What Are Rights?

Ed Brayton fails to make a very important distinction in his 9th Amendment article. This is best highlighted by the following very short excerpt.

The thesis here is that the “unenumerated rights” referred to in the 9th amendment refers to the entire mass of individual natural rights. Those natural rights, argued Madison, Jefferson and many others, were pre-existing (that is, they exist prior to the formation of governments, which are instituted for the purpose of protecting those rights) and inviolable (that is, it is always unjust to violate those rights, regardless of what form of government violates them).

So what distinction is missing? The one between rights and desires. The 9th Amendment says that the federal government may not “deny or disparage” unenumerated rights, but that doesn’t mean everything that’s unenumerated is a right. What Brayton presents is a common false dichotomy: anything not explicitly delegated to the federal government is automatically allowed. Well, no. The natural rights to which he refers are not infinite. The works that defined the term “natural rights” provided specific reasons why specific rights should be considered pre-existing and inviolable, but did not include everything one might wish to do. The “right” to expose yourself to children, for example, is not a natural right. It’s not automatically sacred because the constitution doesn’t mention it. An argument must still be made, and not just assumed, that something is a natural right before it can be protected.

Brayton says that “Where in the constitution does it say you have a right to do that?” is the wrong question, and he’s correct, but after that he goes astray. The problem with that question is that it limits itself to the constitution, but the constitution was never the source of rights and that’s what the 9th amendment recognizes (even according to Brayton’s preferred interpretation). Instead, we should be asking, “Is that a right or just a desire?”

There Oughtta Be a Law

…stating that a “rebate” cannot require the purchase of further goods or services, in particular those that involve recurring charges that must be stopped by calling some hard-to-find phone number and navigating through an absurd menu system to get to some overworked guy who has been specifically trained to be as unhelpful as possible when you try to cancel the useless no-value “membership” you had purchased to avail yourself of the so-called rebate. Come to think of it, maybe there already is such a law. Isn’t “bait and switch” already illegal? Yeah, “mach speed technologies” – I’m talking about you.

I wonder how many of these scams – bad enough as they are all by themselves – are also just a way to make someone pay for having their own identity stolen.

The Age of Scarcity

Michael Klare has written an excellent article called Consumption Has Finally Caught Up With Us, about the way that unsustainable consumption is beginning to catch up with us.

For most of the 20th century, global stockpiles of vital materials like oil, natural gas, coal, and basic minerals expanded as giant multinational corporations (MNCs) poured billions of dollars into exploring every corner of the Earth in the drive to locate and exploit valuable deposits of extractible materials. This permitted consumers around the world to increase their consumption of virtually everything, safe in the knowledge that even more of these commodities would be available next year and the year after that, and so on infinitely into the future.

But this condition no longer prevails. Many of the world’s most promising sources of supply have been located and exploited, and all of the additional billions spent by MNCs on exploration and discovery are producing increasingly meager results. Ever since the 1960s, the most fruitful decade in the worldwide discovery of new oilfields, there has been a steady decline in the identification of new deposits, according to a recent study by the U.S. Army Corps of Engineers. Even more worrisome, the rate of oil field discovery fell below the rate of global petroleum consumption in the 1980s, and since then has fallen to approximately half the rate of consumption. This means we are increasingly relying on deposits found in previous decades to slake our insatiable thirst for petroleum — a pattern that cannot continue for much longer before we will begin to experience an irreversible and traumatic decline in the global supply of oil.

The same is true of other vital resources, including natural gas, uranium, copper, and many minerals. There may be adequate stocks of these materials on global markets today, but the MNCs are not finding enough new deposits of these commodities to replace what we’re consuming. So future shortages are inevitable.

I think this idea – that resource-depleting behavior is bad even when resources seem abundant – is one that a lot of programmers tend to grasp intuitively, and a lot of non-programmers tend not to. Early in their career, almost every programmer has to deal with resource depletion, whether the resource is CPU cycles or memory or communications bandwidth. What programmer doesn’t know what a memory leak is, or why it’s bad even when you have lots of memory? No real programmer, in my opinion. One example that particularly comes to mind is the memory cache inside a modern disk array (or the Revivio appliance). This cache is like a flood-control reservoir, necessary because I/O traffic tends to be bursty, and increasing the end-to-end performance of such systems tends to be expensive. The cache allows a slower device to act like a much faster one, for the duration of one or several I/O bursts, by letting the incoming writes accumulate in the “reservoir” during the burst(s) and then letting them drain later. But what happens when the I/O coming in the front end is continually more than the device can retire out the back end? Think of the reservoir again. The reservoir fills, and overflows, and all hell ensues. Early versions of the Revivio appliance actually had the problem that if this cache ever did fill, performance would actually end up being worse than if we’d never tried to use the cache in the first place. Far worse, even to the point of total collapse. To avoid this “running off a cliff” problem, it is necessary to exert “back pressure” – start slowing down responses, or even allowing requests to time out, well before the saturation point is reached.

In the real world, this “back pressure” comes in the form of rising prices. As demand outstrips supply, prices go up until effective demand falls back to supply levels. Unfortunately, in our actual economy, resource extraction – e.g. mining, forestry, oil and gas extraction – tends to be heavily subsidized, and this has a ripple effect throughout the economy. Because materials are so cheap, manufactured items are designed around replacement rather than repair. They truly don’t make ‘em like they used to. Have you ever seen a celphone-repair shop? Of course not. You don’t repair celphones; you just go out and get a new one. But what happens to the old one? The many exotic materials that went into it – sometimes tons of ore for a few ounces of metal – end up in a landfill, where they’re a toxic liability instead of an asset. It’s just not economically worthwhile to extract and reprocess those materials, compared to the subsidized extraction of more. This is clearly not sustainable – sooner or later those resources will run out – but the equation works out the way it does because of the subsidies. When we’ve mined the last vein of tantalum or niobium, then there will be nowhere to get it except for those landfills, but by then getting it out of the landfills will be more costly than getting it out of the computers and celphones would have been before they were all jumbled together with all sorts of other trash. Today’s distorted economy makes tomorrow’s prospects even bleaker than they had to be. This is exactly the same “running off a cliff” that I just talked about, but with much more serious consequences.

The really sad thing is that this distortion of our economy is mostly done in the name of capitalism. The costs of resource extraction are diffuse, so who pays? Proximately nobody, but ultimately everybody. If somebody dumped a ton of mercury-sodden leach-mining residue on your property, you’d sue and win millions. If they leave it on “public” land, to which they’ve been granted exclusive rights for a pittance, and it seeps into the aquifers and ends up affecting millions, nothing whatsoever seems to happen. Modern (and particularly American) pseudo-capitalism is based not only on government-subsidized land (and ocean) use, but also on the false premise that approximations of costs are not acceptable. Unfortunately, for costs like pollution and resource depletion, approximations are all we have. We can approximate the cost of an activity to each person, and the share of that cost for each entity incurring it, and levy taxes or fees or fines accordingly, but the laissez-faire types resist that at every turn because there’s no perfect proof that this exact cost was incurred by this exact party. So the cost gets dropped on the floor, which we can all tell is an even worse approximation than what we could come up with otherwise, but it’s one acceptable under the prevailing dogma so it’s allowed to happen.

Until it can’t happen any more, because nature and the laws of thermodynamics won’t allow it. There is no free lunch. You can’t get something for nothing. All you can do is defer the cost until later. If anyone ever wonders why I hold the laissez-faire crowd in such utter contempt, it’s because they’re mortgaging my and my children’s future for the sake of their own greed and sloth. Worse, they try to sound noble doing it, misappropriating terms like “liberty” and “freedom” to support their agenda while denigrating a more rational one. They’re stealing not just physical resources but intellectual ones as well. Well, sooner or later someone has to clean up their messes, and it begins with sweeping them onto the trash heap. We need to kill this stupid idea that resources will always be abundant, that there will always be some new frontier to exploit or some technological miracle that will save us from ourselves. Technological miracles are in large part how we raised our consumption levels, but they haven’t affected resource levels. All three laws of thermodynamics still apply, no matter how advanced our technology becomes. We must start performing our calculations based on an assumption of finite rather than infinite resources, and under that assumption small or diffuse costs can no longer be ignored. If there is a technological miracle that will save us, it’s one that will allow us to make better use of those finite resources, and to turn our technology in that direction we must turn economics and society in that direction. Such technology will not develop in an economy distorted to make it unprofitable, and it might no longer even be possible in the economy that will come after. Just say NO to confusing waste with profit.

UPDATE: After I wrote this (but still before it got posted) I found a new report about how water pollution from mine wastes almost invariably exceeds that predicted in environmental impact statements – like that’s news. The figure given for cleanup works out to about $2 for each man, woman and child in the US. That’s only for the sites on the Superfund National Priorities List, and is probably a drastic underestimate even then. How do you like having your taxes pay the bills for companies that pollute our groundwater?

Keep Them At Home

I was reading Liberal Street Fighter’s article about congressional work habits (or perhaps the lack of them) and it got me thinking. The article mentions how many recesses congresscritters currently enjoy. Well, how many recesses should they have? After all, it is good for them to go back and visit their constituents once in a while, isn’t it? How can they get back in touch with those constituents if they’re tied down in Washington practically all the time. Then it hit me: LSF has it exactly backwards. We shouldn’t be keeping them in Washington; we should be keeping them out of Washington. We should make more recesses mandatory. We might not be at the technological point where we can have direct participatory democracy, but we are certainly at the point where videoconferencing and such can allow congresspeople to do the vast majority of their work remotely – just as many others, particularly in high tech, already do. We can get them closer to their constituents, and further from K Street. Make it so that regular people can walk down the street to see their congressman, and lobbyists must get on a plane, instead of the other way around.

This might conflict with congressional rules, of course, but congressional rules are frighteningly anti-constitutional and anti-democratic anyway so screw them. More significantly, there might be ways that a work-from-home legislature might conflict with the constitution. For example, the term “assemble” is used many times and could quite reasonably be interpreted as requiring physical presence. For the most part, I think such issues can be worked around by having congress meet in person monthly (or less). If they can’t, then let’s change the constitution. Clearly the founding fathers did not anticipate modern communications technology, and if they could see it and if its use yields a result more in keeping with their ideals then I’m sure they’d approve. Where’s the down side?

Picture Friday

Here are the November pictures.

Amy among some rocks Amy sitting among some rocks at the Great Meadow, November 18.
Amy lying down in a leaf-pile Amy lying down in a leaf-pile in the back yard, November 21
Amy and Grampy Amy reading a book with Bob, November 23.

Network Algorithmics, part 2

In my previous post about this book, I took issue with some of George Varghese’s conclusions but generally spoke favorably of the book. Now, having read chapters 9-14, I’m not convinced the book as a whole is that good after all. While chapters 1-8 contained a lot of fairly objective and generally applicable ideas, the author’s biases really start to show in the middle section. For example, 90% of the techniques described in chapters 11 (Prefix-Match Lookups) and 12 (Packet Classification) are so domain-specific that they’re no longer useful as case studies. Contrary to the cover blurb’s claim that the material will benefit “anyone involved with network implementations” these chapters are very specific to high-end IP routers such as are made by a mere half-dozen companies. Anyone working on Fibre Channel or InfiniBand networks, or even IP gear that’s COTS-based and not filled with custom hardware, will probably find little of value. The few academics and Cisco/Juniper architects who care about this material probably know Varghese personally anyway, so there’s little point to putting it in book form.

Chapter 13 (Switching) was a lot more interesting, particularly from the perspective of someone who will shortly be working on a switching network even more exotic than those Varghese mentions. Chapter 14 (Scheduling Packets) is where it gets really bad. An entire subsection is devoted to Random Early Detection, but what little is said about alternatives is either misrepresentative or just plain wrong. You’d think RED-PD, CSFQ, FRED or Blue – all cited in Sally Floyd’s excellent list – would deserve at least passing mention by Varghese. Stochastic fair queuing in general gets a mere two paragraphs, with one citation to a 1991 paper; Stochastic Fair Blue is the best-known member of this family, and that papers’s from 2001. Then Varghese has this to say:

Stochastic fair queuing has two disadvantages. First, different backbone routers can hash flows into different groups because routers need to be able to change their hash function if the hash distributes unevenly. Second, SFQ does not allow some flows to be treated differently (either locally within one router or globally across routers) from other flows, a crucial feature for QoS. Thus, SFQ only provides some sort of scalable and uniform bandwidth fairness.

Wrong. Inexcusably wrong, because Varghese clearly knows better but isn’t telling the reader. The fact that routers might hash differently is not a weakness. In fact, some might consider it a strength, for the same reason that multiple hash functions in SFB are considered a strength. An ill-behaved flow might “slip by” one router or hash function, to be caught by the next, so its packets are still more likely to be dropped than those of well-behaved flows – which is exactly what active queue management is supposed to achieve. The part about QoS is just FUD. Varghese does get around to mentioning weighted RED, and it doesn’t exactly take an algorithmic genius to see how similar modifications could be made to SFB or practically any other queue-management algorithm.

Ignoring over a decade’s worth of research and raising bogus objections, just so he can claim SFQ “only” provides “some sort” etc., is simply shabby. So, by the way, is a Wikipedia page, written by a research collaborator and business partner, that looks more like a corporate “about the founders” page than something that belongs in an encyclopedia. I understand that authors have a tendency to write more about stuff they’ve worked on than about competing ideas (I’m not immune either) but that’s no excuse for misrepresenting the state of the art.

Things I Want

This is kind of part Christmas gift list, part notes to myself, and part just links to things I think others might also find interesting.

  • Tomy Flip Flap. A solar-powered plastic plant, perfect for cube living.
  • Despair.com Irresponsibility desk photo. If you don’t know despair.com, take some time and check them out. Many of these are hilarious. “Meetings” would be good for a conference room, “Destiny” or “Dysfunction” for an ex-lover, etc. I like “Irresponsibility” because, besides the cool picture, it’s actually a pretty important principle. Nobody thinks they are responsible for destroying the environment, for screwing up the product they work on, for electing George W. Bush, for the prevalence of junk food and bad TV/movies – and yet millions of people are responsible for those things. All those actions and decisions and votes and purchases do add up. The positive side of that is that small good things add up too. If more people thought about the relationship between raindrops and floods the world would be a much, much better place. I like the picture as a reminder of that.
  • Piano Tribute to Pink Floyd. That just sounds like an awful idea, doesn’t it? Well, I happened to pick up the tribute to Dido for cheap, and it turned out to be remarkably good. In fact, I like the tribute versions of some songs better than the originals, especially where the piano can convey a particular feeling better than Dido’s own not-so-strong voice. Not all of the tributes seem to be of the same high quality, but Vitamin does provide samples and the Pink Floyd one seemed to be of the same excellent quality.
  • Disappearing Civil Liberties Mug. I don’t actually want this one for myself because I’m not that much of a hot-beverage drinker and I already have a rare Angst Technology mug I like to use, but I thought other people might like this idea. They also have one that shows the effects of global warming on sea levels.
  • A good online wish list that’s not tied to a specific vendor. I use my Amazon wish list for whatever I can – mostly books, music and games – but none of the items on this listare available through Amazon. Sometimes I don’t even have a specific make or model of something in mind anyway, but I’d like to put it on a gift list anyway. There seem to be bunches of sites offering this functionality, but I have no idea if they’re any good. Some of the online to-do lists are nice, and Ta-da List’s random list of lists is kind of fascinating in a voyeuristic way, but that’s not quite the same thing. Something like Amazon’s priorities and quantity-wanted/quantity-purchased handling (to prevent getting two of the same thing without spoiling surprises) would be really nice. If anyone has any experience with such a thing, please let me know.

Musical Tastes

I had a tooth extracted today. That’s probably of little interest to anyone but me, except for the potential humor value of the song that was playing on the radio at the time.

Well, everybody hurts sometimes,
Everybody cries. And everybody hurts sometimes
And everybody hurts sometimes. So, hold on, hold on
Hold on, hold on, hold on, hold on, hold on, hold on
Everybody hurts. You are not alone

Yes, really – “Everybody Hurts” by R.E.M. Cute, huh? The dentist (oral surgeon, whatever) was humming along too. He even commented at one point on this being the ideal song for a tooth extraction. If I hadn’t been in a situation where laughing might be extremely dangerous, I might have allowed a chuckle or two.


My PHP script for copying a directory tree directly from a web server (like this one) to Amazon S3, tentatively called PHP2S3, has reached a point where it might be worth sharing with other people. It was actually a lot simpler than I thought it would be. Writing the code was easy; it was finding the right tools and libraries and such to facilitate development that really took all the time. Now that I’ve climbed that “learning” curve, I could probably do something twice as complex in half the time – and yes, I’m open to suggestions. Just for fun, I’m distributing the package available via S3 itself. The tar file contains the script plus the S3 library pre-patched as mentioned in my previous S3 post.

What’s Good for Richard Epstein

Richard Epstein had a piece in the Boston Globe’s “Ideas” section Sunday, called “What’s good for pharma is good for America.” He seems terrified that “the new Congress will seek to intervene on such hot-button issues as FDA oversight of drug safety, patent protection, and drug pricing.” Horrors! How dare the US Congress take an interest in drug safety or patents? As one might expect from someone who has just published a book on the subject and (according to the biographical note buried where most people won’t read it) “has from time to time consulted for Pfizer and PhRMA” Epstein’s presentation of the facts is a little one-sided and his arguments a little less than completely forthright. I’ll try to address the worst of his transgressions against civilized discourse.

The huge profits of major drug firms are often tied to one or two drugs, such as Pfizer’s Lipitor or Viagra – profits that evaporate when their patents expire and generics enter the market.

No, not really. For example, I can still find brand-name Zantac at the supermarket, and it’s still priced at a premium over generic equivalents. Brand loyalty is a powerful thing, and the company is still making plenty of money from it even though the patent has expired.

products with $21 billion in US drug sales are going off patent in 2006, wich another $24 billion to follow over the next three years – a sharp dent for an industry that today generates about $250 billion in revenue.

That’s losing something less than 20% in revenue over four years. Sounds serious, doesn’t it? Well, not really. For one thing, as noted above, not all of that $45B will simply evaporate. For another, much of it is likely to be replaced by revenue from other drugs as they hit the market. Lastly, is that 5% revenue “loss” per year the difference between an overall net profit and loss, or the difference between a modest profit and an obscene one? Notice how Epstein doesn’t pause to consider any of these issues. He has an agenda to push, and a family to feed from the proceeds for pushing it.

All the while, the pharmaceutical houses also must absorb the legal and business risks needed to identify, patent, test, license, and market any new drug.

Yet again, not really. In fact a substantial portion of those costs, particularly during the test phase, is borne by others. The National Institutes of Health in the US, and similar bodies elsewhere, spend billions funding various aspects of this activity. The pharmaceutical is quite heavily subsidized by US taxpayers, which fact surely should be accounted for as Epstein tries to raise the alarm about how vulnerable drug companies are.

surely oncologists can do a better job calculating the odds [of a drug's effectiveness] than the FDA, which has to deal with averages, not individual cases

One of the first things most people learn in their first statistics class is the danger of small samples. “Better” when applied to calculating odds means smaller margin of error, and the margin of error decreases with increasing sample size. The fact is that a particular oncologist, presented with only one or even a dozen cases, can not calculate the odds better than the FDA can with a larger sample. What Epstein is attempting is appeal to emotion, putting the image before us of a patient and doctor making hard decisions, and not appeal to the reason of statistical reality. One could counter with an image of a patient harmed by a drug that was inappropriately put on the market, but then Epstein would bring out the trusty double standard to accuse others of what he has already done.

Often they [personal injury claims] are propelled by inflammatory trial techniques that obscure the scientific evidence, which lay juries find hard to assess in the first place.

What “inflammatory techniques” are these, exactly? How are they different than the techniques that drug-company lawyers might use to bamboozle juries the other way? Epstein doesn’t say. Yes, general scientific and statistical illiteracy is a problem, but it’s a problem for both sides. What kind of evidence of harm, exactly, would Epstein suppress? What other evidence would he prefer to see accentuated, that’s not just as obfuscatory as what he’s complaining about? Does he really want anything other than license to mislead and intimidate jurors?

It is stark evidence of how dire the situation is for pharmaceutical companies that the FDA, typically no friend of the drug companies on safety issues, has now actively intervened on their side in personal injury suits that attack the adequacy of FDA approved warnings.

It’s evidence of no such thing. The FDA is countering claims that they didn’t do their job. Why is any explanation other than self-defense necessary? If we were to dismiss that obvious motivation, we still might not go where Epstein wants because, contrary to how he puts it, the Bush-administration FDA (like every other Bush-administration department) has actually been quite “friendly” to big business. Their first motivation is to protect themselves, their second to protect their friends. Sadly, serious concern for the truth or for the impact of such cases on the public is likely to be a distant third on the list of reasons why the FDA might involve itself in such cases.

Critics also naively assume that investors and firms will continue to make huge investments in new products without any assurance of recouping their costs in the marketplace.

On the general issue of price controls, I almost agree with him. The market, not the government should set prices. It should not hold them artificially low, nor should it push them artificially high by subsidizing drug trials or offering patent protection that contravenes the public-interest reasons for having a patent system in the first place. Here, though, Epstein simply overplay the only decent hand he has seen. It is not the government’s job to shift corporate risk onto taxpayers, or to guarantee markets for anything. “Making huge investments … without any assurance” is exactly what companies are supposed to do, and if the risk/reward ratio is good enough they’ll take the risk. After all, we certainly hear enough “ownership society” rhetoric about how it’s bad for the government to assume personal risk e.g. in the form of Social Security or health care. Why should the government do more to eliminate financial risks for corporations than for flesh-and-blood people?