Comments from Frankston, Reed, and Friends
Tuesday, November 29, 2005
DPR at 9:58 AM [url]:
Some musings regarding the <proposal by the CDC to require tracking the net of passengers sharing flights for contact tracing purposes>_._
One of the sources of difficulty with proposals like the CDC's here is that there is a knee-jerk reflex in systems that contain networks (here the airline transport network) to assign functional responsiblity to the network for functions that are fundamentally "end-to-end" goals.
Though the end-to-end argument is well-understood in the Internet community (though still the subject of amazingly hotheaded controversy, as I can attest based on the flame mail I and others get), it is rarely applied outside the context of the Internet.
Tasking the airline systems with solving epidemic health issues has all kinds of side-effects (not just privacy, though its advocates will be the first to react, I'm sure).
The fundamental problem is that while contact tracing is a valuable tool for epidemic tracking and control, nevertheless the airlines are just a small part of the system, and the problem cannot be defined as a function of the airline net alone.
An end-to-end style analysis of the best way to manage epidemic transmission would not start with the airline network. Instead, it would start with the assumption that the functions placed in the airline network should be viewed as optimizations only of the epidemic detection and control process. As an optimization only, the impact of the airline's role on its function should be both small and as generally useful to ALL of the functions of the airline network as possible. And if that implementation does not integrate with or enhance current "edge-based" epidemic transmission control mechanisms, the function should definitely NOT be included. (I.e. contact tracing only works for some epidemic processes and some epidemic control techniques, which we are largely unable to execute because of lack of capability in our current PHS and WHO).
This proposal is clearly subject to "mission creep" as it is way overbroad in the data it collects. All that should be required is for the airlines to be able to quickly reconstruct a list of contact traces, *after the fact*, and only in the case where an epidemic is in its early stages (late stage epidemics do not even USE contact tracing) and only in the case where the epidemic is actually transmissible by air in the context of an airline trip (very few epidemics have this property).
Maintaining a vast database of all travelers all the time is fairly inexpensive in $, but has a huge cost in the potential for misuse or expanded functionality, of course, but also in reducing the flexibility for new and better solutions to the underlying public health problems (new systems create new bureaucracies that have a stake in keeping themselves funded).
Contact tracing that involves where a passenger comes from (as opposed to who shared the airplane with him/her) is already quite straightforward, and has less broad implications.
Unfortunately, and this is frustrating to me, there are many, many people in government (and this is not partisan), who argue from social benefits (managing epidemics) to "obvious" solutions that scale vary badly. The "end-to-end" style of systems thinking doesn't deny that solutions exist, but like the early designers of the Internet, by focusing on simple, scalable solutions that push solutions away from the core of the network as much as possible, it is possible to address problems in a way that has lower impact and is far more scalable and efficient.
It seems likely that we could be far safer and healthier if we thought more clearly about how to engineer the networked systems we live in, recognizing that network scaling offers alternative ways to approach problems. Instead we seem doomed to think in terms of heavy-handed solutions, which create opposition by their very heavy-handedness, and in the heat of the debate we rarely ask whether we are even solving the real problem we set out to solve in the first place - i.e. reducing pandemic susceptibility and impact.
Sunday, November 27, 2005
DanB at 11:53 AM [url]:
Truthful TV ads?
Watching the Sunday morning interview shows with their corporate ads, I was struck by the alternate ways you could hear their tag lines in light of other discussions about those companies.
Every week there seems to be another example of Microsoft being involved in some effort to stop the adoption of Open Document Format in Massachusetts. They appear so passionate in their zeal to find a way, any way, to prevent the change or discourage others from making a similar change. I also read press reports about the great potential ODF has for advancing things through interoperability, etc. What's Microsoft's tag line? "Your Potential. Our Passion." I see their visuals showing a tiny start up with the potential Microsoft sees to grow into something big. I couldn't help but think Netscape. Hmm. Microsoft wants you to grow, just don't do it in their backyard. Maybe "our tools help you grow" is more of what they meant to say. Strange thoughts -- I guess too much time on the elipitical exercising in front of the TV got to me.
Then there was Verizon. There has been talk about how they are moving to an all-fiber optic network so that they can have pathways into the home that they don't have to share like copper due to some regulatory fine print. Pipes that they can exploit any way they want because they control them. Not the open Internet but rather Verizon controlled paths. You'd think they'd want to counter that with a slogan like "We make you part of the world" or something else open sounding. No such luck. They go with "Our People. Our Network." Sounded to me like "we built it, it's ours, we will do with it what we want." Sigh. Must drink more water while exercising...