Restart required on the BBC Climate Change Experiment

[nfaull]

(superceded announcement: please refer to this URL with Myles Allen's message):

http://www.climateprediction.net/board/viewtopic.php?t=4759

[TC10284]

Not necessarily good news. But, similar to what you said, it does provide some useful information even if it may be in error.
You guys are doing great work! I am not disappointed in the slightest! Please keep up the great work and I will continue to contribute as much as possible!

[herme3]

I can't believe that this program wasn't completely tested before being released to thousands of people around the world. Anyway, as a software programmer I understand how difficult it can be to find all the errors in a program. When the new model arrives, I'll continue to participate in this program and I hope that everyone else will too.

[Dean Morrison]

Whoops!

- still that's science for you.....

I think it is very important to get the message out about what has happened - a friend of mine was looking forward to reaching the year 2000 - and on checking her machine today found that it had been reset. She couldn't find any explanation and was so hacked off she was going to pack it in. Fortunately I checked the boards and found the problem.

Any chance of sending out a mass e-mail before more people pull out???

[astroWX]

Any chance of sending out a mass e-mail before more people pull out???

That's a good way to get put on international "spammer" lists. It actually happened -- once, early on, in Classic CPDN days. Not good to have the main site on 'bad guy' lists.

[Heather T]

So does this mean everyone in the experiment will be starting from 1920 again. I got to 1942.

[grahamd]

Yes it does Heather. My machine has been down for 4 hours but not due to failings in the model. It has been a traumatic day but all I can say is persevere; it's a worthy cause.

[Elaine]

Would it have been too difficult to have proactively given users a warning that the models needed to be re-started? The experiment organisers managed to run a repeat line of text over the model screen telling users to 'keep going, the early years are very important.'

I'm sure it wouldn't have been too hard to write a single line of text alerting users to the re-set issue. Instead, the message boards are filling up with confused users with crashed models; 'unrecoverable errors' messages, and the board moderators are left to explain what's going on.

This re-set has been badly handled and I wouldn't be surprised if many people drop out of the experiment. It's bad enough that it takes so long and the program is prone to crashing. Lack of publicity on the total re-set will be the final straw for some.

[Heather T]

Shouldnt someone post some information on the experiment message board/helpdesk. I cant seem to find any information on there. It must be very confusing for people. I dont know how I found this forum.

[grahamd]

Hello and welcome to the forum Elaine. You are quite correct but it's not a perfect world even on Easter Sunday. There's a huge amount of information and back up help here on the forums. Hopefully, you will persevere, it's for a worthy cause.
Best wishes

[herme3]

How will this incident affect the reporting of the results of this experiment? Is the BBC still planning to announce some of the results in May? Is that even possible now?

[ds50]

This is also what I asked here. (at the bottom)...
It seems so (although I don't know. O.k., I couldn't see the BBC4 anyways, outside U.K.)

[Muglurmath]

I'm afraid to announce that we've recently discovered a major error in one of the files used by the climate model

All crunchers might be feeling bad that they have lost their work. As of now I think a very small percentage of people know that they have to restart.

Going through the posts up to now, everyone is taking the problem in a sportive manner.

Just to announce the fact must have been very painful and difficult. We understand, and we are with you.

Keep up the good work.

[TC10284]

I'm not necessarily shrugging this problem off. It shouldn't have really happened, I believe. But, I do understand how hard this type of thing can be to create and maintain (to a point). I'm surprised this hasn't happened with another project (to my knowledge). I still am in full support of this project and will continue to help with my CPU cycles.

But I do wonder, if I am currently processing a hadcm3l workunit, will I need to abort it and let it DL another or detach/reattach to the project?

[Les Bayliss]

TC10284 (and others)

You can tell by the model name:

hadcm3ln_ for BBC,
and hadcm3lbm_ for cpdn

Note the "n" and the "m" in the new models.

Your previous model will also have a failure of:
User abort: error 99