Talk:Climateprediction.net

From Wikipedia, the free encyclopedia

If someone wants to upload the pictures, they can be found on the BOINC Wiki BOINC Wiki CPDN AV page

---

I think the 2 most important pieces of information that an encyclopia should say about climateprediction.net is that it is modeling climate and it is inviting people to join in and help by running the software.

Wikiacc has twice removed the fact that the project is inviting people to run the software.

The first time it said Anyone with a reasonable PC is encouraged to join by visiting the Climateprediction.net website

Wikiacc may have been correct to remove this under NPOV. I tried to compromise and put in a replacement that said:

Contents 1 Invitation 2 Bigger than the earth sim? 3 Any suggestions for this page? 4 "generated more data than any other climate modelling program"

[edit] Invitation

The project is inviting anyone with a reasonable PC to join this scientific project to improve our knowledge of climate change uncertainties by visiting the Climateprediction.net website to learn more and to download the software to run your own model.

I felt this remained true to NPOV by making it clear that climateprediction.net was doing the inviting. Wikiacc has removed this. I am putting a close replica back in. If you still have issues with it, please explain (or compromise) rather than just removing it.

crandles 19:40 30 July 2005

Put the link into External link section. As it is now it brings low value to the article. (Btw, a more details on /background/ of the project would be useful - what results it brought so fat, what methos it uses etc.) Pavel Vozenilek 23:17, 30 July 2005 (UTC)

[edit] Bigger than the earth sim?

Where does "its bigger than the earth sim" come from? Seems a bit dodgy to me William M. Connolley 19:09:14, 2005-07-31 (UTC)

I took it from a http://www.climateprediction.net/science/pubs/nerc2005allen.pdf which includes a potted history. This and the Climateprediction.net open day were the major sources of the material for the history section. As it was for the NERC council I assume it should be pretty reliable.

Trusting peoples PR? :-))) - William M. Connolley 09:08:36, 2005-08-01 (UTC).

Oops I didn't actually add that to the page but it appears to be verified in that pdf anyway.

As further evidence there is this paper which has been accepted for the 1st IEEE International Conference on e-Science and Grid Computing in Melbourne, Australia this December. 30,000 computers = 2 earth simulators. Model years per host has stayed about the same because later computers are faster but typically shared between more BOINC projects. The first week saw 27263 new users the second week saw only ~3800 and given the publicity on the first day, 15,000 in first day does not look unreasonable to me. I certainly don't see enough evidence to accuse them of lying. crandles 19:33, 28 September 2005 (UTC)

OK OK, I did put a smiley in... I haven't accused anyone of lying (have I? I didn't mean to, only to query PR). William M. Connolley 20:28, 28 September 2005 (UTC).

OK "Seems a bit dodgy" is not really an accusation of lying and taking it as raising a query would be much more appropriate. However, I didn't say you did accuse anyone of lying only that I didn't see enough evidence for anyone to do so. crandles 22:12, 28 September 2005 (UTC)

[edit] Any suggestions for this page?

I am sure this page is not up to featured article status. Any suggestions as to what is needed? Do we, for example, want to include some of this?

I guess better referencing for facts should be a priority.

Anything else? Thanks crandles 21:13, 17 February 2006 (UTC)

Being thrown straight into "the experiements" won't work for people who don't know it. A section on what-is-it-all-about really ought to be upfront.

Also, there should probably be something on the disadvantages of the approach - the limited return bandwidth means that the verification of the models for plausible climatology is very weak (as I understand it). William M. Connolley 21:41, 17 February 2006 (UTC)

In the section where the aims of Climateprediction.net is explained, this article repeatedly refers to 'us' and 'we', almost as if lifted directly from the website. That really needs rectifying to eliminate bias.

JamesOwen 22:34, 4 March 2006 (UTC)

Yes a valid point. It was all lifted from the wiki (which I mainly wrote) for participants and some is lifted from the website. I am doing this with the project thanks for my efforts and therefore do not think there are issues with what has been copied from the website. Here it is aimed at a different audience and the 'us', 'we' and 'our's' needed changing and I have changed some (though there could be others that I missed). crandles 16:36, 5 March 2006 (UTC)

Something on this message from Myles Allen needs to be added. Another discussion on the topic is titled "Restart required on the BBC Climate Change Experiment". --Spiffy sperry 19:35, 21 April 2006 (UTC)

I have done a couple of updates: launch of SAP and the restart. crandles 22:06, 22 April 2006 (UTC)

The "aims" bit seems somewhat repetetive. I've tweaked it. I removed This will allow the project to explore how climate may change in the next century under a wide range of different scenarios because I don't think its right - the focus is on model, not scenario, uncertainty. William M. Connolley 09:52, 23 April 2006 (UTC)

[edit] "generated more data than any other climate modelling program"

Is "generated more data than any other climate modelling program" true? I would have thought that due to the return bandwidth, this is unlikely. CPDN returns far less data per model year than most GCM sims archive William M. Connolley 08:26, 23 April 2006 (UTC)

Hmm surely it depends what you mean by 'generated'. If it is generated and saved even if only for a short period then I would think yes it is true. If it has to be permanently stored in the most useful location then that is a tough test but let us see. 7 MB uploaded compared to 330MB in compressed format is approximately one 50th of the data centrally stored. Ensemble size 157218 slab models compared to what? QUMP's ensemble size of 128 is the largest I have heard of but there could be larger ones around and there could be models that generate and store more than 20 times the data of a slab model. So it may be dubious if you add central storage requirements. The sentence doesn't seem to add storage requirements at all though. crandles 19:23, 23 April 2006 (UTC)

I think "generated" really ought to mean "stored and potentially used". If each run only uploads 7M then thats fairly small. My HadCM3 runs generate about 7M per month run - so at 6M/d on 10 boxes every day I produce 60* one CPDN net. Over 1000 days thats 60,000, and thats just my little cluster. I'm sure the Hadley Centre has 100* that. William M. Connolley 19:44, 23 April 2006 (UTC)

But what does "stored and potentially used" mean? For the slab model participants were asked to keep the data if possible and some participants have been asked to upload further data. If half have kept them then there is 128517*0.5*330 MB or over 21Tb. You are saying you have 420Gb stored from the last 1000 days. I don't doubt that (I'll assume it is compressed), but is it from a single climate modelling program? Hmm. I have interpreted 'program' as 'project'. The context makes 'program' ambiguous and should be changed even if the claim is defensible. crandles 11:01, 24 April 2006 (UTC)

I deleted a 'however' as it didn't seem appropriate after I had made some other changes. Hope this is OK. I wondered about removing the rest of the sentence 'The higher climate sensitivities have been challenged as implausible'. While it is accurate and relevant it isn't really a 'result to date' and it also seems less necessary in view of other changes I made. crandles 19:31, 23 April 2006 (UTC)

I think it should stay; its a comment on the results-to-date. I think it should be noted that the most publically visit result - some v high sensitivities - is doubted William M. Connolley 19:44, 23 April 2006 (UTC)

OK I'll leave it. (You did see the addition that 'The really high sensitivities are being found to be unlikely.'?) crandles 11:01, 24 April 2006 (UTC)