Thursday, 29 May 2008

Peak Oil, Poverty, Moore's Law and Manure

Peak Oil

As oil and with it products we use to fuel our cars and trucks gets more expensive, there is growing anguish about the effects on the cost of living. China alone is adding millions of cars a year to the total, and peak oil theory says we should be hitting limits soon on production growth.

In reality, as prices go up, there will be options of exploiting kinds of oil previously too expensive: under the deep sea, in the arctic (conveniently being freed of ice; what's doing that, I wonder, if there's no climate change?), tar sands, shale oil… so production may not drop off as fast as predicted by peak oil theory. But should we want to squeeze every last drop of oil (and coal and gas) out of the ground?

Not only do we have climate change to worry about, but the economics of fossil fuel has a lot to do with the gap between rich and poor nations.

The Poverty Gap

In the twentieth century, the cost of communication increasingly split rich from poor. By communication, I mean movement of people, goods and information. Part of this was energy; another part telecommunications. The common thing was the distinction between countries with a comprehensive network of roads, rail, electricity and telephone connections, and those without.

Addressing this gap became increasingly hard, as the cost of new infrastructure has to compete with infrastructure created in an era of lower costs (e.g., coal was cheaper because demand was lower). It is this gap which for example makes fruit in a third world country absurdly cheap in tradable currency terms, while making a local phone call is ridiculously expensive. There's a kind of arbitrage, but one where the places where the price differences occur are too hard to connect, to correct the pricing anomaly (as would happen for example if the US$ to euro exchange rate was out of synch with the US$ to pound sterling exchange rate).

Somehow, despite all this, we have arguments from climate change inactivists that addressing poverty is an alternative to addressing climate change. Yet if you accept that poverty is largely structured into communication infrastructure (or lack thereof in poorer countries), new technologies that reduce the need for infrastructure can go a long way to closing the poverty gap.

An example is the cell phone. In many poor countries, cell phone roll-out has been many times faster than predicted, because of communication starvation. People in Nigeria didn't have phones not because they were poor, but because there was no infrastructure. Cell infrastructure is relatively cheap to put in: as long as you have electrical power, you can virtually parachute base stations in. By contrast, a nation-wide wired phone network needs wiring to the home, with extensive local wiring, even if trunk lines are wireless.

This example generalizes to other cases like electricity. If you can generate power locally without a grid, you can make energy accessible a lot faster in countries without infrastructure. Eliminate the cost of consumables, and you also eliminate another huge problem: rapid price increases as supply fails to keep up with demand. If you think this is bad for wealthier countries, what will doubling the fuel price to to someone who can barely afford a car?

Moore's Law

The nice thing about renewables is that technology changes reduce prices. It doesn't matter if coal supply runs low, oil runs out or gas slows to a trickle. The wind will still blow, the sun will still shine. Just as Moore's Law has pulled computer prices down dramatically over decades, photovoltaics and wind are getting cheaper. Eliminate the consumables and you have a real revolution in energy economics, far bigger than Henry Ford's revolution in personal transport.

This is an exciting time once we forget doom and gloom and think of what could actually be done.

Instead of living in terror at consequences of change, how about accepting that we are looking at a change as big as the move from horse and buggy to cars?

Think of it this way: the cost of renewable energy sources only depends on the technology, not consumables. Once we get this right, we can make energy cheaper with every new development. As long as we are stuck with fossil fuels, prices can only go up as demand overtakes supply.

Why are we so scared of this great new concept? Once we solve the energy storage problem (there are already good ideas like compressed air) poorer countries will benefit too. What's the downside? Unless you own a coal mine ... the horse industry lost big time when the Model-T appeared, but think of the advantages to society as a whole: personal mobility on a level never experienced before.

Now we have the next level: energy with radically lower constraints on supply and infrastructure. If you can do local microgeneration with efficient storage, you no longer need a grid. In rural Africa, for example, you could almost parachute in (where have we heard that before?) a solar or wind microgenerator.


Back in the nineteenth century, it's alleged that someone predicted that London would be metres deep in horse manure in a few decades. It's a nice story, even if it's improbable that it's true (I have yet to find a direct source for it, and there are several variants – the hallmarks of a myth or urban legend). It is certainly true that getting rid of horse manure in the streets of major cities was a growing headache – just as cutting carbon emissions is today.

So why, today, are we staring down the problem of curbing carbon emissions when there are far superior alternatives – alternatives that only need a little development to be viable?

The only alternative is to consume fossil fuels like there's no tomorrow, then it's back to the stone age. Why are we even debating this?

Tuesday, 27 May 2008

The Great Mailing List Removal Spam Bomb

Every now and then someone creates a mailing list to announce something and makes three rather fundamental errors: they set it up so that the reply-to address is the whole list, they allow anyone to post to it, and they don't make it as clear and obvious as possible how to unsubscribe.

What really boggles me is that this seems to happen most often in announcements of technical conferences of interest to people who should have an above-average understanding of how email and mailing lists work (though I suppose this could be a factor of the company I keep).

Inevitably, every now and then, when one of these announcements is sent out, someone replies to the entire list asking to be unsubscribed. This triggers a cascade, where others, who would have ignored one unwanted announcement, ask to be unsubscribed too – all such requests going to the entire list (the lesson that they were annoyed apparently having failed to sink in).

The most recent such instance I've encountered was a conference for Field-Programmable Technology (ICFPT 08), about as technical a concept as you can get, and one only hard-core computer scientists and computer systems engineers would be interested in. Unless the original mailing list was misconfigured and included a batch of classics professors who had been inducted to the Internet the day before, it's hard to see how this list could have gone to a significant number of people who would fall into the unsubscribe spam error. But it did.

The result? At least 35 unsubscribe requests to the list, including two (apparently unable to see the irony in their response) which included requests to others not to post to the entire list, and about 15 requests to stop spamming the list.

You have to wonder how this could happen in a technical field. There is of course a fundamental flaw in the way email works which makes it impossible to prevent this sort of thing. There is no limit on how long you can make a recipient list, so there is no way to force mailing lists to be set up according to a specific design that would eliminate this sort of problem.

Of course there are much better solutions out there. A proper mailing list manager can be set up which puts clear and obvious instructions on each message on how to subscribe, and announcement lists should never be set up to allow anyone but a short list of authorized posters to post messages.

What's to be done?

Until we can replace email by something better that would prevent this sort of thing by design, all I can do is point out the idiocy of this sort of behaviour and hope the message gets out. But not repeatedly to the same mailing list. Next time this happens, I'll post a pointer to this article.

Saturday, 24 May 2008

Where is Apple Headed?

One end of Apple's range is being ignored and it's not the iPhone.

There's something of a feeding frenzy going on around iPhone 2.0, with claims that inventory has dried up (all models at time of writing show "Currently Unavailable" at the Apple Store, a sure sign that something is up), Apple has imported unusual consignments to the US, various phone providers leaking alleged attributes of the new model, etc.

That there will be an iPhone 2.0 is a given. That it will appear soon looks very likely. But is this all?

If you mosey over to the Apple store, you will notice that the Mac Mini has ceased to make much sense. The fastest model comes with only a 2GHz processor for $799 with 1GB RAM and a 120GB hard drive, while the base iMac has a 2.4GHz Intel Core 2 Duo and a 250GB hard drive and significantly better graphics hardware. Option the Mini up with a 160GB drive, and you have an $849 machine before you've bought a screen, mouse and keyboard. Go cheap on those e.g. via Dell, and you add about $280 to get USB mouse and keyboard, and a 20" screen, total $1129, only $70 less than the base iMac – despite a significant loss of features and speed.

So I predict the Mini is due for an overhaul.

What else?

Apple TV remains a half-baked product. Who heard of a TV without a tuner? You can buy devices like EyeTV for a Mac then export recorded video over to an "Apple TV" to play. But this is silly. You want this part in the box called the "TV" – and you want to be able to watch live TV, not only recordings, from your EyeTV. Why not support a developer interface to Apple TV that would allow devices like EyeTV to be implemented even if Apple isn't interested in the concept? After all, many people have made a big deal of the absence of an FM tuner in iPods, a need that's been filled by third party developers.

In any case, since no one else is talking about this, I propose that Apple do something a bit more creative: merge the Apple TV and Mac Mini into one low-end product that can be configured as a pure black box media device, or used as a more general-purpose computing device with options to add on your own features.

All that it would take is opening up the interface to Apple TV, and converging the feature set of the two devices with some creative build to order options. With a bit of a stretch the feature range could go from a relatively simple low-end box to something close to an iMac without a screen.

The overall lineup then would look something like this:

  • iPod lineup – Shuffle, Nano, Classic, Touch: eventually Classic will go when flash is cheap enough; no big change except Touch will get any appearance overhaul to match the new iPhone
  • iPhone lineup – soon to be clarified but I would be surprised if the variations widened significantly before Apple had a worldwide presence
  • entertainment Mac lineup – Apple TV with options up to an attractive Mini, comparable to an iMac
  • iMac – no change since there has been a revamp recently
  • Macbook – no big change
  • Mac Pro – no big change

So that's my prediction: not only an iPhone overhaul but something at the low end. The iPhone is obvious; the other more a question of fixing some flaws and inconsistencies.

Friday, 23 May 2008

Smart Alternatives to Fossil Fuels

In 1894, the Times of London estimated that by 1950 every street in the city would be buried nine feet deep in horse manure.
– Eric Morris, From Horse Power to Horsepower, Access no. 30, Spring 2007

This prediction has something of the character of predictions of the effect of climate change. Of course, what happened in the meantime was that trains and cars largely replaced horses (and despite the apparently exact reference to the Times, it is more than likely that this quote is a myth*, unlike climate change for which there are many verifiable quotes – but still, it's a nice thought to build on). So what will the equivalent changes be in the twenty-first century?

The nineteenth century, despite the ubiquitous horse, was the century of coal – first in factories, later in ships. While the concept of steam power went back further, with the major developments in the eighteenth century, it was the nineteenth century when coal and steam became dominant modes of energy use and conversion.

The twentieth century was the century of oil. While coal continued to play a major role especially in power generation, oil took over as the fuel of choice for locomotion. With the exception of electric trains (and minor exceptions like golf carts and milk floats), almost all transport by the end of the twentieth century was powered by oil.

Now that climate change is becoming a concern and the oil price is hitting record highs almost on a daily basis, we have to contemplate: what next? The answer, according to climate inactivists (thanks to RealClimate for this word), is to carry on as before. Why? Aside from a charming belief that if humans could effect the climate in the past, we can't now, there's the argument that substantial reduction in greenhouse gas emission would trash the economy.

The Australian Broadcasting Corporation (ABC) has a science show, Catalyst, which often has good stuff in it. On 24 May 2008, they ran a segment on how London is planning to cut its greenhouse gas emissions by 60% by 2025.

How are they planning on achieving this? Back to horse and buggy, lights out 2 days out of 3, medieval torture for anyone who fails to comply?

No, explains Allan Jones. The trick is to understand the "rubbish way things are done" and do better. What are his tricks? One of the big ones is combined heat and power, actually an old idea, before power utilities managed to persuade everyone that remote power generation was the answer. How does this work? Burning fuel to generate electricity wastes most of the energy produced as heat. Much of the electricity produced is then used to produce heat. Combined heat and power replaces large remote power generators by smaller local ones, placed where the "waste" heat is useful. While the theoretical efficiency of the smaller power station is lower, being able to use the waste heat more than compensates.

Switching from coal to gas completes the picture in terms of greenhouse gas (and general pollution) reduction.

Overall not only are you saving by making use of the two thirds of the energy in the form of "waste heat" but you are saving on losses through the grid.

The result? Not only lower emissions overall, but lower cost.

So much for the notion that radical cuts in greenhouse gas emissions has to be at the expense of the economy.

What else is Allan doing? Another part of his recipe is extensive installation of solar power: expensive in the past, but increasingly competitive with fossil fuel-based power.

This is of course but part of the bigger picture. Reducing car usage, moving to modes of public transport that don't burn fossil fuels and addressing industries that produce greenhouse gases by other means (e.g. cement, steel) would all have to go into a more comprehensive solution. All of these are engineering problems, potentially solvable with known techniques. An alternative geo-polymer cement, for example, can be made using a different chemical process, which does not produce CO2 – reported in the same Catalyst show as the London report.

So why is this so hard in most of the world?

Entrenched interests in the way things are now don't help. If you replace decentralised coal power by centralised gas and solar, existing coal-fired power plants will have to be written off at great cost to the operators. Coal mines will also be closed – again, a cost to the operators.

In Australia, the state of Queensland has massive reserves of coal seam methane which can be used at lower cost than coal to generate power to make it possible to replace every coal fired power plant with much cleaner power. The coal lobby of course doesn't want us to know this because they don't own gas infrastructure, and don't want to risk losing their investment in existing plant and equipment.

What the Catalyst segment has done is to show what can be done if you don't put oafs in charge.

As for oil, we should be preserving it for jet fuel, the one application for which there isn't a clear replacement other than oil from coal, which is a very dirty option.

Bring on reliable commuter transport, high-speed inter-city rail, practical electric cars ... in 50 years, oil-based transport will be as quaint as a horse and buggy today. And the notion of CO2 levels rising to double the pre-industrial level will be as quaint as the idea of a city 3m deep in horse manure.

* Why is this likely a myth?
Horse manure did not just lie in the streets. It was flushed down sewers. The true twenty-first century analogue to failure to plan for paradigm change would have been a massive project to upgrade sewers to handle more horse manure – just as today some cities are still building new roads and tunnels on a large scale, as if the private car will be with us in increasing quantity for decades.

Thursday, 22 May 2008

Science, Governments and Industrial Impact

The still new Rudd government in Australia has announced deep cuts in funding for the Commonwealth Scientific and Industrial Research Organisation (CSIRO), one of the few research bodies to receive consistent funding under the previous anti-intellectual Howard government.

The CSIRO and to a lesser extent universities are suffering from the misplaced desire by governments for research entities to function as commercial operations. If what they were doing had a clear and obvious low-risk commercial outcome, business would be doing it.

The purpose of government funded research is longer-range outcomes which may not have immediate economic impact but when they do turn into something economically viable are game-changing. At Stanford, rightly acknowledged as a world leader in not only blue sky research but in industrial outcomes, the university does not try to hoard IP. If a grad student or academic has a great commercialisable idea, they are encouraged to go out and start a business.

By trying to merge the concepts of blue sky research (needs the stability of a large organisation with deep pockets) and innovative start-ups (need the nimbleness of a small organisation without deep bureaucracy) you end up with neither.

As to the question of the Rudd government's commitment, look no further than the "means test" on solar panels. This is a government which is focused on what it takes to win the next election, not what makes sense for the environment, academia, or anything else long term – despite anything they say to the contrary.

But back to the CSIRO.

From what I know of the organisation, it does some really good work, but has an excessively bureaucratic culture, with an emphasis on booking everything against a project, whether it makes sense to do so or not. This emphasis arises from a desire to appear to be operating "commercially".

What should a government wanting to make best use of a resource like this do?

I would take away the imperative to commercialise, and institute benchmarks similar to those for research academics. I would base funding on outcomes against these benchmarks. Some of these benchmarks would include long-term impacts (e.g., the fraction of research that resulted in a commercial outcome; the fraction of research that was highly cited). But I would not require that the organisation itself do the commercialising. If the CSIRO publishes some breakthrough concept which results in a major new business being started, why should the government care who owns the business, as long as there's a return to society?

The same applies to universities. The University of Queensland proudly compared its income for patents and royalties with that of Stanford, at a talk I attended a few years ago. However, that's not how Stanford operates. The university seldom tries to claim ownership of IP. Academics and grad students have formed many successful start-ups (HP, Sun, SGI, FedEx, to name a few).

The big difference in the US which makes the Stanford approach work well for them is the generous tradition of alumnus donations. However, even without this, a switch to separating commercialisation from research would be beneficial. In the absense of donations, the government could add successful spin-outs as a factor in funding research.

By making this change, the CSIRO – and universities – could revert to doing what a large government funded organisation does best: providing job security for those developing ideas for the long term. Unfortunately, cutting funding without a structural change is likely to have the opposite effect: an increased emphasis on commercial outcomes.

Thursday, 15 May 2008

Sunspots and Climate

It is bizarre that anyone is claiming that we should be basing climate models on sunspots at all. What is important is how much energy is incident on the planet, and how the overall energy balance changes. Sunspots may be a proxy for solar output, but we have direct measurements in the modern era, specifically satellite-based measures of total solar irradiance (TSI).

Nonetheless the thing persists.

In April 2008, an article appeared in The Australian alleging there had only been 3 sunspots since January and hence we were headed for an ice age. This article was rebutted by Melbourne academic David Karoly. On 15 May, a letter appeared in The Australian, claiming that Karoly had confused sunspot number (a calculated value) with number of sunspots. It's possible Karoly did, though he was correcting the previous article's manifestly wrong claim that there had only been 3 sunspots in total over a period when there had been many more.

This latest letter claims:
One visible sunspot is represented by a sunspot number of 11. Karoly writes that “the average number of sunspots a day last January was 3.4, followed by 2.1 in February and 9.3 in March’’. In fact, those numbers show months with many days of zero observed sunspots - a very quiet period on the Sun that has now extended longer than expected. Whether that affects the weather cannot be so easily disregarded as does Professor Karoly.

There does seem to be an interesting correlation between the length of sunspot cycles and climate change going back over centuries, compared with the 150 years that is all the CO2 warmers can work with.

There are several problems here.

For a start, it is incorrect to say that a sunspot equates to the number 11 in sunspot numbers. Let's see what the actual definition of sunspot numbers is:

An observer computes a daily sunspot number by multiplying the number of groups he sees by ten and then adding this product to his total count of individual spots... each daily international number is computed as a weighted average of measurements made from a network of cooperating observatories.

That is a nitpick of a nitpick so I will not dwell on it. The substance of the matter is his claim that sunspot cycles are a much better indicator of climate trends than modern models. Aside from the issue of more direct measures of solar output, the last sentence is bizarrely off target. Why should it be better to have centuries-long data of a phenomenon that did not exist centuries ago?

Anthropogenic carbon dioxide emissions began in earnest in the 18th century; by the close of the 19th century, they were less than 10% of current levels.

Saying that sunspot counts stood us in good stead for centuries so we should disregard other measures is like saying horse riding techniques stood us in good stead for centuries, so someone learning to drive a car should go to a riding instructor. In any case, the "centuries" claim is a tad inaccurate. The modern method of counting sunspots was invented in 1848.

Then there's the question of whether sunspot numbers actually correlate to the modern temperature trend. The IPCC has always stated in its reports that climate is a function of natural and anthropogenic influences, and even shows them separated out (here's a paper that does this nicely – see Figure 2(d) in particular, reproduced here). We would expect that solar output could be a major influence on climate before CO2 emissions became a major factor. But in recent years, this is not the case. In fact the trend in solar output is the opposite to that required to explain the warming trend in the late 20th to early 21st century.

If anyone asks nicely, I may even add in a few correlations.

But in the meantime, here is where you can find the data: