The Box: How the Shipping Container Made the World Smaller and the World Economy Bigger is, not very surprisingly, a history of the shipping container. While a shipping container itself is just a rectangular steel box, the changes it has wrought in the world economy are spectacular. Marc Levinson details the many changes, which range from the economic effects of changing freight rates to the utterly different demands that traditional shipping and container shipping place on ports. Living in today’s world, it is surprising to read the accounts of how international shipping worked in pre-container days, with the loading and unloading absorbing almost half the cost of a transatlantic voyage before containerization.

Longshoremen were prosperous and had a strong union, which held a tight grip on the loading business, bordering on the absurd:

One formal rule provided that, once assigned to a job at a particular hatch of a particular ship, a worker would do only that specific job until the ship sailed; if loading was complete at one hatch but not at another, an idle worker from the first hatch could not be shifted to help out at the second. An important “hip pocket” rule, codified nowhere but enforced by the gang foreman as required, provided that a trucker delivering palletized cargo to a pier would have to remove each item from the pallet and place it on the dock. Longshoremen would then replace the items on the pallet for lowering into the hold, where other longshoremen would break down the pallet once more and stow each individual item – all at a cost so high that shippers knew not to send pallets to begin with.

Still, the history of the shipping container was by no means a straight success story – the book recounts a long and fitful history, because container shipping alone was not enough to realize the full potential – politics and infrastructure had to adapt, standards had to be set, shippers had to learn how to optimize their cargo. The book also makes clear that we have not yet reached the end of its development:

Where vessel size had once been limited by the locks in the Panama Canal, containerships had grown so large that twenty-first century naval architects were contrained by the Straits of Malacca, the busy shipping lane between Malaysia and Indonesia. If a containership ever reaches Malacca-Max, the maximum size for a vessel able to pass through the straits, it will be a quarter mile long and 190 feet wide, with its bottom some 65 feet below the water line. If it should sink, it will take nearly $1 billion of cargo with it. Its capacity will be 18,000 TEUs or 9,000 standard 40 foot containers, enough to fill a 68-mile line of trucks when it arrives in port.

(Topical link: the Panama Canal expansion project just got under way, in a bid to redefine Panama-Max to be larger than Malacca-Max.)

Marc Levinson does a commendable job collecting and presenting a huge amount of facts and exploring the wide-ranging changes wrought by containerization. Even if his description of Malcom McLean sometimes seems to border on hero worship, the later context makes it entirely clear that a single genius was not enough to change the world (and that he was fallible, to boot). Though “The Box” is rich in data and has a large bibliography, Levinson’s style is far from boring or overly scholarly. Recommended to anyone with an interest in economic history, especially for its illustration of how seemingly mundane developments can have profound influence on the ways of the world.

Hey, we went almost two month without a rant post here, so it’s definitely time again…

The recent purchase of a ThinkPad X61 Tablet once again made me angry at the software world. I’m not even going to talk about the preinstalled Windows Vista, I’ll restrict myself to stuff I noticed on Linux: Is x86-64 a new invention or is it the platform that pretty much any PC sold in the last two years is based on? If the latter, then why is it that I can’t get a Java plugin, Java Web Start or Flash for my x86-64 system (without installing a parallel i386 environment, which is inacceptable)? This is annoying. I can see Adobe facing some work in porting the Flash VM, but the Java stuff is just plain stupid – they have the entire JRE working, but not those two simple bits (which happen to be critical for 95% of the things I’d use Java for)?

The real fun started when I tried to install the version of Maple 10 that my university so nicely provides at deep discount. Turns out that doesn’t come with the x86-64 Linux version (the installer thinks there should be one, but it’s just “file not found”). I took that as a sign that it was time to finally switch to Mathematica. In the past, I had considered buying the student version of Mathematica a couple of times, but had always been dissuaded by the necessary offline paperwork – apparently it was necessary to fax documents to the Austrian distributor. Now, I was greeted by a friendly offer to “Download Mathematica now!” – I was intrigued.

Letdown #1 was that I was apparently required to buy the Linux and the Windows version separately, unlike the procedure on my student Maple that just had a “Win/Mac/Linux” license. I cursed, but then decided to give up on Windows mathematics for the moment.

I made my way through the shopping system and actually ended up with a download link – hooray! I got the installer, even the mail with the license code arrived. While it was downloading, I started reading Mathematica documentation in anticipation of the fun to come. With the download complete, I launched the installer and went to input all the necessary information until I was a bit surprised to see a

Password: 

prompt. Password? Wtf? Turns out I have to register (again, type in all the information I gave to the web shop half an hour before) to get a password so that I can actually install what I just bought. By the way, this password will be valid on only one machine. More curses (I like to have the same software on my desktop PC, private laptop and business laptop – and what the hell is wrong with that?). I fill out the web form, in an increasingly bad mood. I submit the form. The response page thanks me for registering, but does not contain a password. Wtf? The answer is in my mail box:

Thank you for registering on the web.

A Wolfram Research Customer Service representative
will contact you regarding your registration.

If you have any questions about Wolfram Research or our products,
please contact Wolfram Research Customer Service.

Yeah, I have a question. Have you nitwits ever heard of Bittorrent? Hey, let’s have a look at mininova.org:

So, if I really wanted to “steal” your software, maybe I wouldn’t have started by paying you 128 euros. Is it really necessary to make the purchase of software so painful? I feel reminded of the worse commercial music download services.

(And yeah, I know that I’m just an “apt-get” spoiled brat and can actually remember the days when buying software always involved mail-ordering a box with 35 diskettes, of which one was invariable damaged.)

Update: One day later, I received an email from Wolfram customer service cheerfully informing me that my order “had invoiced”. No password. In my desperation, I go to the “register” page again, and enter all my info for a third time. This time I’m rewarded with a message saying my password will be emailed to me. I get an email, and it contains a license key that will expire on September 21. To get a permanent license, I have to send them proof of my student status before that. So much for the new and simple download experience.

Update2: Just to annoy me, my university now offers a student version of Mathematica that comes on a Win/Mac/Linux DVD and costs all of 13 euros. Nyech.

Pet peeve, as triggered by the Financial Times, here:

How computers routed the experts

By Ian Ayres

[...]

In Freakonomics, Steven D. Levitt and Stephen J. Dubner showed dozens of examples of how statistical analysis of databases can reveal the secret levers of causation. Yet Freakonomics didn’t talk much about the extent to which quick quantitative analysis of massive datasets – call it “super crunching” – is affecting real-world decisions. In fact, decision-makers in business and government are using statistical analysis to drive a wide variety of choices – and shunning the advice of traditional experts along the way.

Instead of simply throwing away the know-how of experts, wouldn’t it be better to combine super crunching and experiential knowledge? Can’t the two types of knowledge peacefully coexist? There is some evidence to support this possibility. Indeed, traditional experts are shown to make better decisions when they are provided with the results of statistical prediction.
But evidence is mounting in favour of a different and much more dehumanising mechanism for combining human and super-crunching expertise. Several studies have shown that the most accurate way to exploit traditional expertise is merely to add the expert evaluation as an additional factor in the statistical algorithm.

My problem with this is the framing as “human expert vs computer” by people who don’t really understand what computers do. Here’s my story:

In all of the cited cases, as is adequately explained by the article, computers don’t make up the models. What we’re seeing is not human knowledge replaced by computers but simply newly established human knowledge competing with “traditional” human knowledge.

Cheap computing power has made statistical analysis as a tool for testing and creating hypotheses much more competitive, and so it’s not very surprising to me that we see it gain an increasing mindshare over the more traditional “human” expertise. While the foundations are nothing new, the computerized way of infering and testing hypotheses on large datasets is increasingly able to beat the traditional “synthesize knowledge from what the regularities of your observations” way. However, the rules are still made by human experts – it’s just that the experts now use new tools and perform heavy computation.

The emotional component of this “dehumanizing mechanism” is just that there’s a component that is apparently mysterious to people who don’t know the mathematics and/or technology involved. To me, if someone is able to predict the quality of wine just by observing the amount of rainfall in the year of the harvest, that person is as much an expert as someone predicting it through a poorly understood process of extrapolation from the taste of an intermediate product. There is no dehumanization of the process just because the hypothesizer used a computer to create and test his model or because a computer is used to calculate the score.

The boundary of dehumanization is a bit unclear to me, but I’d say that what we’re currently facing are much more organizational threats than technological ones. For example, a credit score is something that is algorithmically created for a person and has a huge influence over their lives, but the details of the algorithm and its parameters are secret, so I can understand that it makes you feel powerless to grope in the dark and always have the fear of “blemishing your credit” over your head when considering something as innocent as cancelling a credit card you no longer want. Still, I assume there is still someone who actually knows how this thing works. The real boundary of dehumanization is probably reached when the system reaches a point of complexity where it’s not possible to come up with a verifiable justification for a certain input/ouput combination because there are so many tunables and cross-references in the system (come to think of it, I’m not convinced that that point hasn’t been reached in the credit-scoring system, yet). I don’t even think it takes “true” artificial intelligence to reach such a point, just a huge software system that grows until nobody really understands everything that went into it. I really enjoy the thought of executives dancing around a glowing mainframe that houses their ancient decision support system that always gives the right answer, but that nobody really understands. I just hope we can keep stuff like this out of our laws.

Update: Here’s Marginal Revolution on the author of the article (who apparently knows what he’s talking about – so I wonder why the article is framed like that?)