where a Pentium-200 is considered a luxury item, it’s doing a lot of good.

So I got to the banquet Sunday night and sat down at a table. There was a program sitting there at every place. I looked at it. “That’s nice,” I was thinking. “Star of Bethlehem Awards.” There were two people listed. Then I saw people were picking up the program and flipping pages. So I picked up mine, turned to the inside, and saw there were more than two people listed. Two more on page two, and then I turned to page 3 and saw my name. With a really kind write-up to go with it.

They read the write-up, along with everyone else’s writeup, after dinner. They gave each of us plaques and asked us to say a few words. I don’t remember exactly what I said–I’m not very comfortable giving impromptu speeches. It was Pastor Schmidtke who had the vision and who got the money. And it was Cathy, a member of the congregation, who made all the phone calls and made all the runs to Office Depot to get things like power strips and network cables when I ran out of power outlets or didn’t have quite enough reach. Maybe I could have done it all without them. But chances are I wouldn’t have. No one would have. One person can’t take on a project of that magnitude alone. It’ll kill you.

The speaker who read the write-up on me was interrupted by applause a couple of times. I got a round of applause as I walked up and another one as I sat down. Helping people like them is easy, because they appreciate it so much.

I hung the plaque up right after I got home. I guess that says something about priorities–I have an expensive Jesse Barnes print I bought more than a week ago that isn’t hung yet. But the sentiment behind that plaque is worth more than a room full of Jesse Barnes prints. It’s a nice plaque. It reads:

New Birth at Bethlehem

We Thank God For You

David Farquhar

For your ongoing support, encouragement, and Christian love to the ministry of Jesus Christ through Bethlehem Lutheran Church. You are God’s Star for the ministry of Bethlehem.
…Daniel 12:3

December 16, 2001
Bethlehem Lutheran Church, St. Louis, MO

Daniel 12:3 reads as follows:

“Those who have insight will shine brightly like the brightness of the expanse of heaven, and those who lead the many to righteousness, like the stars forever and ever.”

There’s just one more thing I wish I’d said Sunday night. They’re a group of people trying to make a difference in north St. Louis. A lot of them are there by choice. They didn’t have to give me an opportunity, but they did. I’m glad they did.

A different Monday, but not much better…

Dave Farquhar General , , , , , , , , ,

Moves at work continue, but unfortunately the electrical contractors we have are as incompetent as ever, and of course IT takes the brunt of the attack when computers don’t work. They don’t care if it’s an electrical problem or not; all they know is their computer doesn’t work, and of course it’s always IT’s fault if the computer doesn’t work. And with one person to keep 300 desktop PCs in tip-top shape, I usually can’t be up there and have the problem solved within five minutes.
In the last three weeks, we’ve lost three power supplies, two printers, an expensive proprietary modem, and a network card. In two instances, there was an honest-to-goodness fire, with flames and everything.

I think it’s time we sent an electrical contractor or two packing.

Meanwhile I’ve got incompetent department directors who plan moves without giving more than a half hour’s notice, and of course they throw a fit when the move falls to pieces and I’m off solving another problem. I also find myself not caring. Go ahead and yell. Davey’s not listening, la la la, and his boss isn’t listening, and his boss’ boss isn’t listening, and if his boss’ boss’ boss listens and says anything, he’ll have two, maybe three raving lunatics at his door in a heartbeat and I think he knows it.

Deep breath. OK. I feel better now. Kind of.

Let’s see what kind of hints The Big Guy may have been dropping with the day’s other events, shall we?

I had a meeting at church at 7 p.m. So I headed out to my car at 10 ’til 6, put my key in the ignition, and the engine coughed, and then nothing. No electrical system. Hmm. Time to find out how good Chrysler Roadside Assistance is, eh? Well, I called, waited an hour and a half, and they never showed up. So I paced in the beautiful October twilight, waiting for a driver who’d never arrive, thinking there are a number of things I’d love do at twilight outdoors in St. Louis in October (and waiting for a tow truck is very near the top of that list, let me tell you!) but it sure beats sitting in a meeting after dealing with irate, high-maintenance people at work for 9+ hours.

And I noticed something. I wasn’t at the meeting, and yet the world failed to fall apart.

Finally I gave up on the tow truck driver and asked one of my coworkers for a jump. Maybe the problem was a dead battery, even though I didn’t leave my lights on or anything. Indeed it was. I drove home, and about halfway there my battery light came on. I guided the car home, called Chrysler again, and asked them what to do.

On my answering machine, there was a pair of messages waiting for me. It was actually one message, but my answering machine is extremely rude and cuts you off after about 10.5 seconds. OK, maybe 30. But it seems like 10.5 seconds to everyone else but me. So most people leave a message, get cut off, then call me back. Sometimes they call me back a third or even a fourth time. Usually by then they’re pretty steamed. But I digress, as always. The message messages basically boiled down to, “Hey Dave, I understand you’re planning to teach Friday, but I hear things are really hectic so there’s no need for us to stay on the regular schedule. I’ll teach for you if you want.”

I had no idea when I’d get a chance to put a lesson together, to be completely honest. So I called her back and said if she wanted to teach, she could go right ahead. And I thanked her.

Hints taken. So much time doing stuff for God there’s no time to spend with God. So I skipped out on the meeting and now I’m not teaching Friday. I might even show up a little late, for good measure.

And now something completely different. This is starting to sound like the Stress Underground, not the Silicon Underground. So let’s talk about silicon.

Dan Bowman sent me a link to a suggestion that businesses buy old Mac clones, then dump $600 worth of upgrades into them so they can run Mac OS X and avoid paying $199 for a copy of Windows.

Yes, I know I’m teetering on the brink of mental illness here. So I’m assuming that if I were completely sane, this would make even less sense.

The best-selling software package for the Macintosh is (drum roll please)… Microsoft Office. So all you’ve accomplished so far is paying a little less money to Microsoft.

I’ve seen Mac OS X. I’ve tried to install Mac OS X. It wasn’t a pleasant experience. And this was a copy of Mac OS X that came with a brand-new G4. Mac OS X is not production-quality software yet. Not that that’s much of a problem. There’s precious little native software to run on it. For native software, you pretty much have to download and compile your own. If you’re going to do that, you might as well just run Linux, since it’s free for the asking and runs on much less-expensive hardware.

Most businesses are a bit hesitant to put Linux on the desktop yet. Some are starting to see the light. But a business that’s reluctant to put Linux on brand-new desktop PCs even when they can pay for good support they’ll probably never need isn’t too likely to be interested in buying a four-year-old Mac or Mac clone, plus 128 megs of obsolete and therefore overpriced memory plus a hard drive plus a disk controller plus a USB card, from five different vendors who will all point fingers at one another the instant something goes wrong. (And we’re talking Apple here. Things will go wrong.)

And yes, I know there are thousands of people who’ve successfully put CPU upgrades in Macintoshes, but it’s very hit-and-miss. I spent two of the most frustrating days of my life trying to get a Sonnet G3 accelerator to work in a Power Mac 7500. It either worked, failed to boot, or performed just like the stock 100 MHz CPU. Any time you turned it on, you didn’t know which of the three you would get. The local Mac dealer was clueless. I called Sonnet. They were clueless. I struggled some more. I called Sonnet back. I got a different tech. He asked what revision of motherboard I had. I looked. It said VAL4, I think. He told me he was surprised it worked 1/3 of the time. That accelerator never works right with that revision of motherboard. He suggested I return the card, or do a motherboard swap. Of course a compatible motherboard costs more than the accelerator card.

And of course there was absolutely no mention of any of this on Sonnet’s web site. At least you can go to a manufacturer of PC upgrades and read their knowledge base before you buy. Sometimes you can even punch in what model system you have and they’ll tell you if they work. Not that those types of upgrades make any sense when you can a replacement motherboard and CPU starts at around $150.

Suffice it to say I won’t be repeating that advice at work. I just got a flyer in the mail, offering me 700 MHz Compaq PCs preloaded with Win98, with a 15-inch flat-panel monitor, for $799. With a warranty. With support. Yeah, I’d rather have Windows 2000 or Windows XP on it. The only reason Compaq makes offers like that is to move PCs, so I’m sure they’d work with my purchasing guy and me.

Think about it. I can have a cobbled-together did-it-myself 400 MHz Mac refurb without a monitor for $700-$750. Or I can have that Compaq. That’s like getting a flat-panel monitor for 50 bucks. As far as usability and stability go, I’d rate Win98 and Mac OS X about equal. But for the time and money I’d save, I could afford to step up to a better version of Windows. Or I could bank the bucks and run Linux on it.

If you’re already a Mac zealot, I guess that idea might make sense. I’ve spent several years deploying, operating, and maintaning both Macs and PCs side-by-side in corporate environments. I have no great love for Microsoft. Most people would call my relationship with Microsoft something more like seething hatred.

But the biggest problems with PC hardware, in order, are commodity memory, cheap power supplies, proliferation of viruses, and then, maybe, Microsoft software. You can avoid the first two problems by buying decent hardware from a reputable company. (No, Gateway, that doesn’t include you and your Packard Bell-style 145-watt power supplies.) You can avoid the third problem with user education. (It’s amazing how quickly users learn when you poke ’em with a cattle prod after they open an unexpected attachment from a stranger. The biggest problem is getting that cattle prod past building security.) Microsoft software doesn’t exactly bowl everyone over with its reliability, but when Adobe recommends that Mac users reboot their machines every day before they leave for lunch, you know something’s up. Even Windows 95’s uptime was better than that.

A nice Labor Day.

Dave Farquhar General , , , , , , , , , , ,

Yesterday was nice. I got up late, then bummed around all day. I did a couple of loads of laundry, and I put a different hard drive in my Duron-750. Then I ignored my e-mail, ignored the site for the most part, and installed Wintendo (er, Windows Me) and Baseball Mogul. Around 6 I went out and bought a CD changer. My old 25-disc Pioneer died around Christmas time and I never got around to replacing it until now.
I knew I didn’t want another Pioneer. I’ve taken that Pioneer apart to fix it before, and I wasn’t impressed with the workmanship at all. And current Pioneer models are made in China. So much for those. I looked at a Technics and a couple of Sonys. Finally, swallowing hard, I dropped $250 on a 300-disc Sony model (made in Malaysia). I still suspect it’ll be dead within five years, but maybe it’ll surprise me.

I am impressed with the sound quality. It sounds much better than my Pioneer ever did. It’s really sad when you can tell a difference in sound quality between two CD players, but I guess that just goes to show how many corners Pioneer cut on that model. Next time I go CD player shopping, I’m going to bring a disc or two along to listen to in the store so I can hear the difference.

Anyhoo, I played two seasons of Baseball Mogul and guided Boston to two world championships and a boatload of money. But something happened that made me mad. I noticed over in the AL Central, Tony Muser’s Losers, a.k.a. the Kansas City Royals, were above .500, with essentially the same team that’s looking to lose 100 games this season. Well, there was no Donnie Sadler, Muser’s secret weapon, currently batting about .137 (which also seems to be about Tony Muser’s IQ, seeing as he keeps playing the guy). So the Royals minus Muser and Sadler were a .500 club. That’s nice to know.

Then, for 2002, Kansas City went and got the biggest free-agent bat they could afford. They also didn’t trade superstar right fielder Jermaine Dye, and they re-signed shortstop Rey Sanchez. And what happened? Well, the first round of the playoffs was a Boston-Kansas City affair, that’s what. I’d used the previous year’s windfall to buy myself an All-Star team, so I rolled over Kansas City in four games. I felt kind of bad about that, but it was partly because of my record against KC’s rivals that year that they made it that far, so not too bad.

It’s all I can do to keep from e-mailing Royals GM Allaird Baird and asking him why, if Tony Muser insists on playing Donnie Sadler every day, he doesn’t consider letting the pitcher bat and have the DH hit for Sadler instead.

And shocking news. HP is buying Compaq. I didn’t believe it either. Compaq’s recent problems, after all, were partly due to its purchase of Digital Equipment Corp. and its inability to digest the huge company. The only benefit I see to this is HP getting Compaq’s service division and eliminating a competitor–Compaq’s acquisition of DEC made more sense than this does.

Monday, 2 July 2001

Dave Farquhar General , , , ,

Some lucky people get a five-day weekend this week. Not me. I’m off Wednesday for Independence Day. About 30 years ago, my dad and his med school buddies used to go to the Missouri River and shoot bottle rockets at barges to celebrate. I’m not sure what I’ll get to do yet. Last year I had to work the 4th. That was a very nice paycheck, since I worked 60 hours that week anyway, on top of 8 hours’ holiday pay.
I found a use for absurdly high-speed CPUs this weekend. My Duron-750 can simulate a 30-team, 162-game baseball season in its entirety in roughly 3 minutes. Of course a faster CPU will do it even faster. Baseball simulation is very CPU-intensive and very disk-intensive. This 750 has a SCSI disk subsystem in it too. It’s old, but I suspect SCSI’s ability to re-order disk requests for speed helps. I haven’t swapped in an IDE drive to see if it makes a difference. So if you’re a statistical baseball junkie, you can actually justify an insanely fast CPU. It feels strange to call the cheapest CPU on the market today insanely fast, but for most things, the Duron-750 really is.

The other use I’ve found for these CPUs is emulating a 50 MHz 68060-based Amiga at full speed. A Duron-750 isn’t quite up to that task.

I talked about PartImage last week. I used it over the weekend to clone 7 PCs. My church’s sister congregation bought 8 Compaq Deskpro EXs earlier this year and just finished the room they’re going in. So I went in, set one of them up (and tweaked it out, of course–the first reaction of one of the members: “Wow, that sure boots fast!”).

Sadly, many companies seem to use non-profit organizations as a way to just get rid of their junk. Here are some of the jewels this church has been “blessed” with: two 386sx laptops with dead batteries and no power adapters, two XTs, two 286s, a pile of 386sxs, and three 486s. Two of the 486s are old Compaq ProSignia servers with big SCSI hard drives, so I can slap in an ISA NIC and install Linux on one of them and make it a file server. The only thing remotely useful that anyone’s ever given them is a pair of Pentium-75s. But one of the 75s had a 40-meg hard drive in it. That’s the better of the two, though. The other had no hard drive, no memory, and no CMOS battery.

Oh, and I shouldn’t forget the large quantities of busted monitors. They’ve got a room full of monitors. About three of them work. What’s anyone going to do with a bunch of monitors that don’t work? Legally, the church can’t throw any of this stuff away (and shouldn’t) because of all the lead content, which makes them hazardous waste. But the church can hardly afford to pay someone to take it away and dispose of it properly either. We’re talking an inner-city church here. Can you say, “blaxploitation?” I knew you could.

The Pentiums did at least come in standard AT cases though, and nice ones at that. They look like Enlights, but they had Sparkle power supplies in them, Whatever the make, they’re nice and thick so they don’t slice you, there’s lots of wide open space inside, and they have 7 drive bays. So I grabbed the diskless Pentium to make into a router/Squid server/content filter. I ripped out the P75 board and dropped in an AT Soyo Socket 370 board with a Celeron-366 on it. It’ll be fabulous.

The best I can do with most of these systems is to try to make X terminals out of them, assuming I can find a machine beefy enough to host StarOffice for a half-dozen systems. It may not be worth the bother.

One of the 386s had a 420-meg hard drive in it for some reason, so I pulled that drive, hooked it up to the first of the Compaqs, and used PartImage to dump it. I used 480 megs on the drive, so with Gzip compression, the image left just 12 megs free on the drive. Tight fit, but we were OK. Then I just ran around to each of the others, hooked up the drive, and pulled the image. I took the drive home with me so I could burn a CD from it.

That’s good use of free software.

Fare thee well, Alpha…

Dave Farquhar General , ,

Fare thee well, Alpha, if Intel will allow it… I’m not holding my breath. The greatest CPU of all time, by a long shot, looks to be no more. Compaq pulled the plug on Alpha, essentially selling it out to Intel. Another case of far superior technology dying because its owner had no idea what to do with it.
But Intel, with its terminal case of NIH syndrome, isn’t about to let some other company’s technology thrive if history is any indicator.

One story in The Register likened replacing Alpha with Intel’s Itanium as replacing a Ferrari with a Yugo.

In my ill-fated Integrating Linux Servers with Windows Networks, I stated that we’ll miss NT on Alpha, and I presented Linux as a way to preserve a shop’s investment in Alpha-based servers. Indeed, an Alpha running Linux ought to be useful for many, many years, with or without Compaq’s blessing.

I guess I should be happy to see that IBM isn’t the only company that comes up with great ideas and great technology and then has no idea what to do with it. But I’m not. I hate seeing good engineering go to waste, especially when the beneficiaries of that waste are perpetual underachievers like Microsoft and Intel.

Minesweeper is murder.

Dave Farquhar General , , , , ,
Minesweeper is murder.

Minesweeper is murder. An activist group is asserting that the Windows game Minesweeper is disrespectful of victims of land mines and should be removed and replaced with a game about flowers. I have no idea if these guys are serious or not. I never liked the game anyway and just always wanted an excuse to say “Minesweeper is murder.” So now I’ve said it three times. I’m happy.

Read more

Craig Mundie’s infamous speech

Dave Farquhar General , , , , , , , , , , , , , , , , ,

I haven’t said anything about Microsoft Executive Craig Mundie’s speech yet. Everyone’s heard of it, of course, and the typical response has been something along the lines of “Now we know Microsoft’s stance on Open Source.”

No, we’ve always known Microsoft’s stance on that. They’re scared of it. Remember the stereotype of open-source programmers: college students and college dropouts writing software in their basements that a lot of people are using, with the goal of toppling an industry giant. Seem far-fetched? Friends, that’s the story of Microsoft itself. Microsoft became an underground sensation in the late 1970s with Microsoft Basic, a programming language for the Altair and other kit computers and later for CP/M. And while we’ll probably never know the entire story of how and why this happened, when IBM decided to outsource the operating system for the IBM PC, they went to Microsoft and got both an OS and the must-have Microsoft Basic. Ten years later, IBM was just another hardware maker–really big, but getting squeezed. Today, 20 years later, IBM’s still a huge force in the computing industry, but in the PC industry, aside from selling ThinkPads, IBM’s a nobody. There may be hardware enthusiasts out there who’d be surprised to hear IBM makes and sells more than just hard drives.

Ironically, Microsoft’s response to this new threat is to act more and more like the giant it toppled. Shared Source isn’t a new idea. IBM was doing that in the 1960s. If you were big enough, you could see the source code. DEC did it too. At work, we have the source code to most of the big VMS applications we depend on day-to-day. Most big operations insist on having that kind of access, so their programmers can add features and fix bugs quickly. If Windows 2000 is ever going to get beyond the small server space, they really have no choice. But they do it with strings attached and without going far enough. An operation the size of the one I work for can’t get the source and fix bugs or optimize the code for a particular application. You’re only permitted to use the source code to help you develop drivers or applications. Meet the new Microsoft: same as the old Microsoft.

Some people have read this speech and concluded that Microsoft believes open-source software killed the dot-com boom. That’s ludicrous, and I don’t see that in the text. OSS was very good for the dot-com boom. OSS lowered the cost of entry: Operating systems such as FreeBSD and Linux ran on cheap PCs, rather than proprietary hardware. The OSs themselves were free, and there was lots of great free software available, such as the Apache Web server, and scripting languages like Python and Perl. You could do all this cool stuff, the same cool stuff you could do with a Sun or SGI server, for the price of a PC. And not only was it cheaper than everybody else, it was also really reliable.

The way I read it, Microsoft didn’t blame OSS for the dot-com bust. Microsoft blamed the advertising model, valuing market share over revenue, and giving stuff away now and then trying to get people to pay later.

I agree. The dot-com boom died because companies couldn’t find ways to make money. But I’m not convinced the dot-com boom was a big mistake. It put the Internet on the map. Before 1995, when the first banner ad ran, there wasn’t much to the Internet. I remember those early days. As a college student in 1993, the Internet was a bonanza to me, even though I wasn’t using it to the extent a lot of my peers were. For me, the Internet was FTP and Gopher and e-mail. I mostly ignored Usenet and IRC. That was pretty much the extent of the Internet. You had to be really determined or really bored or really geeky to get much of anything out of it. The World Wide Web existed, but that was a great mystery to most of us. The SGI workstations on campus had Web browsers. We knew that Mosaic had been ported to Windows, but no one in the crowd I ran in knew how to get it working. When we finally got it running on some of our PCs in 1994, what we found was mostly personal homepages. “Hi, my name is Darren and this is my homepage. Here are some pictures of my cat. Here’s a listing of all the CDs I own. Here are links to all my friends who have homepages.” The running joke then was that there were only 12 pages on the Web, and the main attraction of the 12 was links to the other 11.

By 1995, we had the first signs of business. Banner ads appeared, and graduating students (or dropouts) started trying to build companies around their ideas. The big attraction of the Web was that there was all this information out there, and it was mostly free. Online newspapers and magazines sprung up. Then vendors sprung up, offering huge selections and low prices. You could go to Amazon.com and find any book in print, and you’d pay less for it than you would at Barnes & Noble. CDNow.com did the same thing for music. And their ads supported places that were giving information away. So people started buying computers so they could be part of the show. People flocked from closed services like CompuServe and Prodigy to plain-old Internet, which offered so much more and was cheaper.

Now the party’s ending as dot-coms close up shop, often with their content gone forever. To me, that’s a loss only slightly greater than the loss of the Great Library. There’s some comfort for me: Five years from now, most of that information would be obsolete anyway. But its historical value would remain. But setting sentiment aside, that bonanza of freebies was absolutely necessary. When I was selling computers in 1994, people frequently asked me what a computer was good for. In 1995, it was an easier sell. Some still asked that question, but a lot of people came in wanting “whatever I need to get to be able to get on the Internet.” Our best-selling software package, besides Myst, was Internet In A Box, which bundled dialup software, a Web browser, and access to some nationwide provider. I imagine sales were easier still in 1996 and beyond, but I was out of retail by then. Suddenly, you could buy this $2,000 computer and get all this stuff for free. A lot of companies made a lot of money off that business model. Microsoft made a killing. Dell and Gateway became behemoths. Compaq made enough to buy DEC. AOL made enough to buy Time Warner. Companies like Oracle and Cisco, who sold infrastructure, had licenses to print money. Now the party’s mostly over and these companies have massive hangovers, but what’s the answer to the Ronald Reagan question? Hangover or no hangover, yes, they’re a whole heck of a lot better off than they were four years ago.

I’m shocked that Microsoft thinks the dot-com phenomenon was a bad thing.

If, in 1995, the Web came into its own but every site had been subscription-based, this stuff wouldn’t have happened. It was hard enough to swallow $2,000 for a new PC, plus 20 bucks a month for Internet. Now I have to pay $9.95 a month to read a magazine? I could just subscribe to the paper edition and save $2,500!

The new Internet would have been the same as the old Internet, only you’d have to be more than just bored, determined, and geeky to make it happen. You’d also have to have a pretty big pile of cash.

The dot-com boom put the Internet on the map, made it the hot ticket. The dot-com bust hurt. Now that sites are dropping out of the sky or at least scaling operations way back, more than half of the Web sites I read regularly are Weblogs–today’s new and improved personal home page. People just like me. The biggest difference between 1994 and 2001? The personal home pages are better. Yeah, the pictures of the cat are still there sometimes, but at least there’s wit and wisdom and insight added. When I click on those links to the left, I usually learn something.

But there is another difference. Now we know why it would make sense to pay for a magazine on the Internet instead of paper. Information that takes a month to make it into print goes online in minutes. It’s much easier and faster to type a word into a search engine than to leaf through a magazine. We can hear any baseball game we want, whether a local radio station carries our favorite team or not. The world’s a lot smaller and faster now, and we’ve found we like it.

The pump is primed. Now we have to figure out how to make this profitable. The free ride is pretty much over. But now that we’ve seen what’s possible, we’re willing to start thinking about whipping out the credit cards again and signing up, provided the cost isn’t outrageous.

The only thing in Mundie’s speech that I can see that Linus Torvalds and Alan Cox and Dan Gillmor should take offense to is Microsoft’s suspicion of anyone giving something away for free. Sure, Microsoft gives lots of stuff away, but always with ulterior motives. Internet Explorer is free because Microsoft was afraid of Netscape. Outlook 98 was free for a while to hurt Lotus Notes. Microsoft Money was free for a while so Microsoft could get some share from Quicken. It stopped being free when Microsoft signed a deal with Intuit to bundle Internet Explorer with Quicken instead of Netscape. And there are other examples.

Microsoft knows that you can give stuff away with strings attached and make money off the residuals. What Microsoft hasn’t learned is that you can give stuff away without the strings attached and still make money off the residuals. The dot-com bust only proves that you can’t necessarily make as much as you may have thought, and that you’d better spend what you do make very wisely.

The Internet needs to be remade, yes, and it needs to find some sustainable business models (one size doesn’t fit all). But if Mundie thinks the world is chomping at the bit to have Microsoft remake the Internet their way, he’s in for a rude awakening.

More Like This: Microsoft Linux Weblogs Internet Commentary