Troubleshooters.Com Presents

Troubleshooting Professional Magazine

Volume 4 Issue 11, November 2000
Annual Linux Showcase!
Copyright (C) 2000 by Steve Litt. All rights reserved. Materials from guest authors copyrighted by them and licensed for perpetual use to Troubleshooting Professional Magazine. All rights reserved to the copyright holder, except for items specifically marked otherwise (certain free software source code, GNU/GPL, etc.). All material herein provided "As-Is". User assumes all risk and responsibility for any outcome.

[ Troubleshooters.Com | Back Issues ]

Steve Litt is the author of Troubleshooting Techniques of the Successful Technologist and
Rapid Learning: Secret Weapon of the Successful Technologist.


Editors Desk

By Steve Litt
Whew! This was the third biggest Troubleshooting Professional Ever. Annual Linux Showcase (formerly Atlanta Linux Showcase) was a big show leading to big ideas. The size of this issue caused a few sections to be a little rough -- please forgive me. And my interviewing skills obviously need to be upgraded :-)

All you Troubleshooting Process readers, hang tough. The December issue will discuss the broad topic of problem solving, together with problem solving theory and some specific problem solving methodologies, situations they're optimized for, how they relate, and how they differ. I've been writing the December issue for the last 3 months, so you know it's going to be good.

The January issue will be on the topic of perceived similarities, and how that can lead to enhanced learning, adaptivity and productivity. So stay tuned.

And of course, even this issue is of vital interest to Troubleshooters. As I've repeated consistently since the 1/1998 issue of TPM, you can't fix intermittent junk. GNU/Linux gives us an alternative that can be troubleshot (when necessary) using standard Troubleshooting procedures (such as the Universal Troubleshooting Process), instead of the traditional Windows Reboot,Reapply,Reinstall methodology. So whether you're an Open Source advocate, a GNU/Linux enthusiast, or a system-independent Troubleshooter, this is your magazine. Enjoy!

Steve Litt is the main author of Samba Unleashed. He can be reached at Steve Litt's email address.

ALS -- Yes!

By Steve Litt
The morning after an all day drive, I walked into the Cobb Galleria to register for the 4th Annual Linux Showcase. Immediately, Monica Ortiz introduced herself to me. Monica is the marketing and publicity person for Usenix, and it was she who had arranged my press pass.

Monica's a genuinely nice person. ALS was incredibly intense, but Monica made everyone feel welcome and comfortable. It turns out all of ALS was like that.

ALS used to stand for ATLANTA Linux Showcase, but it grew to the point where Atlanta Linux Enthusiasts (ALE) could no longer run it alone. So they brought in Usenix to help. Usenix is the Advanced Computing Systems Association, headquartered in Berkeley California. They put on shows all the time. That's one of the many things they do. Anyway, that's why ALS now stands for ANNUAL Linux Showcase -- it won't always be in Atlanta. The current plan is for ALS to alternate between the east and west coasts.

I sat next to two of the ALS volunteers during Dave and Busters party on Thursday night, and what a story they told. Did you know that ALS started as an installfest? Yep, in 1996 it was an installfest. And it just kept on growing. These guys remembered the bad old days when there wasn't enough money to hire people for the registration desk, so they had to man the registration desk while everyone else enjoyed themselves. Talking to these two guys (and I sure wish I could remember their names), my opinions on the registration fees for various Linux shows mellowed. Booking hotels months or years in advance, booking speakers and companies long in advance, paying hotel room guarantees -- it's not easy.

But ALE and Usenix sure made it look easy. Great exhibitors, great Birds of a Feather sessions, and, I assume from the 1 or 2 tech sessions I managed to attend, great tech sessions.

ALS was incredibly dense in terms of things to do and people to meet. Many times I found myself needing to choose between a keynote speech, a Birds of a Feather session, a party, or dinner :-). Life was interesting from 9:00am til midnight. And the people! I'm not going to tell you "everyone" was there, because the days of "everyone" being at a Linux show are gone. But Ken Coar of Apache fame was there, as was Perl's Larry Wall, and Eric Raymond of Cathedral and Bazaar fame. Red Hat, IBM, Debian project, Borland, Sun, SuSe, VA Linux, VMware, XFree86 project, AbsoluteValue Systems, Inc., ActiveState, Cambridge Computer Systems, Compaq, FSF, the Free Standards Group, and HP were there. Those were the tip of the iceberg -- there were many, many more. Many of the Progeny Distro's developers were there, in the Debian booth. LUG members met, hooked up, and planned the future.

I couldn't attend the Saturday sessions -- Exhaustion had set in.. I transcribed my notes, started this issue of Troubleshooting Professional Magazine, and hit the road. And I smiled all the way back. In the rich tradition of LinuxExpo, Apachecon and ITSA day, ALS had been an outstanding experience.

You could get an exhibit-only pass for ALS free of charge. The tech sessions fees were $260 for three days, less for Usenix members (you could become a member on site and take advantage of the lower price). The tutorials were a little pricey, but considering the subject matter and the caliber of the instructors (many from Red Hat, including Bryan Andregg, Red Hat's Director of Networks), I'd imagine it would be money well spent.

If you EVER find an opportunity to attend an ALS show, do whatever you have to do to get away, and go. You'll be glad you did.

Steve Litt is the documentor of the Universal Troubleshooting Process. He can be reached at Steve Litt's email address.


By Steve Litt
Bradley Kuhn's interview was conditional on my calling the operating system we've all grown to love "GNU/Linux" in the article containing his interview, but not necessarily in other articles. A very fair position. In fact, a moderate position.

GNU/Linux is comprised of the Linux kernel originally crafted by Linus Torvalds, plus many, many utilities, a large number of which were utilities from the original GNU project. "GNU/Linux" is probably the most accurate moniker one can give to the operating system. Please be aware that in all of Troubleshooters.Com, when I say "Linux" I really mean "GNU/Linux". I completely believe that without the GNU project, without the GNU Manifesto and the GNU/GPL license it spawned, the operating system the press calls "Linux" never would have happened.

I'm part of the press and there are times when it's easier to say "Linux" than explain to certain audiences that "GNU/Linux" is the same as what the press calls "Linux". So I abbreviate. Additionally, I abbreviate in the same way one might abbreviate the name of a multi-partner law firm. But make no mistake about it. In any article in Troubleshooting Professional Magazine, in the whole of Troubleshooters.Com, and even in the technical books I write, when I say "Linux", I mean "GNU/Linux".

There are those who think FSF is making too big a deal of this. Nothing could be farther from the truth. The GNU General Public License, combined with Richard Stallman's GNU Manifesto and the resulting GNU-GPL License, are the only reason we can enjoy this wonderful alternative to proprietary operating systems, and the only reason proprietary operating systems aren't even more flaky than they are now. Last but not least, it's significant to note that in Infoworld's October 6, 2000 article entitled "E-business innovators", author Mark Leon names GNU's Richard Stallman as the innovator associated with GNU/Linux.

Steve Litt is the author of Rapid Learning: Secret Weapon of the Successful Technologist. He can be reached at Steve Litt's email address.

Vote for Your Favorite Troubleshooting Professional

By Steve Litt
Now is the time to vote for your favorite Troubleshooting Professional issue (by year and month) and/or article (by year, month and title). Submit your entry to Steve Litt's email address. The results will be announced in our fourth anniversary issue, which comes out in January.

Also, we're having our first annual Troubleshooting Professional Magazine trivia contest. Simply name the issue (by year and month) and article names for the ten different articles containing the these ten phrases:

  1. Gary Kildall has reached out of the grave and grabbed Bill Gates' ankle.
  2. Then, just for a second, a big red thing with three silver eyeballs and huge yellow fangs appears in the closet door.
  3. A few days later, as Communism crumbled throughout Europe, The Attitude became the third Troubleshooting Tool.
  4. I've been the worst, and I've been the best, and the only difference is what I learned about Troubleshooting Process.
  5. There's slushy snow on the ground, and like a couple fourth graders the tycoon and the pauper trudge through the muck.
  6. Like so many before me, I rumbled down the highways paralleling the old route 66 with nothing but an ancient overpacked car and the age-old dream of jobs and sunshine where old 66 ended at a sign reading "Santa Monica Pier".
  7. The girls are all high school seniors (some for the second time), who look 30, walk with too much slink and smile with too much smirk.
  8. Smiles becomes wide-eyed surprise as the wiseguys in the Mustang see me pass, with the front tires a few inches off the ground.
  9. And above all, if you don't want to see him cry, don't tell him you set up an Intranet with virtual hosting, DNS and full CGI (which can interface to Oracle or Sybase or free PostgreSQL), in four hours on a $600 computer.
  10. Richard Stallman wrote the manifesto. Linus Torvalds proved it worked.

Anyone answering all 10 correctly will be mentioned in the January 2001 issue, and if they'd like, will have an email link and an http link from the 1/2001 issue. If nobody gets all 10, the person with the most correct will get their name, and if desired email and/or http links.

Steve Litt is the author of Troubleshooting Techniques of the Successful Technologist. He can be reached at Steve Litt's email address.

Jon "maddog" Hall

By Steve Litt
Jon "maddog" Hall had made himself available for press interviews from 9am to 10am on Wednesday, 10/12/2000, and I had just one question for him. Basically, I wanted to know how he viewed Linux's future, especially in light of Corel's "strategic alliance" with Microsoft. I expected a 5 minute answer. What I got was 20 minutes of good news.

Maddog said that last year there were 20 million Linux systems, and that IDC said it would be almost double this year. Right now IDC estimates there are 30 million. There's been a huge explosion in embedded device drivers, due to Linux stability and modifiability.

Maddog mentioned that IDC has said Linux surpassed Apple on the Desktop. If you believe that Macintosh is viable on the Desktop, then logic says that Linux is more viable on the desktop.

Maddog discussed the International front, where Linux usage is seeing huge increases in India, where Internet usage grows at double-digit rates, and in Israel. China has chosen Linux and works with four vendors. Linux is very strong in Germany, where the Linuxtag show drew 17,000 attendees. Mexico's Scholarnet initiative also came up in our discussion.

Maddog didn't know all intricacies of the Corel "strategic alliance", musing that maybe it's just a strategy for Microsoft's .Net initiative. He hopes Corel continues to turn out Linux products. It was during the Corel discussion that maddog cut loose with the best quote I heard during the course of ALS. After mentioning that he never thought a Corel distro was good or necessary, and that their WordPerfect Suite and CorelDraw products were very beneficial to Linux, he said you can't make money selling Yet Another Distro. Then he cut loose with the following:

"The real way to make money in Linux is not to create another distribution, but to use the existing Linux distributions in a value added model."

Jon "maddog" Hall, 10/12/2000

At that point I asked him for an example. He asked if it wouldn't be nice to have an Internet connected refrigerator that could call the repair company when the motor performance was beginning to wane, so they could fix it before your $500 worth of groceries went bad.

After a minute of thought a lightbulb went off. I asked him if he meant that you don't sell Linux or Linux software to obtain money, but instead Linux is a vehicle for making money. He assured me that's what he meant. I'll further discuss this point in the next three articles of this magazine.

We wrapped up the interview with maddog's assessment of the state of Linux. Maddog was extremely positive about Linux's progress into the Embedded, Supercomputer, and Server arenas. His lone area of reticence was the desktop, which he felt was still waiting for apps, and that there is really no way to rush that -- it will happen slowly.

We concluded the interview with a little small talk and went our merry ways. Jon "maddog" Hall is a really nice guy. He's the type of guy who makes the Linux community a community.

Editors Note: In a later email Maddog commented that you can make money selling a Linux distribution IF that distribution has value added, but not just by selling Yet Another Distro (YAD).
Steve Litt is the editor and content lead for Troubleshooting Professional Magazine. He can be reached at Steve Litt's email address.

Further Thoughts on Jon "maddog" Hall's Value Added Model

By Steve Litt
Let me repeat Jon "maddog" Hall's quote:
"The real way to make money in Linux is not to create another distribution, but to use the existing Linux distributions in a value added model."

For those of you who are long time Troubleshooting Professional Magazine readers, you know the sharp distinction I make between tools and solutions. I, and many other Open Source advocates, have been treating Linux as the solution. Wrong! It's the tool you use to obtain the solution. When your product needs a good operating system, what are you going to pick -- an overpriced, underperforming piece operating system, or a free, powerful and stable OS for which you have complete source access. It's your product, your market, and your money on the line -- which tool will you choose? More and more, Linux is the tool of choice.

This is great news for those of us living outside the big tech metropoli -- California, Colorado, Omaha, Chicago, Texas, Atlanta and the northeast USA. The "give away the product, sell the service" model doesn't work too well if you need to fly to the big city to deliver your services. Let's face it, a lot of us want to live in nice small town, and make money shipping a product around the world. Now we can. Linux can be added to hardware to produce a shippable product, or in many cases it can be added as a platform to your proprietary software. And for further customer convenience, it can be preloaded on hardware so that you deliver an "appliance". Most such Linux "appliances" run for months or years without attention.

Of course, Linux desktop apps aren't ready for the value added model yet, because the vast majority of desktop users use Windows. But traditionally, few have made money on Windows desktop apps. Instead, add value with a back end product.

In this era of web apps, why not make your slick time and billing a Linux based web app, using SQL to hit the database (any database) and HTML/Javascript, CGI, Python or Java for the user interface. Works just great on your customer's Windows desktops, but exhibits the speed, power, and ease of development you'd expect from Linux. And you don't become Bill Gates unpaid MS license salesman.

If you're making embedded devices, consider Linux. Linux can now be pared down to a 4MB footprint. That could make a mighty interesting cell phone, complete with a file system and extensive customization capability. And no license fees. With a little improvement in voice recognition technology, computers and cell phones could merge.

Steve Litt is the author of Troubleshooting Techniques of the Successful Technologist. He can be reached at Steve Litt's email address.

Google, the Perfect Value Added Example

By Steve Litt
It was Friday the 13th. I was wandering around after the SSH and Wine tech sessions, and I happened to wander into Jim Reese's talk on Google. Nothing unlucky about that!

After a quick review of Google basics (search engine, only a search engine, the best search engine), Jim described Google's overall architecture. Google uses 6000 boxes in three datacenters -- one on the U.S. East Coast, and 2 on the U.S. West Coast.

In each data center, a load balancer farms out search requests to appropriate web servers. Their software decides which "server" is appropriate, based on many things including proximity to the incoming request, machine load, etc. The request can be sent to a server in a different datacenter. Each web server connects to multiple index servers and multiple page servers. The page servers hold cached copies of every spidered page. The index servers hold a section of the indexes. Google algorithms quickly determine which index server holds the appropriate search term and query that index server. Jim called the distribution of the index database between many servers "sharding", and stated that Google's indexes were too huge for any one box to hold.

Just one more thing. Everything described in the preceding paragraph is made redundant multiple times. Jim stressed redundancy, redundancy, redundancy. Power problems, natural disasters, security attacks and just plain malfunctions can reduce Google's throughput (making it slower than its usual <0.5 second), but the redundancy guarantees that nothing will shut Google down.

This is important because some of Google's revenue comes from special indexing and special search capabilities they give large customers. Basically, one of the things Google does for money is outsource search functionality of large sites. These large sites needn't hire developers to create a site specific search engine, nor devote machine resources for that site specific search engine. They simply outsource the entire search functionality to Google.

Jim spent some time on Google's unique benefits to searchers. The main benefits are speed and relevancy. He demonstrated how an Altavista search could pull up a bunch of fluff before the home page of the search term he used, while Google had the search term's home page on top. And anyone who has used Google knows its speed, which is typically less than a half a second. In fact, Jim mentioned that Google has various internal reporting functionalities that alarm when search times start exceeding a half second.

Jim outlined the creation of Google. A couple Ph.D candidates, Larry Page and Sergey Brin, created a superior search engine as a part of their studies. It ran on a couple of boxes under their desks. And of course I86 boxes running Linux were their platform of choice. Cheap, reliable, customizable. Linux as a tool -- value added! Pretty soon everyone at their school was using their search engine. It was superior.

A professor handed them a check and told the Ph.D candidates they were now in business. A business plan was made, employees hired, hardware procured, and the search engine was scaled massively. And a platform needed to be chosen.

I don't remember Jim mentioning the possibility of Windows as a platform. Sun and SGI were considered, but those choices would have been fabulously expensive. Also, those computers cannot be packaged as densely as Linux boxes. When renting datacenters, you pay per square foot. They chose commodity priced 1U Linux rack mount boxes. They can load 80 of them into a 7 foot rack, 2 to a 1U unit. They mount fans on the top to suck air up a shaft in the center, cooling the computers.

Litt Reads Between the Lines

Now let me interpret Jim Reese's talk using Jon "maddog" Hall's value added model.

Google had a job to do. They needed to economically create the world's finest search engine. They had some venture capital. They had (and soon hired more) the in-house brainpower to write the software. But they needed a platform on which to deploy their software.

I suppose they could have used NT or one of the Windows2000 server editions. If they'd wanted to spend the money, and then code around all of Windows' quirks.

They could have used Sun or SGI boxes. Jim Reese seemed to imply those platforms were at least briefly considered. But it would have been expensive iron, scaling wouldn't have been as granular, and they couldn't have mounted them 80 to the rack. And then those operating systems are not free, and when you're talking 6000 computers, even an inexpensive OS can be expensive.

What tool could they use to deploy their system on inexpensive i86 boxes? OK, let's all say it in unison -- GNU/Linux! Free. Configurable. And if there's something you don't like about it, you've got the source code and the legal right to modify it.

If I read between the lines, Google didn't ask "how can I make money with Open Source". Sounds to me the plot went something more like this -- they knew they could build a killer search engine, and asked themselves "I need hardware and OS to host my system. Where can I find a hardware/OS combination that's inexpensive, reliable, powerful, granularly scalable, capable of massive redundancy, and free from proprietary licenses that might someday get in the way".

I think the message we can all take from Jon "maddog" Hall, as proven by Google, is that Linux and Open Source Software has given us a huge, cheap, quality factor of production we didn't have before. It's the equivalent of finding a massively abundant source of cheap, clean and safe energy -- it changes everything, and if you use that factor of production and your competitors don't, you win.

So make a list of all the ideas you once cast off as impractical because the hardware or operating system software would have been too expensive. That list is a gold mine.

Steve Litt is the maintainer of the UMENU and htmlslides GPL projects. He can be reached at Steve Litt's email address.

Yet Another Smooth Example

By Steve Litt
So back at LEAP-CF during our monthly meeting in Orlando, Phil Barnett gives a presentation on Smoothwall. Smoothwall is a Linux distro tweaked to be a firewall, a whole firewall, and nothing but a firewall. Installation is via CD, whose image you can download and burn. You then boot the cd on an old cheapo box (min 386 with 8meg ram and 100meg hard disk), answer a few trivially easy questions, and you have a firewall. Make sure all clients using the firewall list the firewall's IP address as their gateway.

Under the hood, Smoothwall is a GPL product. It's a self-contained, stripped down Linux distro derived from the VA-Linux 6.2.1 Linux distro, which is itself derived from Red Hat 6.2.

Smoothwall is a very young product, so right now it supports only modems. But soon enough it will support cable modem, dsl and isdn. "Soon enough" isn't the same as proprietary software's vaporous "real soon now". Smoothwall development runs on free software time -- there are people working on it day and night. And if one of you readers decides to go in and patch the source for cable -- that's it, it's done. So even if you can't use Smoothwall today, stay tuned to Smoothwall's website so when the time comes, you're the first on your block to use it for cable or dsl.

Look at the opportunity here. In certain environments, 8meg 386's are available for $20 apiece. Pop in a $20 network card, install Smoothwall, and you have a *high quality* firewall for $40.00. Unlike some of today's cheap firewall appliances, this is a firewall to which you can apply the latest security patches. This is one that you can modify yourself, if the fancy strikes you. If you offer networking services to small business, this is your chance to offer firewalls without becoming the unpaid salesman for a proprietary outfit. Whatever you charge, you pocket. Simply let your clients know not to dispose of their antiquated boxes -- you'll turn the next one into a firewall.

Don't worry about maintenance costs. The Smoothwall installer CD (plus boot floppy if necessary) asks a half dozen questions, detects your network card, and forms a firewall. You don't even need a CD drive -- you can do an NFS install off a CD iso image loopback mounted on a different Linux box. After completing the Smoothwall install, you need to set the gateways of client machines to the IP address of your Smoothwall box. Life can be even easier if the clients are DHCP clients. Just tell Smoothwall to be its own DHCP server, and it gives the clients that gateway address. Plug and play, where do you want to go today? To Open Source, for enhanced profit.

Another example of the Jon "maddog" Hall "business model". Once again,  the cheap price, freedom, and minimal hardware requirements of Open Source create an incredibly cheap factor of production that you can exploit to lower prices while increasing profit. And if your competitors refuse to use this new cheap factor of production, you can walk on their graves.

Steve Litt is the author of Rapid Learning: Secret Weapon of the Successful Technologist. He can be reached at Steve Litt's email address.

Peter Salus, The Historian

By Steve Litt
ALS was very nice to the press. They had a press room set up, and they published a schedule of people available for interview in the press room. That schedule listed Peter Salus as the USENIX Historian. I always listen to historians. Those who forget history are bound to...

Before getting to the interview, I should tell you that Peter is a former Vice President of the Free Software Foundation, the author of "A Quarter Century of UNIX" (Addison-Wesley 1994) and "Casting the Net" (A-W 1995). He is currently the Director of The Tcl/Tk Consortium. Peter's a busy man, but besides all that he's also the Chief Knowledge Officer of Matrix.Net. You probably want to know what Matrix.Net is. So did I.

Matrix.Net evaluates throughput on the Internet. That means they:

  1. Detect latency problems
  2. Find the bottleneck
  3. Find alternative routes
Matrix.Net has sensors and injectors. They call the injectors "beacons". The beacons are placed strategically all over the net. Using mathematical software, they continually analyze the packets issued by and detected by these beacons, forming a sort of "sonar" image of the net -- where it's moving fast, and where it's moving slow. They also use ping lists to evaluate the 'net.

Get this. Right now Matrix.Net can detect an Internet slowdown within 15 minutes of inception. They are working very hard to lower this figure to 5.

If you're a large company whose business depends on low latency communications with customers or business partners, you can hire Matrix.Net to evaluate your performance. Not just the performance of your boxes, network, or outgoing trunk lines. Anyone can do that. Matrix.Net can evaluate the entire route your communications take to and from their destinations. And by discovering the bottlenecks, they can tell you how to fix the bottleneck or route around it. In a world where surfers give up after 10 seconds, and a lost minute in a stock trade can cost thousands and create lawsuits, this is an absolutely vital service.

My one other observation about Peter is that he's a nice guy and extremely fun and interesting to talk with. Now that you know something about Peter and the many hats he wears, here's Peter's interview:

TPM: My question is, I'd like to hear your perspective on Linux, where it's going, from a historical point of view, kind of extrapolating forward, and you know, I see the the thing that just happened with Microsoft and Corel. To me that's alarming. I see the thing with a license change on Python as alarming, those are my opinions of course. But anyway I'd like to get your perspective on this.

Peter: First of all let me say that the Corel thing is alarming. Python is not as alarming. And the analogy I would make is one that's just over 20 years old. 1979 in June, at a Usenix meeting in fact. AT&T announced its repricing structure and its more restrictive licensing, pushing academic institutions up to $7500 per CPU, which in 79 meant a hell of a lot of money because you didn't have desktop workstation type things. The result of it was indeed the proliferation of things like Berkeley Unix etc. in order to get out from the restrictions that AT&T was placing on things. So that where the Python question is concerned, I think that it may well cause things like perl,  tcl, etc., other scripting languages, other ways of doing things, to have greater success, and it may cause a wane in Python because the movement to just use tools -- you're not going to be able to go out and patent the screwdriver and the hammer. Tools are there for people to use. And implementations of tools are important to the community.

But let me go back, you asked for history, that's why 1979 is in there. The 1979 thing in fact gave direct rise to Linux in general. Admittedly a decade intervened, but it's direct, because one of the people there in Toronto in that audience was Andy Tannenbaum, who had been planning to use UNIX for teaching operating systems in Free University in Amsterdam. He knew he couldn't get the administration in Amsterdam to pay for 30 or 40 licenses at $7500 each. And so he went home and he wrote a replication of UNIX, license free, called Minix, which ran on the 286 box and on a number of other things, and which was a very very successful teaching tool for a decade. It was Minix that Linus Torvalds was unhappy about in 1991, that caused him to sit down and write Linux. So that one has a straight stream for the imposition of stupid licensing regulations and the creations of other stuff. I think the really important thing to recognize is that the bright kids, and I call them kids because I'm an old person in the field, but the thing is that the ones in high school, in college, in graduate school, who work as volunteers for all sorts of things -- for Apache, for Perl, for Linux in general, for the GNU people, for FSF, are much more interested in creating something interesting and useful than they are in caring about who owns the license on X, Y and Z, and I think the kids are brighter than the suits who do the other things. The result of that is that they will always think of ways around. The licensing restrictions are not going to cause a lack of innovation. Now, the thing is that Microsoft as a corporation is an enormous and successful corporation. It makes gazillions of dollars even when if I look at the stock pages its overall  price has dropped by over 50% in a year, in the last year, it nonetheless is an enormously profitable enterprise. However, when one then looks at rather than control of the desktop, control of systems, they're actually not being able to make inroads into it. The interesting thing is that when you look at the big web surfers, there's an overwhelming flow towards Linux, UNIX as a second thing, one of its ilks, usually its Solaris but it can be others as well, and Microsoft has about a 20 percent market share there, not the 90 percent they have on the desktops. It's because of stability and reliability. There are no bug fixes that don't get made, in something like a free operating system, because you've got 10,000 guys out there, working on stamping on those bugs every minute of the day because they're all over the world. And this means that while you can have an economic piece of control of things like the desktop PC, you can't control those systems.

You go on, I've done way too much talking.

TPM: Comment on the Corel Microsoft thing.

Peter: It's a bitch. [laughter] I don't know what else to say.

TPM: Do you think this is the beginning of an what do you call it, extend, extinguish, I forgot, the three E's...

Peter: Yeah, I mean basically, I think that within this, the whole last century the basic thing that very big corporations do is to buy up and help out smaller opponents in order to step on them and grind them into the dust. And I think this is exactly the way Microsoft has been successful. It has bought up a very large number of other corporations, it has hardly ever incorporated the software into their own product, they just sat on it. It's not illegal to do that. The thing is that Corel has served as an interesting entity in the marketplace, and I don't think it will be able to do that in the future, and I think, that Microsoft will just squelch them.

TPM: Do you anticipate Microsoft trying to use Corel as a beachhead to attack the rest of the open source community, one at a time, a sort of a domino theory type thing?

Peter: I don't think so. I mean it's an interesting hypothesis, but I don't really think that that's the Microsoft way of doing it. If Microsoft wants to, they'll just go out gulp up other pieces. They don't need to apply what you refer to as the domino effect. The trouble is that a lot of those pieces are not gulpable. How does one gulp up the FSF and Richard Stallman. How does one get at SuSe for example, which is basically in Germany, and not a place where either where Microsoft has a tremendous beachhead, nor on the other hand a place where the economic end of the European Community is in fact going to tolerate that squelching that we do in the US. So I don't think it will happen that way.

TPM: I think you've answered my questions.

Peter: Great, that's what I wanted to do.

Editor's Interview Analysis

There's nothing like a historical perspective. Those who forget history are bound to repeat it.

As you can probably tell by my questions, I've been very agitated about this Corel thing. Peter has helped set my mind at ease. If there's one thing I can take from Peter's analysis, it's that "one monkey don't stop no show". Everything that Microsoft has tried has been tried before. In 1979 AT&T tried to jack their loyal following with a $7500 price tag. But that didn't end UNIX, or end the academic UNIX community, or the hacker UNIX community. In fact, the money grab begot Minix and GNU, which then begot GNU/Linux, and the community came back stronger than ever. The thing I take away from this interview is that, as a law of human nature, source code sharing and free software cannot be squelched. Remember Peter's words:

I think the really important thing to recognize is that the bright kids [...] are much more interested in creating something interesting and useful than they are in caring about who owns the license on X, Y and Z, and I think the kids are brighter than the suits who do the other things.

Let Microsoft have Corel. It's Corel's misfortune but not that of the free software community. One monkey don't stop no show. Even if Microsoft somehow bought Red Hat, Caldera, and Slackware, well, oops, there's still Debian, and its progeny, Progeny. Let Microsoft make Windows-specific Linuxes, attack with FUD, vaporware, and deliberate incompatibilities. Let them embrace, extend and extinguish. Yes, Microsoft's big and such actions can delay the free software conquest. But make no mistake -- we cannot be beaten and eventually we will triumph. As long as we keep the faith and support our community, we will escape any box the bad guys try to put us in. We will win. It's a historical imperative.

Steve Litt is the main author of Samba Unleashed. He can be reached at Steve Litt's email address.

Linux and Inline Skates

By Steve Litt
I wish I got a dime each time someone said Linux would "never make it" on the desktop, citing numerous reasons boiling down to "no apps, too complex". And of course, the underlying message is that "no apps" and "too complex" are somehow intrinsic properties of Linux.

Aren't they? Could it be possible for Linux to be capable of hosting apps, and capable of user friendliness, and yet not display such properties currently? How long does a superior technology take to prove it's superior?

I've seen it take 10 years...

In 1980 I joined a speedskating team called the Venice Speed Demons, from Venice, California. We competed in all local outdoor races in distances 10K, 20K, and 26.2 miles. For a team with no coach, no sponsorship, and no training facilities other than the Venice boardwalk and bike path, we did just fine. And in 1980, not a one of us had inline skates. As a matter of fact, we laughed at skaters with inline skates. Uniformly, "inline skaters" were either roller hockey wannabes without the endurance to go 10K, or recreational skaters who had learned on ice rinks. And uniformly, they were slow. We joked about inline skaters. And we made the obvious assumption -- inline skates were slow.

As the 80's continued, we started seeing a few fast people on inlines. Our reaction -- imagine how fast these guys would be on classic skates. But in about 1986, the Rollerblades company sent a fast guy to one of our training runs. He scampered away from our fastest team members. Watching him it was obvious the lack of trucks and truck flex enabled him to put more of his power on the pavement. Some of our guys began experimenting with inlines.

And failed. Inlines require a different skating style than classic skates, and our guys experienced a speed decrease on inlines. So they raced on classics, and once in awhile practiced with inlines. This pattern was repeating itself on every speed team in the country. And slowly, the classic guys learned inline skating style and began winning. By the turn of the decade all competitive outdoor speed skaters were on inline skates.

Technological development is on Internet time, but human habit is not. First the new technology must surround itself with a number of skilled people to demonstrate the technology. Additional time is required for new converts to change their habits in order to realize the advantages of the new technology. A technological conquest can require several cycles of conversion and education.

Some assume the fact that Linux has not won over the desktop in the last 2 years means it's not suited for the desktop. I'm not so sure.

In the late 80's and early 90's, former Venice Speed Demon Jonathan Seutter set several world records for most miles skated in 6 hours, 12 hours, and 24 hours. On inlines!

Steve Litt is a former outdoor speedskater turned geek. He can be reached at Steve Litt's email address.

Jacob Farmer Talks Backup

By Steve Litt
Jacob Farmer from  Cambridge Computer Services led Wednesday's 7pm Birds Of a Feather session titled "Backup and Restore Technologies for Linux". Thanks to Jacob, I won't be making a big mistake I might have made otherwise. Early in the BOF, Jacob asked the following multiple choice question:

Do you get better backup speed from a faster tape drive, or a slower tape drive:
A. Faster tape drive
B. Slower tape drive
C. It depends

But before discussing Jacob's answer, let me digress. Although it doesn't seem relevant right now, please allow me to talk about the process of transcribing an interview from tape. I promise you will soon see the relevance.

I taped my interviews of Peter Salus and Bradley Kuhn with a little Radio Shack tape recorder. Transcribing the tapes was brutal, taking approximately 50 minutes for Peter Salus and approximately 3 house for Bradley Kuhn. Here are the statistics:
Words Interview
Peter Salus 1112 8:47 126.6 40 50 22.24
Bradley Kuhn 1182 6:36 179.1 40 180 6.6

The interviewees could each talk in excess of 100 words per minute. I could type about 40 words per minute. So why is the final product (transcription speed) half that of the weak link (my typing speed) in one case, and one sixth of the weak link in the other. And why the huge variation in transcription speed? Answer these questions, and you'll understand Jacob Farmer's discussion of tape backup speed.

I should start by telling you that the interview with Bradley Kuhn, which took place in the exhibit hall, was noisier than that of Peter Salus. But both interviews were easily discernible, so I don't believe that was the major problem.

Anyone watching me transcribe would have seen the problem immediately. I'd listen to the tape, type as fast as I could, and then the tape would outrun me. So I'd need to reposition the tape (actually I repositioned a .au file made from the tape). Bradley talked faster, so he outran me after just a few words, whereas with Peter I could get a half a sentence or more before rewinding. So with Bradley's talk, I spent a more time rewinding. Much more time.

Bradley's increased spoken speed resulted in a decreased transcription speed. Remember that for later.

Bradley spoke at 179.1 wpm, and Peter spoke at 126.6. If I had been able to type 179.1 wpm or faster, I could have transcribed Bradley at 179.1 wpm and Peter at 126.6 wpm. If my typing speed fell below 179.1, I would have needed to rewind. At a typing speed somewhere below 179.1, the rewinding would have caused Bradley's transcription to become slower than Peter's. At typing speeds below 126.6, Peter's would have started to decrease also, but still not as badly as Bradley's. At incredibly slow typing speeds, I would have needed to rewind after every single word, but there is no speed at which transcription speed would have fallen to zero. Instead, transcription would have basically leveled out, as shown in the graph below:

Do I get faster transcription with a fast speaker, or a slower speaker? The answer is, "it depends". If my typing speed is comparable with the fast speaker, the fast speaker allows faster transcription. If my typing speed speed is significantly slower than the fast speaker, the slower speaker yields faster transcription. The reason is simple -- the tape stops for no man, so the more the tape outruns my typing, the more I need to perform a time consuming and unproductive tape reversal and cueing.

Now you probably see where where I'm going with this. Streaming tape drives stop for no man. If they outrun their data buffer, they need to rewind and re-cue, which slows backup drastically. Let's say you're backing up from a computer that can deliver a data stream at 5Mb per second. If you back up to a 3Mb per second streaming tape drive, you'll get a backup speed of 3Mb per second. If you back up to a much faster streaming tape drive (say 10Mb/s for example), the tape drive will probably rewind and re-cue on a constant basis, producing a backup speed much lower than the 3Mb/s being delivered by the computer.

In the streaming tape drive world, they even have a name for this constant rewinding and re-cueing. They call it "shoe shining". A pretty descriptive name. Basically, the tape buffer fills up, after which the tape moves to record the buffered data. When the buffer runs out, the tape tries to stop, but it can't be done instantly, especially since some streamers operate at speeds as high as 150 inches per second. So it must rewind, and then re-cue, and begin all over again.

In fact, the interview transcription graph above bears an amazing resemblance to the graph of computer data throughput, tape drive speed, and backup speed. That's because there are some very close equivalancies:

speaker wpm
<==> tape speed
typing wpm
<==> computer data delivery
transcription speed
<==> backup speed

Jacob Farmer detailed the result of not understanding these streaming tape drive facts. There are people out there who buy the biggest, fastest streamer they can get, and wonder why it performs poorly. Their fix -- they purchase several similar units and run them concurrently, splitting the already inadequate data stream and sliding each tape drive further left on the curve. Eventually the curve levels out, very similar to the fact that transcription levels out at the cue-per-word point, and they're able to back up with many, many high power streaming tape drives. But of course, they likely could have achieved the same result with one or two lower speed tape drives.

If your streaming tape drive seriously underperforms, look at the data stream coming into it. Back up from the highest throughput box you have. If your streamer's speed exceeds 1.25 Mb per second, don't back up over a 10mbps network. Similarly, if the streamer is faster than 12.5Mb per second, don't back up over a 100mbps network. Even so, most computers cannot deliver data that fast. In the end, you may need backup software that builds a huge data queue to minimize tape overrun.

EDITORS NOTE: Although Jacob didn't delve deeply into the subject of variation, it's obvious that data throughput is never constant. In fact, a few small gaps in the otherwise constant throughput can require significantly increased buffer capacity, or a significantly slower streaming tape drive.

Unlike my transcription, streaming tape drives don't fall off drastically at first. They have large buffers (the equivalent to my being able to "remember" a few sentences while still typing). And some tape drives can stop the tape pretty fast, minimizing the shoe shining. Nevertheless, a seriously undersupplied streaming tape drive will underperform considerably below the input data. At lower data flows, the fast streamer becomes the bottleneck.

Jacob discussed helical type drives, an alternative to streamers. The leading helical scan drives move the tape at a speed of about 1" per second.  They can start and stop on a dime and the tape drive buffering keeps it from needing much start and stop.  The Ecrix drive ( has even more fancy ways of coping with variations in data speed.  You can get further details at their web site.

A large part of Cambridge Computer Services' market distinction is education. By educating their customers and prospective customers, they can sell their customer the system that works best in the customer's environment, not the system with the largest spec numbers and the biggest price tag.

Steve Litt is the author of Troubleshooting Techniques of the Successful Technologist. He can be reached at Steve Litt's email address.

The Guy from VA

By Steve Litt
Fresh from the spectacular Google presentation, I went to the exhibit hall for what I thought would be a five minute visit. Just to the right of the entrance was the VA Linux area, containing about ten black-shirted VA people. But now there was a crowd around the area, and one of the black-shirted VA guys was wearing a microphone and giving a speech. I caught a few words of the speech, which went something like this:

"Don't think of management as idiots, it will show through. Instead, think of them as smart people who optimize for different things than you do".

Hey, that's an interesting point of view. I stopped to listen. The more the guy talked, the more interesting it got. I had to move right up next to the speaker system to hear the guy, because somebody else was playing an accordion loudly (go figure).

The guy with the VA shirt spoke of management's need to maximize profit and minimize risk. They don't need a "better operating system", or "Open Source". They need to maximize profit and minimize risk. Start by asking them if they really want to take the risk of placing their vendor in a monopoly position over them. Explain the downsides of working with software whose source is available only to the vendor. Then explain the risk alleviation of having the source yourself, either to have someone else fix problems or do it in house.

It went on and on, the information getting better and better. I blew off attending another talk to continue listening. The more this guy spoke, the more intriguing it got. Who the heck was this guy wearing a VA shirt with the microphone speaking to the people at the VA Linux booth?

The talk concluded and I walked over to Central Florida's favorite VA marketing rep, Jeff Rabin.

"Jeff, who was that guy who just spoke?", I asked.

Jeff smiled a surprised smile, like he thought I was pulling his leg. Finally he answered "That was Eric Raymond".

Steve Litt is the main author of Samba Unleashed. He can be reached at Steve Litt's email address.

Absolute Value Systems, Inc.

By Steve Litt
If you're read this far in the magazine you know that Central Florida's AbsoluteValue Systems, Inc., won the Best New Product Award, and that we Central Floridians celebrated appropriately. I want it clearly understood that I had asked AbsoluteValue for an interview BEFORE the award, and they agreed to the interview BEFORE the award. TPM interviews interesting products and interesting people, not just celebrities.

To save time, we conducted the interview by email. Here it is:

TPM: Congratulations on receiving the "Best New Product" award at ALS.

Mark: Thanks Steve, it was an unexpected honor.

Jo-Ellen: Thanks Steve.  I don't recall any awards being presented at ALS in the past, so I was completely surprised when awards were being distributed.  This was our first time having a booth of our very own, so I was already on cloud nine just from standing under a banner with our company name on it.  Receiving the award was an unexpected bonus.  Words can't describe how good it feels to be recognized for our efforts after having worked so hard for so many years in the Linux community.

TPM: How old is AbsoluteValue Systems? What led up to your company's formation?

Mark and Jo-Ellen: In 1993, we were extremely dissatisfied with our jobs and decided we wanted a change.  We founded Mathews Software Services with the intent of developing Windows based diagnostics software.  Our forays into the Windows software market were less than successful.  In 1994, we needed to replace a misbehaving NetWare file server and decided to try Linux and Samba.  We've been a Linux shop ever since.

In December of 1996 we received two pcmcia wireless LAN cards from a relative at Harris Semiconductor.  We needed a way for the wireless LAN to talk to our wired LAN and Linux routing seemed ideal.  All we needed was a Linux device driver for the wireless LAN cards.  Hence, the Linux WLAN Project was born.  The goal of the Linux WLAN project was and is to develop a complete, standards based, wireless LAN system using the GNU/Linux operating system including a Linux device driver for IEEE 802.11 wireless network cards, access point and portal software and user mode utility software. What differentiates this project from the Linux wireless extensions and other Linux wireless projects is that we base our development entirely on the IEEE 802.11 standard.

From 1996 until February of this year our company, renamed AbsoluteValue Software, focused on software development services.  Our specialty was applications of wireless LAN technology.  Our contribution of our wireless LAN device driver to the Linux community served as an excellent form of marketing.

In February of this year, we decided to switch the focus of the company from being a service company to being a product.  Our first product, the AVS Wireless LAN Developers Kit ( was released in June of this year.  The developers kit is a targeted embedded development environment that provides the foundation for companies and individuals who wish to develop wireless LAN products or add wireless LAN capabilities to embedded systems.  Our next line of products, due out soon, consists of the components necessary to create wireless LANs with and for Linux.  The product line is called which is short for Personal Intranet.  The first product in the line is a PCMCIA wireless LAN card that supports 11Mb/s operation and ships with driver support for Linux.  The second is a wireless LAN access point (or base station) that actually has Linux running inside it!  This product line is the source of winning the "Best New Product" award at Atlanta Linux Showcase 2000.

Some users will note that there have been Linux drivers for wireless LAN products for some time now.  Each of those drivers has been developed by a third party, not the vendor of the hardware.  In our case, we are the developers of the wireless LAN Linux driver for our products, we use Linux in our Access Point and we support Linux users and customers.

TPM: At what point did you decide to work with GPL software?

Mark: When we replaced that Netware server with a Linux box, I became intrigued with the notion of Free Software.  I did some research and read most of Richard Stallman's and Eric Raymond's writings at the time.  It became apparent to me that Free Software/Open Source Software (the OSS term hadn't yet been coined) was an unstoppable force.  It wouldn't take over the world overnight, it might take decades, but it will take over a significant part of the software industry.  I decided that we were best served by trying to figure out how to earn a living in the Free Software world sooner rather than later.  Jo-Ellen on the other hand was a little more skeptical...

Jo-Ellen:  When Mark wanted to install Linux on one of our computers to function as a server and firewall, I didn't care since it wouldn't be used as a desktop machine.  I was at a point of despair from both a user perspective and a business perspective.  I was sick and tired of my desktop machine crashing, and I was devastated by our lack of success in developing software for the Windows market.  Thoughts of updating my resume and searching for a job became more frequent.  Even changing careers haunted my thoughts because I was beginning to believe there wasn't any potential for earning money in developing software unless you were one of the well recognized software giants.

In casual conversation, Mark had talked to me about the underlying philosophy of the Free Software movement.  I vividly recall our last discussion about Free Software where I was skeptical about the philosophy.  My response to Mark was, "That all sounds good in a perfect world, but we don't live in a perfect world.  Even if you could make money on GPL'd software, I would have to believe the results would be even worse than shareware.  Look at how many people use shareware without paying for it, so how is there really any revenue potential on giving away software, especially in source code form?  What's to stop companies like Microsoft from stealing it?  No individual can afford to take on a company like Microsoft in court even if they could prove Microsoft was not acting in accordance with the GPL?"  Mark optimistically and enthusiastically addressed my concerns, but I maintained my role as Scrooge in the matter and just wouldn't buy it... "Bah Humbug!"

Mark wanted to attend the first Atlanta Linux Showcase.  I was excited at the opportunity "to get away", so I went with him.  I promised I would remain "open minded".  Around a dozen vendors were present, but no more than two dozen.  Seeing what they had to offer didn't take a long time, so the only thing left to do was attend talks, and talks were plenty.  I was particularly interested in the business track since I was so skeptical about how one can make money by giving away source code.  After listening to Eric Raymond's talk on his paper, "The Cathedral and the Bazaar," I was forever changed.  Mark was attending a talk in the technical track, and I couldn't wait to tell him what I had heard.  From that moment on, I was sold on developing Open Source Software.  I think Mark was shocked at my enthusiasm and rapid change of heart.

TPM: The word "business model" is overused, so let me ask you, what is your strategy for deriving income from a GPL'ed product?

Mark: We get asked this alot.  The easiest way to say it is to refer to Eric Raymond's paper "The Magic Cauldron"

Eric would have categorized our software development services business as a "Give away the recipe, Open a Restaurant" type business.  Our service business used the GPL'd (actually MPL'd) software as a marketing hook and clear demonstration of our expertise in our specific niche.  Basing our business on products is clearly an instance of "Widget Frosting".  We're selling wireless LAN hardware that is supported by Open Source software, and we're the only vendor in the wireless LAN market who is doing so.  We provide the most value for customers who need a complete and supported solution for Linux.

Jo-Ellen: Our Open Source project, linux-wlan, is a guerilla marketing tool.  Demonstrating our expertise through Open Source Software is like a client magnet.  Customers came to us wanting a port of our driver to their platform.  This strategy was the foundation of our business when we were a service oriented business.  However, customers didn't come to us so easily.  We had to "prove" to the big players in the wireless LAN industry that we were technically competent with regard to wireless technology and earn their respect.  Our Open Source project did just that.  Now that our focus is on developing products, we strongly feel we have the support of the Linux community.  We have many users of our wireless LAN Linux drivers, and we feel that users would like to have their software and hardware come from the same source and truly receive Linux support from the hardware vendor.

Editor's Interview Analysis

I've known Mark and Jo-Ellen since 1998. They were and remain one of the very few companies I know that truly succeeds with the orthodox "give away the product, sell the service" Open Source revenue model. As such AbsoluteValue Service, Inc. serves as an important case study. It's interesting that they thoroughly studied Eric Raymond, and then consciously set out to follow ESR's procedures and reproduce the result ESR describes. They studied, they planned, they executed, they succeeded. Could it be possible that the reason there are so many failures in the "Give away the recipe, Open a Restaurant" model is because would-be Opentreneurs don't have the faith to follow the formula?

It actually makes sense. Remember Jo-Ellen Mathews' words:

"Demonstrating our expertise through Open Source Software is like a client magnet."

This software was a killer portfolio piece opening normally skeptical doors. Back in the bad old days of proprietary software, you couldn't offer an app written for personal use as a portfolio piece -- it was considered "unprofessional". Remember, everyone was asking for three years of paid programming experience. How different it is now, when Open Source is considered highly professional.

One thing the Mark and Jo-Ellen modestly don't mention is that the software they wrote is very difficult and sophisticated, and also a very needed product. You can't make money by writing a cute little 3 day app. It must be substantial.

So they studied Eric Raymond, found a difficult problem needing a solution, solved it by authoring Open Source, and parlayed that into a "client magnet" to obtain software development contracts. In hindsight, it's absolutely obvious :-)

With everything going so well for them, you have to wonder why they started selling hardware. I didn't have time to ask them that question, so let me throw out a possibility that strikes me as plausible.

When you sell service, you're selling your time, and that's scarce. Considering that you often need to hop a plane to deliver that service, it becomes even scarcer. And it's especially scarce for Mark and Jo-Ellen, because their portfolio piece is for their skills, not for the skills of developers they might hire. To move beyond the income level of hourly rate multiplied by the 90 hours a week they each might be able to deliver, they had to sell hardware.

And whose hardware are you going to buy? You could go for Giganticorp's wireless hardware, and hope you can find a reliable driver. Or you can go turnkey with AbsoluteValue Systems. If you value your time, the choice is obvious.

As Mark mentioned, Eric Raymond has called this type of hardware sales "Widget Frosting". I don't think that name does justice to the melding of hardware with Open Source. I'd call the hardware more a main course than a frosting. And you know what? I bet Jon "maddog" Hall would call it the Value Added Model.

Yep. It's a perfect example of Jon "maddog" Hall's value added mode. Because there was no 802.11 wireless driver, AbsoluteValue wrote one. And then used that free factor of production to sell hardware.

One might object and say that by giving away their software, they are allowing their hardware competitors to offer all their advantages. Well, all advantages except technical trust. The competitors can adapt the AbsoluteValue's driver to work with their hardware, and you hope the hardware/driver combination will be highly reliable. AbsoluteValue Systems is the one with the brains and quality to have created the software. Which one are you going to trust with your data?

AbsoluteValue Systems, Inc. is successful with two different "open source revenue models". When all is said and done, what they've really done is parlayed brains, and ability to deliver superior product, into revenue. Mark and Jo-Ellen have proven what we all knew all along. In any "business model", the cream rises to the top.

Steve Litt is a director on the LEAP-CF Executive Committee. He can be reached at Steve Litt's email address.

Phil Barnett

By Steve Litt
After the Friday the 13th Bash discussions about a possible LEAP-CF Linux exposition, I knew I had to get an interview with Phil Barnett, president of Linux Enthusiasts and Professionals of Central Florida (LEAP-CF). The following is a post-show email interview done with Phil Barnett. The subject of the interview is the creation of a regional Linux show hosted in Orlando by LEAP-CF.

TPM: Phil, at ALS there was extensive discussion of a LEAP-CF sponsored Linux show in Orlando, Florida. What do you see as the strategy to start and grow the LEAP-CF Orlando Linux show?

Phil: Well, Steve, I think it's important that we don't start out trying to create Comdex for Linux in Orlando. As much as I'd like to do exactly that, I know that we need to start small and grow over time.  We need to create this as a regional event initially.

And, even though LEAP-CF has emerged as a leading advocate for Linux in Central Florida, we have other Central Florida resources at hand, and formed some excellent liaisons with other Florida LUG's at the Atlanta Linux Showcase. With that in mind, I feel this should be a multi LUG participation event with LEAP-CF carrying the flag due to our non-profit incorporated status.

There is mounting interest in Linux. We are now at the point where we can start bringing everyday computer users in contact with our outstanding operating system. We already know that we have the best server operating system around, we know that we are reaching into imbedded technology, we see Linux appliances on retail shelves and open source projects for Linux are scorching a path across application space. The desktop is at hand. We aren't there yet, but it's coming and it's coming fast.

With this in mind, I believe that this effort should showcase "Linux from A to Z". We have some of the brightest minds in Linux right here in Central Florida. Let's give our community a view of what Open Source and Free Software is all about. Solving problems without boundaries. From A to Z.

TPM: Works for me. I understand ALS started as an Installfest. Do you feel Central Florida Linux gurus have the drawing power to anchor a regional event?

Phil: Not knowing the answer to this question is exactly why we have to start small. I mean, the problem is not whether Linux is interesting or exciting nor whether we have LUG members that understand the depths of Linux. At the most basic level, we will face a problem attracting big names. Big names are important to an event like this. The public barely knows anything about Linux at this point, so we need one or more marquee names to get their attention. I believe this is the first thing we should try to accomplish. Get a few big names to commit. Without a few big names in the Linux industry to draw in the public, we might as well just hold a very large installfest/educationfest.

TPM: Even a regional show requires extensive planning and logistics. What would you say to skeptics arguing there are very few LUGs that could pull this off?

Phil: Frankly, I view creating a regional Linux show as requiring a lot of planning and time consuming hard work. And, virtually all of our LUG members are full time employees (and beyond). If we have a problem bringing this off, it will be because of a labor shortage, not a capability or talent shortage. In our favor is the fact that our board of directors has spent a lot of time learning how each other work and what our limitations are. I believe that starting small is the only way we can do this. We just don't have the resources to start big.

Editor's Interview Analysis

The LEAP-hosted regional event is obviously a well thought out plan, not a pie in the sky dream. Notice how Phil ticked off the challenges: Difficulty drawing big names, and limited volunteer time. Conspicuously absent is hotel reservations and the like -- this is a regional, not a national event.

We have a good start on the big names issue, because we have some, for want of a better phrase, second-tier names. In LEAP-CF alone we have the main author of Samba Unleashed, and Mark and Jo-Ellen Mathews of AbsoluteValue Systems, newly crowned "Best New Product" at ALS. I have a feeling that in less than a year we'll have considerably more second-tier names in our group. If we get a very few top tier names, that should be sufficient. Maybe not sufficient to pull from Chicago and Boston, but certainly enough to get people to drive 200 miles.

I liked Phil's statement:

Without a few big names in the Linux industry to draw in the public, we might as well just hold a very large installfest/educationfest.

A very large installfest/educationfest just might be an ideal "training run". It worked in Atlanta. Remember that the first ALS was an installfest. A large, successful and well publicized installfest just might draw some big names next time around.

Then there's what Phil calls "a labor shortage". This is always a problem until the show gets big enough that someone like Usenix participates. As Phil mentions, LEAP's board of directors (officially called our Executive Committee), has learned how to accomplish much with little. I'm sure the event we set up will be low maintenance. And we'll find a way to bring in volunteers.

For selfish reasons I like the idea of a local Linux show. I want a guaranteed annual show I can go to. But beyond that, I'd like a template that any LUG can follow to create a Linux exposition. Here's why. There seems to be a law of nature (let's call it Litt's Law :-), saying that if a Linux show gets big enough to tax the resources of the LUG running it, a larger partner from a different city will be brought into the mix, and the show will leave town. Traveling mega-shows are wonderful, as long as we can still count on local shows. The LEAP-CF template will assure plentiful regional shows.

So if you're in Florida or Georgia and want to help, email me and I'll put you in touch with the proper people.

Steve Litt is a director on the LEAP-CF Executive Committee. He can be reached at Steve Litt's email address.


By Steve Litt
I really need to apologize to ActiveState, because I gave them a very hard time. Especially David Asher. First I gave him a hard time about the changed Python license. Like he could do anything about Guido's ex-employer's actions :-). I gave both David and Mike Smith a hard time about ActiveState's work to implement Microsoft's .Net strategy into Perl and Python.

Here's the way my thinking went. I thought that if they supported .Net, that would decrease the chance that .Net would flop. If .Net flops, that's pretty much the end of Microsoft. Microsoft bet the farm on .Net and C#. I pretty much accused ActiveState of aiding and abetting the enemy (of course, ActiveState makes no bones about the fact that they're neutral in the OS wars).

About a week later I realized how INCREDIBLY hypocritical my stance had been. After all, if interoperating with Microsoft protocols is treason, what do you call my writing Samba Unleashed, a book that tells everyone how to interoperate with Microsoft's smb/cifs protocol? Open mouth, insert foot.

In fact, Samba has been a major factor in the migration from NT servers to Linux servers. Because of ActiveState, if .Net becomes a reality, you'll be able to use good programming languages on Linux boxes to interoperate with it, eliminating Microsoft's monopoly over .Net. As far as I can see, this is exactly equivalent to Samba.

ActiveState takes no part in the war between people like me and people like Bill Gates. They're neutral. Mike Smith emphasized that Microsoft is "a very good partner" of ActiveState. That's OK. Python and Perl support Corba too. And if, by chance, ActiveState were bought by Microsoft or coerced by Microsoft to "extend" Perl and Python in OS specific ways to lock out Linux, well, ActiveState's contributions to Perl and Python are Open Source...

I have a warm spot in my heart for ActiveState. The first Perl I ever used was ActiveState on Windows (ok, throw rotten tomatoes now). I loved it. I could do anything with it. It always did what you programmed it to do. It was quick to develop and reasonably quick to run. So I started using Perl on the HP9000. The rest is history.

And now ActiveState has Python that runs on Windows. Their Python 2.0 came out recently. Python is the most readable language on earth, and it develops quick as lightning. That's why companies like Google use it. Python is a work of art. It has hooks for Corba and XML, and thanks to ActiveState, if Microsoft's .Net initiative succeeds, you won't need to have VB or Windows to interoperate with it.

But ActiveState isn't just Perl and Python. They have a plugin called PerlMx, that's an SMTP level Perl wrapper for the Milter API. I couldn't understand everything ActiveState's Mike Smith said about it, but it sounds like you can write your own filter classes for very fine control of message filtering. After all, Perl has regular expressions -- a significant improvement over the filtering capabilities of most packages. You can build yourself a pretty sophisticated autoresponder with PerlMx. But here's the really kewl thing. It filters at the SMTP level, meaning that filtered out spam never makes it to the server. PerlMx comes bundled with six filters already written. These serve as both a starting point for the administrator, and as a template so the administrator can make his own filters. Behold the power of Perl.

They also have an interesting IDE called Komodo. Komodo includes language sensitive editor, on-line help, remoted debugging, code and data browsers. It does not, however, support drag and drop development a-la Delphi, VB and Powerbuilder. But it does support Perl, Python, and Javascript. Yes, Javascript, complete with syntax highlighting and method folding!

ActiveState is a non-combattant. They're not on anybody's side. And yet ActiveState Perl for Windows was a major factor in bringing me into the UNIX world. Interoperability favors quality.

So if you want 2 great cross-platform languages, a killer SMTP level filtering language, and an IDE that supports Python, Perl and XML, visit the ActiveState web page.

Steve Litt is a Central Florida Perl/CGI developer. He can be reached at Steve Litt's email address.

The Free Software Foundation

Given the fact that the GNU project set the stage for the triumph of GNU/Linux and all of its associated tools, I just had to interview the Free Software Foundation. And what better subjects to ask about than the troubling Corel/Microsoft thing, and the licensing questions associated with Python 1.6 and PHP4.

TPM: This is Bradley Kuhn of Free Software Foundation and Steve Litt interviewing him on 10/13/2000. Bradley, I'm a little concerned about some of the events that have been happening lately. One of the events I'm concerned about is the -- what I hear about the changed license for Python. Another one is -- that's kind of old -- is PHP4. Another the third thing, which may or may not impact, or which you may or may not have something to say about, is the alliance between Microsoft and Corel.

Bradley: OK well let's talk about first with the PHP license because that's the one I remember first. What happened with PHP4 is they decided to switch from a dual license, which was a disjunctive license with GNU general public license and then a license they had written, to using only the license they had written. The license they wrote was free software license, so PHP4 is free software, but unfortunately it is not GPL compatible. What that means is you can't take PHP4, combine it with GPL software, and make a larger program out of it legally. You'll be in violation of one of the licenses. So it is very difficult to derive software from PHP4 given that most free software, according to Freshmeat over 50%, is GPL, and you can't combine GPL programs with PHP4, which is a real tragedy. So what we're encouraging people to do is just stick with PHP3, which is under the disjunctive license,  continue to code under this disjunctive license which is GPL and the PHP license, disjunctively, and in turn that person so and use that version. We attempted to convince the PHP developers not to change this, but they wanted to go their own way, which is their right to do as the copyright holders of the software, but we're encouraging people to stick with PHP3.

TPM: How bout Python?

Bradley: The python issue. Yeah [inaudible]. You know Guido -- you might want to speak to him -- he's very in favor of having a GPL compatible license for Python. Unfortunately he's not the copyright holder of the new version of Python. So he's been trying to convince the people who hold copyright on the work, which I guess is his employer -- you can check that out -- I don't know I don't have my computer in front of me to have the details, but he's trying to convince them and we're helping out as much as we can. We've been in discussions with them to try and help out so hopefully that situation will be resolved. Again, most of the people in the situation are very interested in solving it . It's a very friendly discussion.

TPM: Any opinion on Corel and Microsoft?

Bradley: You know Corel has consistently distributed mostly proprietary software and so has Microsoft, so as far as we're concerned it's just two proprietary software companies getting together to do more proprietary software, which is just as bad as all the proprietary software there is. I mean there's no greater harm done, I mean the harm has been proprietary software and that's continuing, and that's -- we see that as unfortunate, but we see that as unfortunate with any company, that's unfortunate for anybody.

TPM: I've been under the impression lately that some of these licenses are getting a little sleazy and I feel more comfortable with GPL licensing. Do you have any comment -- does that sound reasonable to you?

Bradley: Actually we're happy with any free software license. First and foremost. When something is free software that is a wonderful thing no matter what license they use.

TPM: Free software is what, copyleft?

Bradley: No, free software's not copyleft. Free software just means you're getting all the basic freedoms. And they're listed on our website I can bring them to you right now. You have the freedom to help yourself by examining the source code, modifying it yourself and being able to use it for any purpose. You have the freedom to help others by taking your modifications of the software and distributing them as you like to to other people and not being required to distribute, but it's your option to distribute to others if you like, and of course you have the freedom to copy and share non modified versions as well. These freedoms are listed on our website we have a page that lists them very specifically. Any license that provides those freedoms, and we also have a list of those licenses that do on our website, we think that's wonderful. If they're providing all those freedoms, that's the important thing.

Now there is  one other issue we're concerned about, which is whether a license is GPL compatible or not . As I spoke about with the PHP issue. So if the license is GPL incompatible but a free software license it still grants all those freedoms but it creates a very big practical problem when you want to combine software together. So that's the concern we have with that. But the important thing is it's free software and that's what matters most. Now we'd like to talk to these people who are using GPL incompatible licenses, and try to convince them to please, switch to a disjunctive license. For example, Mozilla just did that. They're going to be switching as much as they can to a disjunctive license, GPL or MPL. So that's a case where they were headed GPL incompatible free software license and now they're switching to a disjunctive license which makes the overall license GPL compatible. So we think that's a wonderful thing that they've done and we're hoping that others will do that.

TPM: A disjunctive license is two different licenses for the same software?

Bradley: Basically it's, I often describe it as a meta-license. The license of Mozilla when they change it will then be the license will state that you have the permission to distribute the software either under the terms of the GNU general public license or the terms of the Mozilla public license. So that each user has the choice of which license to accept. They can accept the GPL, or the MPL, or decide to continue distributing under both licenses as a choice. So basically [it] allows you to choose which license you want to accept, or to accept both as a disjunctive.

TPM: Wouldn't that, depending on the other license, wouldn't that open a route to taking it proprietary?

Bradley: Well not in the case of the Mozilla public license cause the Mozilla public license doesn't really permit, if I recall correctly, I need to check the facts but I believe it doesn't permit proprietary forks. Now the disjunctive license that Perl uses, which is GPL or Artistic License, the Artistic License does permit proprietary versions of a software, so in that case it would permit it -- it depends what 2 licenses you're talking about, there's no general rule.

Editor's Note: Bradley asked me to check with him about whether the MPL permits proprietary forks. Via email he has just let me know that the MPL does make some controls on proprietary software forks, but is not a strong copyleft.
TPM: OK, that's it.

Bradley: OK

TPM: Thank you very much.

Bradley: You're welcome.

Editors Analysis

Once again, you can see by my questions that I was quite worried about Corel, and by the PHP4 and Python licensing issues, which I had interpreted as threats to free software. Bradley said that the PHP4 license IS free software. I didn't know that. He recommends starting with PHP3 because it's disjunctively GPL compatible. Bradley also reiterated what David Asher of ActiveState had told all of us at a BOF session, that everyone was working in a friendly manner toward a resolution to the Python license question. As far as Corel, Bradley said the predictable (for an FSF person), which is that proprietary software is equally bad, regardless of whether it comes from Microsoft, Corel, or anyone else.

Bradley mentioned that any free software is a good thing. But I have a concern. Will the free software I contribute to today stay free? I discuss this issue in the "GPL is Goose Friendly" section of the "The Goose that Laid the Golden Egg" article later in this magazine.

Steve Litt is the main author of Samba Unleashed. He can be reached at Steve Litt's email address.

A Thousand Stories at ALS

By Steve Litt
There were a thousand stories at ALS. I've managed to record a handful.

Jeff Rabin, the Human Face of VA Linux

A few days after ALS I had a real forehead slapper thought -- I had neglected to talk with anyone at Red Hat. Whoops!

How could I forget to interview the commercial distro that fights vigorously for free software. The distro I've said on numerous occasions is the best server distro? Heck, these guys almost put GNU/Linux on the map. Why didn't I talk to them?

It was intimidation. They're so big. They had such a large exhibit area. They had so many people, and those people all looked so professional. And those people were always surrounded. Every time I went by Red Hat, I figured I'd talk with them later, when they looked less busy. And the opportunity slipped by.

That didn't happen with VA Linux. VA's exhibit area seemed as big, as populated, as busy. But my buddy Jeff Rabin worked there.

Jeff Rabin is a VA Linux marketing rep in the Southeast U.S. He's a fixture at LUG's from Miami to Memphis.  Every Linux or Open Source show I've been to, Jeff Rabin's been there. I've seen Jeff at meetings of three different Florida LUGs. Several times. VA Linux is a big company now, but to me, Jeff is the human face of the company. And he's a really nice guy.

VA is smart to have guys like Jeff. Guys who make it their business to know their areas' LUGsters personally. When people ask me where to buy a preconfigured Linux box I tell em VA. I know I'm doing right by recommending VA. Because Jeff Rabin has taken the time to explain to me all the internals of their boxes and what a customer can expect from VA Linux. I can recommend VA with confidence.

The bigger and more important a company gets, the less approachable it seems. It's important to put a human face on your company.

XFree86 Project

Did you see what XFree86 version 4 can do? At their booth, they had sideways terminals, videos in videos, double monitor games running lightning fast, and tons of other stuff. The only think I didn't see was game combatants walking out into the isle. Version 4 is incredible, and they've managed to do it with support for legacy ISA cards, at least in non-accelerated mode (and after all, how "accelerated" can five year old cards be anyway).

Borland JBuilder -- all standard Sun Java

Borland was there, demonstrating JBuilder on Linux. It seemed like a much better product than the version 1 I got for Windows. And according to the guy who demonstrated it, the output is 100% standard Sun Java, meaning, I take it, that you don't need to put a bunch of slow loading Borland-specific classes on your server.

Judging from the size of the print on the logo above their booth, Borland is now emphasizing their Borland name, and de-emphasizing "Inprise". That's a good thing. I started using and loving Borland products in 1984, on a Kaypro 2x CPM machine. For awhile in the late 90's, around the time they began calling themselves Inprise, it looked like they strayed and forgot who their customers were. But they appear to be coming back. I look forward to hammering out Java code on my Linux box with JBuilder.

JBuilder has all the goodies, Javabeans, JDBC, all the rest of the stuff you need for quick enterprise Java apps. They have different editions of JBuilder, each with its own price and capability. Delphi and JBuilder on Linux gives us a lot more clout.


Strolling by the Sangoma booth, I talked to Evelyn Ledsham, VP of sales. She showed me a card containing a csu/dsu and router, ready to plug into your system. I couldn't understand everything she said, but this card appears to be an excellent solution for a branch office. Just plug one of these cards into the branch office's server, and it facilitates communications via the Internet with the head office's router. Their WANpipe software supports all IP encapsulated data and all tunneling implementations, as well as offering sophisticated diagnostics. If I understand this correctly, we're talking instant branch office scalability using the Internet. Like I said, I don't completely understand it, so you need to confirm all of this at their website, which is listed in the URL's section of this magazine.

Old Friends

Hanging out at the Debian booth, this guy says "hey, how you doing?". For an instant he looked vaguely familiar -- then I knew -- it was John Goerzen but with shorter hair. John and I had met at Linux Expo in Raleigh, May 1999, after  we worked together on the Linux Unleased 4E team. John's still a member of the Debian team, still writing lots of books, and he's working for Progeny Linux Systems.

I glanced at the BRU booth and who should I see but Marty Connor, the guy who gave me a ping pong lesson at Linux Expo in Raleigh in May 1999. We got together, swapped a few stories. Marty's contributing to several Open Source projects, including the Etherboot Project ( He is also involved in ( Especially interesting is a document at in which Marty is mentioned 10 times in connection to things like ntulip.c, reducing RTL8139 and VIA-Rhine driver footprints, and adding Netgear FA310TX, 3C905C and LinkSys LNE100TX v4 support.

Walking the halls I spotted Art Wildman of the Jacksonville LUG. I'd met Art at ITSA day in Gainsville a few weeks before. I tried to catch him, but the place was so crowded he faded into oblivion. A few minutes later I found him again. Art is an enthusiastic JAX LUG member, one of the key people in forming a solid Florida Linux presence.

Should I call Anthony Awtrey a new friend or an old one? We have a virtual friendship going back more than 2 years, but we just met in person at the Friday the 13th Bash. Tony's running the technical end of Awtrey Consulting and still active in MLUG, Inc. (Melbourne LUG). We tentatively scheduled my Samba presentation at MLUG (I'm not sure he knows I'm the main author of Samba Unleashed :-). Tony, Phil Barnett and I began planning a multi-lug event.

New Friends

I was puttering around the halls when I met Joseph Carter. Joseph works for Progeny Linux Systems and as a member of the Debian Project. He told me Progeny will soon be issuing a Debian-based distro with an easy installer and configuration tool. IMHO that's precisely what Debian has lacked until now. Joseph wants to design a voice communicating handheld computer using embedded Linux. He mentioned the benefit to the visually handicapped. If you've been reading Troubleshooting Professional Magazine for more than a year, you know I'd give my eye teeth for a box allowing me to walk through the woods dictating and editing Troubleshooting Professional Magazine. Go Getem, Joseph!

So I'm getting off the skateboard game at the Dave and Busters party when who should get on but this guy with 2 foot dreadlocks. His name was Smith (not sure if it was his first or last name, too loud), but he and I both spent most of our time on the skateboard and snowboard games. We never got a chance to discuss what we do, but he says he'll be at a lot more Linux shows. Smith -- be sure to come to LEAP-CF's show -- who knows, maybe as early as next year.

Old Friends Missed

ALS was missing some of my Linux Expo favorites. Sadly absent were Digital Creations (the Zope crew), who I heard through the grapevine are now doing well and in the black.
Late Breaking News!

Guido and the Python team have just moved to Digital Creations!

Also not there were the fun and hard working crew from Andover.Net. If Victor Goodman and Marc Spencer were there, I didn't see them. Yall come to the LEAP-CF show when it happens in Orlando.

Steve Litt is a director on the LEAP-CF Executive Committee. He can be reached at Steve Litt's email address.

Who Gets the Money?

By Steve Litt
Jon "maddog" Hall's Value Added Model basically states that you don't sell Linux, you use it as a vehicle to sell other things. That model can be extrapolated to all Open Source. But there's a very basic question. Who gets the money, and if not the original developer, then what's the developer's incentive.

Here are some of the Value Added Model examples discussed in this magazine:
Offering OS Project developers Who got the money
Google search engine Linux crew Google
Smoothwall appliance The Smoothwall crew Any networking consultant
AbsoluteValue Systems WLAN card AbsoluteValue Systems AbsoluteValue Systems

You'll notice in 2/3 of the cases, they guys doing the Open Source coding weren't the guys pocketing the money. Houston, we've got a problem!

What kind of a sucker would code his fingers to the bone so someone else can reap the rewards?

I now call for a moment of silence so everyone can consider the impact of the "who gets the money" factor on Jon "maddog" Hall's Value Added Model, and whether it renders that model useless.

Minute's up. To answer this question, let's go to that most classic of Open Source documents, ESR's "The Cathedral and the Bazaar.

In a Cathedral development methodology, the coding load would be shouldered by a very few developers. For all but the most ideological, writing this much code without reasonable expectation of direct reward would represent an unacceptable alternative cost. They could have bought a car, or a good portion of a house, by contracting out those hours to a client.

But now look at the Bazaar development methodology. There's an "inner circle" of a few developers doing heavy coding, design, and planning. Supporting them are hundreds (in a large project) of folks contributing code to work with specific hardware, or specific OS's, or to do specific tasks or incorporate specific features. This is some serious leverage given to the "inner circle".

There is a non-monetary reward for the supporting people writing small, encapsulated additions or changes. They get the benefit of having their favorite features incorporated, or hardware supported, in their favorite Open Source software. Without the necessity of re-patching each new version. And secure in the knowledge that further development won't break their favorite feature. It's probably something they would have done anyway, but now they can make sure it's integral to the software. Furthermore, these supporting players have an excellent bullet point for their resume, and also are in a unique position to add some value and sell it. So the supporting players do just fine under the Value Added Model, and have every incentive to keep doing what they're doing.

But what about the inner circle, who work day in and day out planning, designing, strategizing, massaging egos, and yes, coding. They work their lives away making a product to which they don't have time to add value. You know, guys like Larry, Guido, Andrew and Jeremy, Linus. They must subsist on their speaker fees, and the high profile jobs offered them due to their reputations, and in some cases IPO money. The inner circle of large projects have leveraged their skills with the contributions of large numbers of small contributors. It's tough, but they scrape by :-).

Somewhere in the middle are moderate contributors to large projects, and the large contributors to moderate and small projects. They may not get the big speaker fees, the very best jobs, or IPO money. But their Open Source activities aren't as time consuming, so they have time to incorporate added value to the Open Source they create, and they also gain enhanced clout marketing their services.

So basically, the more Bazaar-like the development methodology, the more incentive for Value Added Model. Here are the incentives:
Developer Classification Incentive
Inner circle of large projects The best jobs and contracts, speaking fees, book writing opportunities. Ability to leverage the no-cost work of others to create software far beyond their abilities and time limitations.
Middle layer of large projects,
or inner circle of moderate to small projects
"Client magnet" for sales of services, time and energy left over to sell services and value added products.
Rank and file contributors Their favorite software now contains their pet features. They obtain an excellent resume bullet point, and an inside track in selling services or value added product associated with the Open Source software.

This is a Win Win Win Win Win proposition. Among developers, the inner circle, the middle layer, and the rank and file contributors all win. The person adding value also wins. And the end user wins. The Value Added Model is very much logically consistent.

Makes for an interesting career path. Start as a rank and file Open Source coder moonlighting as a consultant. Through extra dedication and quality work rise to the middle layer, moonlighting by selling services, selling value added products, and writing books. Rise to the inner circle, or switch projects to get into an inner circle, and moonlight as a speaker, board of directors member, and IPO stock recipient.

Apollo, you're clear for landing.

Steve Litt is the author of Rapid Learning: Secret Weapon of the Successful Technologist". He can be reached at Steve Litt's email address.

The Goose that Laid the Golden Egg

By Steve Litt
These are wonderful times for Linux. Read my interviews with Jon "maddog" Hall, Peter Salus, and Bradley Kuhn and it's obvious these are great times. See IBM and Sun solidly in our corner. GNU/Linux is making inroads, jobs are being created. These are the golden days of Linux.

But what about the goose that laid this golden egg? What about the rank and file Open Source developers? Are we cultivating them to keep laying golden eggs? Or are they headed for the stew pot? And if they get boiled, can Linux survive?

Turn the clock back to mid 1999, just after LinuxExpo in Raleigh. O'Reilly hosted their "Opensource Software Convention" in picturesque Monterey, California. The cost -- as I remember about $1500 for registration in 4 tutorials and access to the conference. And hotel prices in the vicinity of picturesque Monterey, California were astronomical. Nothing like the $45/nite I got in Raleigh for a clean, quiet room, or the $59.00/nite I got a half mile from ALS at a Holiday Inn. O'Reilly's "Opensource Software Convention" was too expensive for the rank and file Open Source developer to attend. What's up with that?

For years the Southeast has had two Linux shows each year, Linux Expo (1995-1999 in Raleigh, NC) and Atlanta Linux Showcase (1996-2000 in Atlanta). Next year there will be none (although there are vague rumors of a return of Linux Expo). Atlanta Linux Showcase, now renamed Annual Linux Showcase, will be in Oakland CA next year, although the plan is to move ALS around the United States from year to year.

Today it seems like if you want to attend a Linux show, often as not you need to go to places unattractive to a geek on a budget. California's expensive. Lodging in Chicago and Las Vegas are expensive, especially when shows such as Linux Business Expo are attached to hotel filling shows such as Comdex or Networld+Interop. The average guy writing a few subroutines or patches for Open Source software won't be able to hang with his own crowd. How long will it be until the image in his mirror murmurs "sucker", and he quits coding for the projects whose shows are in the high rent districts.

Pricing hintergeeks out of the shows is the first step in alienating them. Peter Salus mentioned that geeks would always create interesting and useful software. I called it a historical imperative. But there's no guarantee that such software will be centered around GNU/Linux, Samba, Perl, Python and the like. Especially if the rank and file geek is disenfranchised.

The solution isn't as simple as having more large shows, but I believe there is a solution. It will be discussed later in this article.

Then there are licensing issues. In the past year there have been instances of what some see as license backsliding. Licenses becoming less free, or less GPL compatible. I think geeks contribute to free software with two expectations:

  1. Their enhancements will always be freely and legally available to them and their associates
  2. No carpetbagger will charge exorbitant prices for their free work
Violate these expectations, and contributions will plunge. Nobody wants the image in their mirror to murmur the word "sucker".

This is where the GNU GPL license comes in. This is discussed later in this article.

Linux is the golden egg. The world's geeks are the goose. Let's not blow it.

ALS Is a Good Template

If you're looking for a template for a show that honors rank and file geeks, look no farther than this year's ALS. It was located in a very nice area amenable to those on an expense account, but there was plenty of reasonable lodging (I got a Holiday Inn room walking distance from the show for $59/nite). An exhibit-only pass was free, and it enabled you to get into those wonderful Birds of a Feather sessions. The tech sessions were reasonable at $260 for all 3 days, $50 less for Usenix members (and you could sign up for Usenix on site). At $760 for 2 days, the tutorials weren't cheap, but they were reasonable for corporate training purposes. And of course, tutorials and training, never are cheap, unless you use Rapid Learning techniques.

The bottom line is that a couple geeks could come to ALS and mingle for 4 days, with new friends and old, for a lodging expense of maybe $120 each, and whatever food and gasoline cost them. ALS honored rank and file geeks.

The LEAP-CF Initiative

Linux Enthusiasts and Professionals of Central Florida are discussing the creation of a regional show in Orlando, FL. We talked extensively to some of ALS's pioneer organizers. We're organized, we're incorporated, we have contacts in the government, in schools, in industry. Our show may not draw like ALS or Linux Expo, but with any luck Florida and Georgia will always be assured of at least one show per year. By the way, if you're anywhere near Orlando Florida on November 16, 2000, Python guru Mark Alexander will be presenting "Zope Web Application Server!". Come early for adequate seating.

I'd like to see LUGs in every state do regional shows. That way, everyone would always have 2 or 3 shows to go to every year. Sure, some of the biggest players wouldn't be there, but we just might meet some local heavyweights instead. A perfect example is that the winner of the ALS "Best New Product" award was none other than Central Florida's own Absolute Value Systems, Inc.

Regional Shows and the Farm System

The real problem is that the Geek-centric regional shows have grown up and moved away. That's natural. A successful show outgrows the administrative abilities of its sponsoring LUG and requires the assistance of an organization having show expertise. And that new organization will of course want to move the show around for maximum exposure. This is all natural.

The trouble is that the shows that grew up and moved away have not been replaced by other regional shows. Regional shows were a wonderful meeting place for the region's local geeks, and also a great opportunity for the region's companies to become exhibitors.

Besides their services to local geeks and local Linux businesses, regional shows also conserve Linux community resources. Even the biggest potential exhibitors and sponsoring companies had budgets, and must pick and choose their national and international shows. The same is true of Linux heroes like Linus, RMS, ERS, Andrew and Jeremy,  maddog, Larry, and Guido. Every large show depletes these resources. A regional show needs only one or two "big name" companies or individuals to draw attendees from a 300 mile radius.

Not only do regional shows conserve these scarce resources, but they can also grow them, acting like major league farm systems. Regional shows enable local celebrities to develop into big drawing worldwide celebrities. Regional shows must be grown in enough quantity and distribution to guarantee every rank and file geek, and every aspiring small Linux business, the chance to hang with their crowd.

Growing a regional show has its own challenges. Local exhibitors are less able to pay big bucks, so paying for the venue becomes somewhat problematic. Lower cost venues can help. It may be possible for big names to advertise, through sponsorship or equipment loans, without the need to actually appear. Souvineers can be sold instead of given away. I don't have the answers.

But I intend to find out. I could not locate a Linux Show HowTo. If it turns out such a howto doesn't exist, I'll write Linux conference documentation based on our experiences planning and executing the LEAP-CF. And other resources, including analyzing the few shows that didn't go to well, and finding out the cause of their failures. Hopefully, the Linux Show Howto will provide a recipe for LUGs across the country to put on regional shows, thereby serving their rank and file geeks, and serving as a farm system to train major league Linux market leaders and Linux advocacy leaders. Once there's a reasonable body of information, I will attempt to make it into an LDP HowTo.

My Challenge to Macmillan

I've done quite a bit of writing for Macmillan Publishing, including authoring Samba Unleashed, and I can tell you Macmillan is very Linux friendly. They're located in Indianapolis, Indiana. I don't remember whether it was Joseph Carter or John Goerzen who mentioned that Indianapolis's LUG meetings are held at Macmillan.

Macmillan -- I challenge you to host a Linux conference. You can do it in the summer when the South is too hot. Such a conference would cover states that don't get too many Linux conferences.. Indiana, South Wisconsin, East Iowa and Missouri, Kentucky, Tennessee, Western Virginia, West Virginia, Western Pennsylvania, Ohio, Lower Michigan, and a small part of Canada are within a comfortable 500 mile 1 day drive, as well as Illinois, which recently played host to Linux Business Expo. Within the radius of a 1000 mile 2 day drive (or a 2 hour flight) are the entire east coast except Maine and Florida, all the gulf states, Oklahoma, parts of Texas including Dallas, parts of Minnesota including Minneapolis, and the midwestern states as far west as Kansas and Nebraska. Indianapolis has a very favorable cost of living, presumably with plenty of economical lodging. But it's a big enough town to have fine restaurants and hotels. I've been there -- it's a nice town.

IMHO O'Reilly shouldn't be the only book publisher with a Linux Conference.

Raleigh and Atlanta, Come on Back!

ALS, don't be a stranger. Atlanta would be great! I hope future East Coast ALS's aren't in super-expensive places like New York. I've heard rumors that Linux Expo is coming back to Raleigh. Do it. Please!

And to LUGs across the land. It's our job to build regional conferences to replace those that have grown up and moved on.

GPL is Goose Friendly

As I mentioned, geeks contribute to free software with two expectations:
  1. Their enhancements will always be freely and legally available to them and their associates
  2. No carpetbagger will charge exorbitant prices for their free work
Fail to fulfill these expectations, and they'll go elsewhere. This is why license is important. We want to make sure that our current hard work isn't kidnapped as somebody else's property a year from now. The GNU-GPL and GNU-LGPL licenses are ideally suited to that requirement. Let's start by examining the Python situation. Python is a language originally created in the early 90's Guido van Rossum while at Stichting Mathematisch Centrum (CWI). In 1995 he began work for CNRI in Virginia, and he and his team moved to BeOpen in, I believe, May of 2000. Now, as of 10/27/2000, they have moved to Digital Creations, home of Zope. Just before the move to BeOpen, CNRI created their own license for Python. There were GPL compatibility problems, which CNRI and FSF are trying to work out. I got the following license info off pythonlabs:
These materials were copied from:
Python 2.0 License Information


Python was created in the early 1990s by Guido van Rossum at Stichting Mathematisch Centrum (CWI) in the Netherlands as a successor of a language called ABC. Guido is Python's principal author, although it includes many contributions from others. The last version released from CWI was Python 1.2. In 1995, Guido continued his work on Python at the Corporation for National Research Initiatives (CNRI) in Reston, Virginia where he released several versions of the software. Python 1.6 was the last of the versions released by CNRI. In 2000, Guido and the Python core developement team moved to to form the BeOpen PythonLabs team ( Python 2.0 is the first release from PythonLabs.  Thanks to the many outside volunteers who have worked under Guido's direction to make this release possible. 


BeOpen Python Open Source License Agreement Version 1

1. This LICENSE AGREEMENT is between ("BeOpen"), having an office at 160 Saratoga Avenue, Santa Clara, CA 95051, and the Individual or Organization ("Licensee") accessing and otherwise using this software in source or binary form and its associated documentation ("the Software"). 

2. Subject to the terms and conditions of this BeOpen Python License Agreement, BeOpen hereby grants Licensee a non-exclusive, royalty-free, world-wide license to reproduce, analyze, test, perform and/or display publicly, prepare derivative works, distribute, and otherwise use the Software alone or in any derivative version, provided, however, that the BeOpen Python License is retained in the Software, alone or in any derivative version prepared by Licensee. 



5. This License Agreement will automatically terminate upon a material breach of its terms and conditions. 

6. This License Agreement shall be governed by and interpreted in all respects by the law of the State of California, excluding conflict of law provisions. Nothing in this License Agreement shall be deemed to create any relationship of agency, partnership, or joint venture between BeOpen and Licensee. This License Agreement does not grant permission to use BeOpen trademarks or trade names in a trademark sense to endorse or promote products or services of Licensee, or any third party. As an exception, the "BeOpen Python" logos available at may be used according to the permissions granted on that web page. 

7. By copying, installing or otherwise using the software, Licensee agrees to be bound by the terms and conditions of this License Agreement. 


Python 1.6 is made available subject to the terms and conditions in CNRI's License Agreement. This Agreement together with Python 1.6 may be located on the Internet using the following unique, persistent identifier (known as a handle): 1895.22/1012. This Agreement may also be obtained from a proxy server on the Internet using the following URL: 


Copyright (c) 1991 - 1995, Stichting Mathematisch Centrum Amsterdam, The Netherlands. All rights reserved.

Permission to use, copy, modify, and distribute this software and its documentation for any purpose and without fee is hereby granted, provided that the above copyright notice appear in all copies and that both that copyright notice and this permission notice appear in supporting documentation, and that the name of Stichting Mathematisch Centrum or CWI not be used in advertising or publicity pertaining to distribution of the software without specific, written prior permission. 



So this license is to be interpreted under the laws of Virginia, I mean California, I mean Virginia, I mean California. Digital Creations is in Virginia, so at least there won't be another state added to this mess :-).

In the past, I've talked to Zope CEO Paul Everitt. He and the entire company have a true love for Python. So I'm hoping that Digital Creations will be able to use their good offices to finally clear this up once and for all, and I certainly hope the final Python product will be GPL compatible.

In the meantime, we can all learn from Guido's trials and tribulations.

None of the preceding applies to Python specifically, because my research shows that Guido never was the copyright holder for Python. Nevertheless, it gives us a valuable lesson.

Now a word about LGPL...

Let's say you want to make a library useful within commercial programs. I'm not a lawyer, but it appears to me that the Lesser GNU Public License (LGPL) allows that within parameters. The library itself appears to be GPL in every respect except that software linked with it is not required to be GPL. The library itself must have full source passed on, and you must grant the same rights to the next guy that you had. But they can link it with their proprietary code, as long as they provide a mechanism such that the recipient of that proprietary software can modify the library and recompile, which basically means they must supply all object files for the entire work. That's my understanding.

It's also my understanding that the LGPL has provisions to prevent its being used in a totally proprietary fashion with a wink and a nod. For instance, if the library refers back to tables in the main program, the library calls must work correctly if the tables are absent.

I kind of wish I had done UMENU this way. I could have made almost the entire thing an API library licensed LGPL, but Joe's Proprietary Medical Software could have incorporated UMENU functionality in their medical management package. As for my UMENU executable, my calling program would have been GPL, or BSD if there were a viral problem between GPL and LGPL.

Editor's Note: The preceding two paragraphs represent my understanding of the LGPL after a 15 minute reading. I could be mistaken. I am not responsible for any problems resulting from my interpretation of the LGPL license. Use my information at your own risk, or don't use it at all. I am not a lawyer.

Whether you want to have a totally and forever free license (GPL), or a license that makes some provision for incorporation into proprietary software (LGPL), I recommend looking first at the licenses from FSF.

The GPL benefits software users because it is the best guarantee that the software they use -- the software they've taken their time to learn, deploy, implement, and use as a cornerstone of their business, will always be available at a reasonable price. There is little reason to fear that one day Microsoft will "buy Linux".

The GPL benefits the software developer by guaranteeing that his hard work in contributing to the software will always be available in a reasonable way. If worst comes to worst and the software maintainer will not place the developer's mods into the distribution or main code tree, the developer can always distribute his own version. Without the need to rename the product, or distribute it via patches.

The Geek is the Goose

It's tempting to credit GNU/Linux's rise to Linus, Andrew and Jeremy, RMS, Eric, Larry and Guido, Red Hat, IBM, and Sun. That's pretty much what the press does. But such a view is inaccurate, and it's dangerous for the future of Linux and all current Open Source. In fact, thousands upon thousands of geeks have built the body of what we commonly call Open Source software. Next time you go to a LUG meeting look around you, and it's a good bet you'll see 5 or 10 Open Source contributors.

For example, Linux Enthusiasts and Professionals of Central Florida is a moderately sized LUG in Orlando, Florida, a city best known for its tourist industry, not technology. But even here we have many Open Source contributors:

These are only the LEAP-CF contributors I know of, but I'm sure there are more. And probably even more have contributed a patch here and there. Many LUGs have more, and many contributors aren't in LUGs. Open Source depends on thousands of contributors.

Why in the world does a person code without pay? What is the incentive?

  1. Solve a problem, then enlist others to better the solution
  2. Ability to procure the best software
  3. Professional networking
  4. Recognition
Licensing backslides undermine #1 and #2. Pricing rank and file geeks out of shows undermines #3 and #4. Nobody wants the image in the mirror to call them a sucker. For the sake of all the software we've worked for, our community must continue to serve its geeks.
Steve Litt is a the author of Troubleshooting Techniques of the Successful Technologist, Rapid Learning: Secret Weapon of the Successful Technologist, main author of Samba Unleashed, and a contributing author to "Linux Unleashed" 4E, "Red Hat Linux 6 Unleashed", and "Red Hat Linux 7 Unleashed". He can be reached at Steve Litt's email address.

Linux Log: The Heart and Soul of ALS

Linux Log is now a regular column in Troubleshooting Professional Magazine, authored by Steve Litt. Each month we'll explore a facet of Linux as it relates to that month's theme.
Celebrities were everywhere. It was great talking to people you'd read about in the magazines. But that was the fringe benefit of ALS.

Beneath the conspicuous celebrity surface, swirling eddies of geeks met, formed groups, planned the future, broke up and reformed. Needed information was procured and strategic alliances were made.

That's why they have parties at Linux shows. Tech sessions, BOF's and the Exhibit floor are a great place to meet the celebrities, but a heck of a lot of the business gets done at parties. And if it hadn't been for Phil Hughes I would have missed my chance.

Friday afternoon LEAP's Phil Barnett introduced me to Phil Hughes (Linux Journal publisher), who promptly asked me if I was going to the Friday the 13th bash that night. I declined, citing notes and interviews to transcribe. But later that day Phil asked again, and this time I said yes. Thank goodness!

When the Authors and Publishers BOF got out at 9pm, I went to the Friday the 13th Bash, flanked on one side by an author whose name I've forgotten (sorry, I didn't transcribe my notes soon enough) and the other side by Coriolis Acquisitions Editor Kevin Weeks. Inside, pounding music made it hard to talk, but we managed. Soon several LEAP-CF members joined us. When Don Marti announced LEAP's own Mark Mathews and Jo-Ellen Mathews' AbsoluteValue Systems as makers of the best new product, the Central Floridians in the crowd went happily berserk. Two of our own had achieved celebrity status.

So five minutes later I'm congratulating Mark and Jo-Ellen's at their table, when a voice says "you must be Steve Litt!". This is spoken by a guy I've never seen before. Then he says "I'm Tony Awtrey".

Tony and I go back a long time. Via email. Two years ago, we discussed his excellent content on how to CGI on a web host who wouldn't give you telnet or ssh. Later, when I became a member of LEAP-CF's Executive Committee, and found out Tony was a major influence in the Melbourne LUG (MLUG),  he and I had corresponded about growing and incorporating a LUG (both LEAP-CF and MLUG are now incorporated). Later I wrote him about going to ITSA Day in Gainsville. And now we meet in the flesh.

At the Friday the 13th Bash, Tony and I first agreed that I'll speak on Troubleshooting Samba at MLUG. Then we discuss organizing a knock your socks off all-Central-Florida (maybe even all-Florida) event. I pulled LEAP-CF President Phil Barnett into the discussion, details were hammered out, plans were made. It's premature to talk about the event, but I can tell you it will be long discussed throughout the Linux community.

ALS. We arrived as several people from several LUGs. We left as a solidified regional presence. This pattern repeated endlessly with various groups at ALS.

Celebrities were everywhere. Some were long standing community celebrities. And some celebrities were last week's rank and file geeks, newly encouraged and empowered through ALS to greater contribution. This, my friends, is the heart and soul of ALS.

Steve Litt is a member of Linux Enthusiasts and Professionals of Central Florida (LEAP-CF). He can be reached at Steve Litt's email address.

Letters to the Editor

All letters become the property of the publisher (Steve Litt), and may be edited for clarity or brevity. We especially welcome additions, clarifications, corrections or flames from vendors whose products have been reviewed in this magazine. We reserve the right to not publish letters we deem in bad taste (bad language, obscenity, hate, lewd, violence, etc.).
Submit letters to the editor to Steve Litt's email address, and be sure the subject reads "Letter to the Editor". We regret that we cannot return your letter, so please make a copy of it for future reference.

How to Submit an Article

We anticipate two to five articles per issue, with issues coming out monthly. We look for articles that pertain to the Troubleshooting Process, or articles on tools, equipment or systems with a Troubleshooting slant. This can be done as an essay, with humor, with a case study, or some other literary device. A Troubleshooting poem would be nice. Submissions may mention a specific product, but must be useful without the purchase of that product. Content must greatly overpower advertising. Submissions should be between 250 and 2000 words long.

By submitting content, you give Troubleshooters.Com the non-exclusive, perpetual right to publish it on Troubleshooters.Com or any A3B3 website. Other than that, you retain the copyright and sole right to sell or give it away elsewhere. Troubleshooters.Com will acknowledge you as the author and, if you request, will display your copyright notice and/or a "reprinted by permission of author" notice. Obviously, you must be the copyright holder and must be legally able to grant us this perpetual right. We do not currently pay for articles.

Troubleshooters.Com reserves the right to edit any submission for clarity or brevity. Any published article will include a two sentence description of the author, a hypertext link to his or her email, and a phone number if desired. Upon request, we will include a hypertext link, at the end of the magazine issue, to the author's website, providing that website meets the Troubleshooters.Com criteria for links and that the author's website first links to Troubleshooters.Com. Authors: please understand we can't place hyperlinks inside articles. If we did, only the first article would be read, and we can't place every article first.

Submissions should be emailed to Steve Litt's email address, with subject line Article Submission. The first paragraph of your message should read as follows (unless other arrangements are previously made in writing):

I (your name), am submitting this article for possible publication in Troubleshooters.Com. I understand that by submitting this article I am giving the publisher, Steve Litt, perpetual license to publish this article on Troubleshooters.Com or any other A3B3 website. Other than the preceding sentence, I understand that I retain the copyright and full, complete and exclusive right to sell or give away this article. I acknowledge that Steve Litt reserves the right to edit my submission for clarity or brevity. I certify that I wrote this submission and no part of it is owned by, written by or copyrighted by others.
After that paragraph, write the title, text of the article, and a two sentence description of the author.

URLs Mentioned in this Issue