Another view of the CRU code leak
Michael Williams writing at
Master of None has a fun observation of the Climatic Research Unit code leak:
Anthropogenic Global Warming Will Lose Geeks Thanks To Bad Code
I'd venture that most software geeks are fairly leftist and generally support the theory that human activity is causing global warming. In that light, the biggest revelation from the recently hacked global warming emails might be the awfulness of the climate simulation code.
I've examined two files in some depth and found (OK so Harry found some of this)
Inappropriate programming language usage
Totally nuts shell tricks
Hard coded constant files
Incoherent file naming conventions
Use of program libray subroutines that appear to be:far from ideal in how they do things when they work
do not produce an answer consistent with other way to calculate the same thing
but which fail at undefined times
and where when the function fails the the program silently continues without reporting the error
AAAAAAAAAARGGGGGHHHHHHHH!!!
More code analysis.
I'm pretty proficient at writing simulation software: it's how I earned my PhD and how I earn a living. I've also worked closely with self-trained programmers who only write code to advance their research in other fields, and I can tell you that their code is almost always terrible. Writing good software is extremely difficult, and it doesn't surprise me at all that the climate modeling software is so bad as to be useless. It is always wise to be skeptical about the outputs of simulations, especially if you cannot see the source code for yourself.
And today, Michael wrote the following in
part two:
Up to this point, it was difficult to challenge the conclusions of AGW-believing climate scientists because most geeks don't have much expertise in climatology. We tend to consider ourselves scientists and to give other scientists in other areas of expertise the benefit of the doubt. Without a great deal of experience in climatology, it's hard for a geek to justify spending much time questioning the modes and methods of professional climate researchers.
However, the email leak has changed all this. Along with a hoard of emails, some source code for the computer climate models was also hacked and released to the public -- and the source code is an unusable mess. It doesn't take expertise in climatology to look at source code and determine that the code is garbage. There are many more geeks with software expertise than with climate expertise, and the geek community will go through every line of code and likely conclude that the computer models are so flawed that any conclusions drawn on them are without merit.
Despite the liberal tendencies of many geeks, I believe that the source code evidence will be insurmountable for most. Some will continue to cling to AGW because of a devotion to left-wing politics, but the majority of geeks will abandon their belief, and that abandonment by geeks will truly spell the end for AGW.
Hat tip to Paul Hsieh at
GeekPress for the link.
Posted by DaveH at November 28, 2009 3:43 PM