The Climategate code

Laatste wijziging: zondag 20 december 2009 om 17:46, 2106 keer bekeken Print dit artikel Bekijk alle nieuws feeds van onze site
zondag 20 december 2009

In addition to the more than 1,000 emails and assorted documents that were leaked to the public a month ago, there were several files containing software code, seemingly source code for the calculations the Climate Research Unit of East Anglia used for processing and adjusting temperatures for their calculations of global warming.

Immediately after the leak, attention was drawn to comments in the code's documentation that seemed to indicate that the programmers had difficulty understanding what had been done previously, did not feel capable of righting the programming wrongs, and on occasion invented weather stations to park data or even invented data.

Now, as computer scientists and programmers have had more time to analyse the code as well as the comments, programming errors that would change calculations or even cause the program to skip data are coming to light.

John Graham-Cumming is one of the software engineers doing this analysis, and he has found at least five errors in the code. When interviewed by the BBC two weeks ago, Graham-Cumming specifically said he wouldn't trust it. (He's not the only one--Chief Io has been working on it as well.)

However, Real Climate, defender of the consensus and home to some of the scientists involved in the leak, have said that the code in the leaks was not the code actually used by the CRU. If this is true, then a lot of people trying to replicate the work of the climate scientists are wasting their time.

It is certainly not too early for CRU to release the actual code for inspection. Some would say it is long past time. As Real Climate is currently demanding code and data from a skeptical scientist so they can replicate his work, a bit of reciprocity would not be unwelcome.

Real Climate is obviously not CRU. But many of its long-time contributors are co-authors with CRU scientists on climate science papers. And the leaked emails show a tight link between Real Climate and CRU, with Real Climate volunteering to serve as a house organ for CRU statements, and offering to act as a buffer between CRU and skeptics. They also have felt free to pronounce authoritatively that the code released is not the same as the code used. So the request is not absurd.

As Steve Mosher fequently says, 'Free the data! Free the code!' To which I can only add, 'And your mind will follow...'

Bron: examiner

Voeg toe aan: