In an earlier blog posting, I mentioned that I was on my way to western China, to give a keynote talk at GCC2007 in Urumchi, which is in northwest China. I’ve been many places in central and eastern China (Beijing, Xian, Guangzhou, Guilin), but I had never been to western China or Tibet.
The mountains and culture of Tibet and the historical significance of Urumchi on the Silk Road were powerful attractors for an inveterate academic tourist, particularly when I could gain some new insights into Chinese science policy and scientific computing priorities from some of my international colleagues from the Chinese Academy of Sciences.
My wife Andrea and I arrived in Lhasa on a beautiful, sunny day and checked into the hotel that afternoon. The oxygen cylinders in every hotel room were a hint of what was to come, but we laughed about it as an unnecessary affectation. I should have known better …
Digital Death Comes Calling
Needing a digital fix and wondering about network connectivity in Tibet, I turned on my IBM ThinkPad. Windows Vista booted normally, and my applications began loading. Life was good. Then, I saw the dreaded blue screen of death, followed by a message that struck terror in my heart:
Disk read error Ctrl-Alt-Del to retry
I typed Ctrl-Alt-Del with trepidation, hoping Vista would boot, but the error repeated. Frantically, I enabled the BIOS advanced diagnostics and launched a disk read/write test. It failed immediately. Knowing what that meant, I reached up to pull out my hair, only to realize that even this venting of frustration had been thwarted. (See Dan; see male pattern baldness.) Arrgggghhhhh, nooooooo!!!
The evidence was clear and ominous; I’d had an old fashioned disk head crash of the worst kind. The real question was how much data had I lost. Little, I hoped, as I thought about IBM’s Rey Johnson, the inventor and patron saint of rotating disks, and the probability that only Vista’s swap space was corrupted. My upcoming keynote talk and several proposal drafts were on this disk.
All that faded from my mind, replaced by a single, haunting scientific fact – the Tibetan altitude, all 3650 meters of it. I had foolishly tried to cheat Messrs Navier and Stokes, and they were not amused. They had reached out and struck me down for hubris, or at least stupidity.
You see, disk read/write heads literally fly on a cushion of air, using the same principles as other airfoils. Unfortunately, commodity disks are rated for operation only to 3000 meters. Above that altitude, the air is not dense enough to provide the requisite lift for flying disk heads without other design compensations. (It’s the same reason airplanes need longer runways for takeoff on hot days and at higher elevations.)
... Then There’s Biology
Do you remember those hotel oxygen cylinders I mocked earlier? The first night in Tibet, I awoke around 3 AM with a massive headache, one of those “Oh, please, bludgeon me into unconsciousness so the pain goes away” migraines from altitude sickness. I was having a second head crash, the biological kind this time.
Thinking I might be suffering from caffeine withdrawal as well, I made and slammed down three big cups of instant coffee – no help, though it kept me awake the rest of the night to experience the pain. (If you were wondering, there aren't any Starbucks in Lhasa.)
In a better living through chemistry moment, I then rummaged through my travel kit and downed all the combinations of acetomenephin (Tylenol), naproxen (Aleve) and ibuprofen (Advil) I could find – no help. Cold towels, temple massage and other measures followed. Finally, realizing what was happening and seeing the edema in my face, I began sitting upright, and the pain eased around sunrise. I really did then enjoy the culture of Tibet after this. We had a great time visiting the palaces, monasteries and the area.
The Navier-Stokes Equations
I’ve been reflecting on the irony that my disk crash and altitude sickness were due to the same physics that dominates much of my professional life: the Navier-Stokes equations. Beguilingly simple to derive, yet fiendishly complex to evaluate, these differential equations are an application of Newton's second law to describe fluid flows in a wide range of physical situations: weather (air), ocean currents (water), airfoils (air), galactic motion (stars), explosions (fluids) and with Maxwell's equations, fusion (plasmas). They also describe biological phenomena of great theoretical and practical importance, notably circulation (blood) and respiratory dynamics (air and particulates).
For all but the simplest cases of laminar flow, the computational fluid dynamics (CFD) solutions are both computationally and numerically challenging. Modeling turbulence in all its forms (e.g., Rayleigh-Taylor, Richtmyer-Meshkov and Kelvin-Helmholtz instabilities) demands complex software structures and high-performance computing systems of great power. From the large-scale structure of the universe (stars, galaxies and the galaxy clusters) through the design of fusion tokamaks such as ITER to numeral weather prediction and disaster response (e.g., in our LEAD and HydroMet projects) to Pringles potato chip manufacturing and inhalant-based drug delivery, CFD continues to be a major enabler of engineering and scientific discovery.
Now Back to Our Story
Let's recap. I’m in Tibet, I have a dead laptop, I’m giving an invited keynote talk in 48 hours to a group whose native language is not English, and my only remaining digital lifeline is a Blackberry.™ Ironically, my GPRS service in Lhasa is better than in Chapel Hill, but the status of broadband networking in the United States is a topic for another day. In a word, I'm hosed, big time.
What I really need is disk forensics, but that must await my return. When I reached Urumchi, I spoke to Professor Zhiwei Xu, the conference chair, and explained my problem. He did what all professors are inclined to do. He found the nearest graduate student conference volunteer, explained the problem and said, “Loan Professor Reed your laptop and help him get on the network.” (Her name was Like; thank you, where ever you are!)
Four hours later, I had recreated my talk using Chinese PowerPoint and realized that being facile with Microsoft Office was language independent. That is an HCI story as well. Life was good.
Despite the disk crash and altitude sickness, I had a fantastic time in China. From the technical and policy discussions with my colleagues to the scenery and culture, it was a great experience. I will write more later about Chinese economic development and science policy.
I’m also very appreciative of the hospitality and friendship I experienced. The great thing about being an academic is that one has friends and colleagues all over the world. It is a fraternity without boundaries, united by intellectual curiosity and a shared perspective on scientific questions.