LinuxWorld Expo Report, Part 2
~ By Marty Connor
What a great show!
As I fly back from San Francisco, I am filled with a sense of both satisfaction and optimism. Our demonstrations went well. We made good contacts, and network booting still holds a fascination both for us, the implementors, and for those who use what we create.
In this part of my LinuxWorld Expo report, I'll share with you how we plan, prepare for, and execute our show experience. Hopefully, you will find it interesting to see how things look look from the inside. This is Open Source, after all, and transparency of mechanism is a central theme in what we do.
After 11 Expos in the .ORG Pavilion, some things are easier. We have a pretty good idea of what to bring and generally what to expect during our time at the show. However much planning we do, there are always surprises, usually pleasant ones, that keep things interesting.
So, let's get to it. There are somewhere around a dozen sub-tasks that we do when preparing for a show. They are all intended to make sure that when the lights come up and the people come by we can demonstrate our technology reliably.
Here is a rough list of the steps:
* Travel there
* Booth Setup
* Meeting up with people
* Packing up
* Shipping back
* Traveling back
Let's talk about planning for a moment. Each .ORG Pavilion exhibitor is given a 10x10 booth, with two 6' tables (one tall, one short), two chairs, a rug, power, and internet access. This may not sound like much, but if we had to pay for just this, it would come to over US $6,000.
This is one reason we put so much energy into the show. Having done (and paid for) booths at MacWorld for a different project, we are keenly aware of what a generous grant this is.
We still have to get there, and between lodging, air travel, and food, we still spend a few thousand dollars, but free booth space is a tremendous help.
Now that we have the booth, the question is, what can we show there? People will be walking around the show floor, stopping at places that catch their eye, and interacting with people in the booths. So we want to make our booth interesting.
Since our software is all about network booting, we want to include everything we need to demonstrate that and to give our visitors a good sense of what network booting allows one to do.
So we'll need several computers, a router, lots of network cables, a ROM burner, and various smaller pieces of equipment.
To make the booth attractive and eye-catching, we'll need a banner, and we'll need flyers to hand out. Fortunately we have friends that do graphic design, and clients that are printers, so we can barter and call in some favors for things we need help with.
We also want to arrange some meetings with other people in the area who are "skilled in the art" of network booting. This involves emails and calls, and we were fortunate this year to be able to arrange dinners, lunches, and other social time with a number of people, even before arriving in San Francisco.
This is very important because getting to do in-person, full-duplex communication with these folks is rare (and a lot of fun). Also, we're here to talk about extending network booting in significant ways, and the thoughts and counsel of these particular people are invaluable. Network booting (especially on PC hardware) is quite a narrow area of speciality, and meeting with people who are skilled in this area is a rare opportunity, one we won't be wasting.
So we have our hardware, our plane tickets, and our hotel reservations. Michael Brown, our lead developer, and Anselm Hoffmeister, longtime contributor to the Etherboot Project, have agreed to work in our booth.
Since we're in active development, it is important to build and test everything before sending it to the show. We want to make sure everything works well.
Show attendees have limited time and attention spans, so our machines must boot reliably and allow the person demonstrating to be able to get a good cadence for the demo. This requires predictability. We have an acrylic-clad demo machine which is eye-catching. Equally important, it allows people to see the inside of the machine that is booting over the network and to verify that there are no obvious bootable media devices (hard drive, floppy, CD) installed in the machine ("nothing up my sleeve...").
The other important thing is to have people in the booth that can answer technical questions as well as be personable enough (and patient enough) to answer questions from people who don't have much idea of what they are seeing. They are our guests after all, and we're honored to be able to show them what we have. Part of what's neat about being at the show is to see the light go on in people's eyes, as they realize what it means if you can start your computer up over the network.
Having people who can be natural and unscripted is good, because you never know what people might ask about. Sometimes their questions will suggest really clever ideas, which sometimes we have implemented at the show.
To make sure everything works ahead of time, we mock up the whole booth in our living room and create a small private network to simulate being at the show. We practice the demos we think we are going to do, while preparing to bring enough extra things to give us flexibility to do things that are unscripted (for example, flashing an EEPROM for someone to take home that night and try network booting themselves).
Once we're comfortable that our demo server (an IBM ThinkPad running RHEL in this case) is doing what we want, and that our demo equipment is running appropriately, we are ready to pack and ship.
Packing and Shipping
Getting 150 pounds of equipment safely from Boston to San Francisco is fairly expensive. With shipping being driven by fuel prices, it pays to ship early and, even then, rates are high. We shipped to our hotel, the San Francisco Marriott, because it has an excellent shipping department, and we can ship far enough in advance to get the lowest Fedex rates. Even with discounts we're looking at a several hundred dollar shipping bill to get our equipment there and back.
I'm not a big fan of air travel, and heightened security concerns meant that extra time was needed to fly. It wasn't so bad for me, but for Michael Brown, our lead developer, traveling from England meant that he would not be able to use his laptop during the flight. London to San Francisco is a long flight, and a great opportunity to write code. In-flight "entertainment" passes time, but is certainly no substitute. Oh well.
I arrived Sunday evening (13 August 2006), and established a LAN in my room, because I knew there would be hacking to do before the show on Tuesday morning. Little did I know how much.
Michael arrived later Sunday evening, after a bit of airline redirection. His hotel was not too far from mine, and we agreed to meet on Monday.
IDG did a nice job with the .ORG Pavilion. Making the Slashdot / SourceForge.net Lounge the focal point, with our 20 booths ringing it was a great setup. It brought people over and gave them a place to relax.
Not everybody does the kind of exhibiting we do in our booth. As you can see from the list of people in the .ORG Pavilion, many groups don't really have the need or desire to demonstrate what their group produces. Some groups produce licenses or specifications, and meeting and talking to people about things is their main goal. Other projects are showing their latest distributions or releases. What was really cool was to be able to put names to faces. Emails to authors. Code to coders.
The .ORG Pavilion exhibitor list this show included:
* Creative Commons
* Eclipse Foundation
* Electronic Frontier Foundation
* Etherboot Project
* Fedora Project
* Free Software Foundation
* Free Standards Group
* Gentoo Linux
* GNOME Foundation
* K Desktop Environment (KDE)
* Linux Terminal Server Project (LTSP.org)
* The Linux Test Project
* Open Source Software Institute
* PostgreSQL Project
* The FreeBSD Project
* X.Org Foundation
Some names you may recognize, others you may not. All are dedicated to promoting the use of Free and/or Open Source software. There is a lot of camaraderie in our area of the show.
As is my custom, I went over on Monday to check out the show floor and to get our exhibitor badges. The show staff was efficient and professional, and I was pleased to see that the booths were up and that our friends in the LTSP booth were already setting up.
Normally, I would have put up our 3x8' banner in our booth on Monday afternoon, but I was a bit pressed for time. We had a special dinner Monday night, and Michael had driven to Sunnyvale to get a 64-bit rackmount computer for a particular demo we wanted to do.
As you will see later, hardware can be a seductive paramour.
Monday Dinner with Markus and hpa
Michael and I had the pleasure of having dinner with Markus Gutschke and H. Peter Anvin (also known as "hpa") on Monday night. Food was definitely not the main point of the meeting, though dinner was nice.
The opportunity to meet up with people like Markus and hpa is another one of the reasons we believe it's valuable to take time to travel to shows like LinuxWorld Expo.
Markus and hpa are longtime network booting experts. Markus was the original author of Etherboot in the early 1990s and hpa (among a great many other things) created PXELINUX, which allows network cards that follow the PXE specification to load Linux and various other programs and kernels over a network. They are both world class.
Michael is the person who added code to support the PXE network booting specification to Etherboot. Getting Etherboot to the cutting edge of network booting has been a project of mine for a number of years. I personally funded the addition of PXE to Etherboot and now wish to see it extended to support a number of features that it currently does not.
So we had dinner the night before the show, from about 5:30pm to 7:00pm. Markus and hpa had never met, Michael had never met Markus or hpa, and I had never met hpa. Yet, it is amazing how quickly we felt comfortable talking. We had read each others' code, communicated via email on technical issues, and had some idea of each others' opinions on various topics.
The conversation flowed naturally and comfortably. Even through points of disagreement, we found ourselves genuinely pleased in the knowledge that we were among a precious few who knew precisely what we were disagreeing about.
One thing we want to do is to arrange for PXELINUX to be able to use the HTTP support we have added to gPXE, our next version of Etherboot. The PXE spec does not include support for this protocol. It uses TFTP, which is UDP based. Basically, we want people to be able to load their operating systems from web servers. It sounds simple enough, but getting PXELINUX and gPXE to talk to each other will require a number of low-level changes that only the implementers can sort out.
We were like a quartet of jazz musicians. Each person playing their instrument, adding notes to the fabric of the musical garment we were creating in the moment. Dreamers, we, but also makers who can make our dreams come true. Moments beyond price.
The Expected Unexpected
I saw it coming. Michael, trapped on a plane all those hours wanted to hack. There was a demo we really hoped to do, but we the code wasn't ready. I had bought a card table and chairs from the hardware store across the street, and we had set up way too much hardware in my hotel room and started working early Monday afternoon. The .ORG Pavilion opened at 10am on Tuesday morning.
I had expected to put the tested equipment back in the crate for transport to the show after dinner with Markus and hpa and get a good night's sleep, but that special demo we wanted to do wasn't working.
So at 7:30pm the night before the show opened, we started hacking again. I was already tired from traveling, setting up equipment, re-testing, and a big dinner, but hacking will not be denied.
We worked until almost 3:00 in the morning, coding, debugging, single-stepping, and getting farther than either of us thought possible but not quite far enough.
You know how it is. A friend needs your help and, though you know there is a price to be paid, you are there for them, as they have been for you. I was fried, but I would stay up as long as Michael needed, helping him debug, because that's what friends do.
Morning, and the start of LinuxWorld Expo, would come all too soon, with the promise of a full day of meeting people and demonstrating our work. And yet, the time spent hacking in-person with Michael was a rare delight. His code is beautiful in both form and effect, and I find working with him to be highly educational and inspiring.
End of Part 2 ...