Lispian Random meanderings on whatever catches my fancy


I’m reading the to-and-fro and what’s shocking from the AGW proponent camp is this fixation that “there is no conspiracy”. That’s not the point of what seems to be evident in the emails and source code. What seems to be the problem is that science was corrupted. Science requires openness to both the theories and the data. If new data comes to light that contradicts a given theory then the theory must be modified.

Same holds for any given model. If data suddenly shows the model to be faulty, the model needs to be reworked. That’s just good science.

A lot of the noise going back and forth can be resolved if both sides would just come the realization and agreement that what we need is a neutral body to simply store all the data, all the algorithms, all the models, etc. That way no one is ever going to be pestered with access to information requests and anyone who wishes to use the data and software can and if they feel there’s something worthy of publication it can be so presented.I do note some on both sides of the argument are already espousing a fully open structure. This is commendable since a hallmark of science is skepticism. In fact, a scientist must be most skeptical about his or her own work. If ever a bias is to sneak into someone’s work, it’ll be one’s own.

Besides, science that cannot be independently verified is not science. It’s dogma. And the sooner the scientists return to science the better it will be for all of us.

And please, find some computer scientists — good ones — to create models that are decent. Computer scientists judge people’s skill and the quality of the code by looking at the source. From the snippets I’ve seen the code is horrid and requires a serious rework by someone who actually is trained to work properly with computers — like a computer scientist. It’s much easier to make statements about the validity of a model if the model as coded is actually well done.  Poorly done code is not maintainable, not extensible, difficult to understand, difficult to prove that it works even remotely correctly, and is a long term liability. So, fix up the code guys. It’s not like we’re short of CS-types in the Western world.

Comments are closed.

December 2009
« Nov   Jan »