L’Ombre de l’Olivier: The Shadow of the Olive Tree

07 December 2009

“Harry” gets a rewrite

The most important news on the CRU leak (aka Climategate) over the weekend was that the UK Met Office is not just opening up its data it is also checking the processing and rewriting the code.

I strongly suspect that this only happened because of the leak and that all the discussions about “HARRY_READ_ME” were influential in this decision. It will be very interesting to observe just how the process will develop and whether they take on board the suggestions of many of us who’ve looked at the code about the choice of a database or of language use.

The really good news is that we ought to be able to build some kind of third party hooks onto the database so as to identify the problem stations and so on. The one interesting thing that I note is that we are not, as far as I can tell, seeing a release of the sea data – only the land data. I’m going to email them to verify whether this is the case or not because it would seem to be quite important regarding the global temperature.

via di2.nu

While I don’t agree with this “Englishman” on all points, I have to agree that the CRU information that the leaker/hacker published is pretty grim when it comes to reproducibility. You pretty much have to have the author present to ensure you’re using the same info as published. Some help from the Computer Science field would seem in order 🙂

Posted via web from TweetingDonal’s Temporary Insanities

Advertisements

2 thoughts on “L’Ombre de l’Olivier: The Shadow of the Olive Tree

  1. Yep, he is… problem I have with all this is that we’re not looking at source, just what some hacker or leaker decided to release. Hell, I’ve got 3,000 emails in my inbox for the last few months, there’s certain to be more that 1,000 emails in the mix in the last 13 years. And I’m pretty sure that’s not all the code or data… got a copy of what was posted and it seems extremely incomplete.. I’ve seen more referred to just in the pubs at Nature that aren’t in this pile of …. bits that got released.

    Also, having done simulations and modeling of real world situations, stray code found laying about means nearly nothing. We need to see what was used to produce reports and charts, not whatever someone thought looked bad. Not sure we’re ever really going to find out without the active help of the people who did the reports… but from a computer scientist point of view, that code is a huge mess.

    I’m hoping we get a reasonable analysis out of the Auld Scot they appointed to review this mess, especially since he’s not part of the Climate Modeling community… we can only hope we get a clear answer…

Comments are closed.