Buy

Books
Click images for more details

Support

 

Twitter
Recent comments
Recent posts
Currently discussing
Links

A few sites I've stumbled across recently....

Powered by Squarespace
« Ciccerone circumspect | Main | Purdue climate confab »
Thursday
Nov112010

CRUTEM code still not fixed

Readers may remember that back at the start of the year, John Graham-Cumming found some errors in the CRUTEM code. He very helpfully notified the Met Office, who subsequently confirmed JG-C's findings.

The odd thing is that, according to JG-C's latest posting, the Met Office haven't actually managed to get round to fixing their live version of the code yet.

PrintView Printer Friendly Version

Reader Comments (15)

Arrogance or stupidity , take your pick.

Nov 11, 2010 at 8:38 PM | Unregistered CommenterKnR

KnR,
Maybe neither one. Maybe they don't care.

Nov 11, 2010 at 8:49 PM | Unregistered Commenterj ferguson

Because previous errors (when corrected) had no effect on the desired outcome, it would hardly seem worth the while to follow up on any new ones.
Humpy would be proud of their studious inactivity.

Nov 11, 2010 at 9:01 PM | Unregistered CommenterTony Hansen

They should load the code and the data on to sourceforge, go home, collect their pensions, and watch the quality of 'their' work improve on a daily basis.

Nov 11, 2010 at 9:12 PM | Unregistered CommenterZT

Maybe they just do not know how to fix the error.

Nov 11, 2010 at 10:17 PM | Unregistered Commentergolf charley

The most likely explanation is that in prioritizing work done at the Hadley Centre the folks there feel this is low on the list. They indicated in their mail to me that the error was small and so I suspect they feel they've got other, more important issues to work on.

Still a pity it hasn't been corrected.

Nov 11, 2010 at 11:25 PM | Unregistered CommenterJohn Graham-Cumming

Maybe they will get the forecast for next week correct. That would be nice. But even that is unlikely.

Nov 11, 2010 at 11:36 PM | Unregistered CommenterDon Pablo de la Sierra

A question to John G-C, how long would it take you to fix and test this error? Are we talking weeks, days or hours? I'm guessing hours considering you understood it back in Feb and have test programs at the ready.

Nov 12, 2010 at 1:14 AM | Unregistered Commenterrc

Source: http://www.quotegarden.com/procrastination.html

Procrastination is like masturbation. At first it feels good, but in the end you're only screwing yourself. ~Author unknown, possibly from Monty Python?

Nov 12, 2010 at 1:55 AM | Unregistered CommenterIanB

Dilbert on bad data: http://www.dilbert.com/strips/comic/2010-11-07/

Nov 12, 2010 at 6:59 AM | Unregistered CommenterAntonyIndia

If it is a medium to low priority error and they employ a "major" release strategy then this is not so unusual. It could be stacked up somewhere.

I would be surprised for this sort of error fix to be pushed through to the production release as an individual item, that would only normally happen with critical error fixes.

I'd be interested to know what are their source management procedures... anyone know how they work?

However I learnt that if there are more visible errors to the world, a comma missing on a piece of text, you get more kudos for fixing that quickly than perhaps more critical errors.

If I was the Met office I would have fixed it, but I can understand the reasons for its current state.

Nov 12, 2010 at 7:07 AM | Unregistered CommenterJiminy Cricket

@rc I would imagine (not having seen the code, or other systems that might depend on it) that it would be a matter of an hour to write some test cases and then fix the code.

@Jiminy Agreed. I suspect this is just too low priority for the Met Office folks to push out a fix quickly. Pity. In terms of understanding the Met Office's software development practices I'd recommend: "Engineering the Software for Understanding Climate Change", Easterbrook, Johns.

http://www.cs.toronto.edu/~sme/papers/2008/Easterbrook-Johns-2008.pdf

Nov 12, 2010 at 9:21 AM | Unregistered CommenterJohn Graham-Cumming

Harried Harry is heard in the background: "COMMENTS, you dumb *$&%^#s! Give me some &#&$^%# COMMENTS!..."

Nov 12, 2010 at 5:07 PM | Unregistered Commentermojo

Isn't this an issue for Quality Assurance or something, knowingly allowing mistakes to remain?

Nov 13, 2010 at 8:36 PM | Unregistered Commentergolf charley

@John Graham-Cumming

Thanks for that PDF link and VERY illuminating document. I have only had time for a once read through but...

Throughout the document the word "simulation" is used... and that I am sure is the word the use to outside world and to the people who pay the bills. But what that document describes is NOT simulation. Not even close.

I better word would be "projection". Their work is just projection.

Ignoring the need for paper's like this to tart up their findings with pseudo-scientific jargon, what the Hadley Centre does is not rocket science or even close it.

There are a million lines of code "organically" grown over 20 years, and it has become its own reality.

It is like building the world's biggest skyscraper (one million lines high) in a bog without good foundations.

I am not sure non-IT Engineers would fully understand that document, but I would recommend some of our IT contributors to read it.

Projection has become simulation has become reality... bloody amazing

Nov 14, 2010 at 11:39 AM | Unregistered CommenterJiminy Cricket

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Post:
 
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>