Dmitry, the WRF model is a freely available and may be downloaded at www.wrf-model.org. To check out BOINC, start here: http://boinc.berkeley.edu/ There is no development documentation available. Thanks.

Doubt very much that anything dictates that improvements changes, additions made privately to GPL / Open Source software are required to be made public. If though it were decided to make any additions, changes, improvements and manuals written for the purpose of operation were to be made available, no monies could be asked for that. That's as much as what i can gather from reading a little about GNU & GPL.

Any experts out there?

As with other software, such as AutoDock and Rosetta, we "grid-anable" the software but do not contribute our changes back to the organization that manages the code.

That's because for each project, such as the AfricanClimate@Home project, we are only working with the code on behalf of the researchers. World Community Grid has no interest in the code except for the fact that we are running it on for the researchers.

WRF was a special case and we did give back something to the folks who manage WRF. As you may know, we do a thorough security review of each piece of software that runs on World Community Grid. We sent sent an email to the folks at WRF to let them know about the security exposures that we found in WRF. We fixed these exposures in the grid-enableb version of WRF but we do not know if the WRF people made the changes.

The bottom line is that we do not share the grid-enabled version of any of the software that runs on World Community Grid.

Just a thought ... knock it back if it's not applicable:

Is there any possibility that your local weather forecasting efforts could be done as a (mini-)WCG project? That would save you having to reinvent the wheel. If you have crunching power at your disposal that you use for your forecasts, I think WCG would appreciate having some of its spare capacity online crunching other projects.

On the other hand, WCG projects to date have been aimed at extending scientific knowledge rather than providing a routine service. But you could probably extract some scientific knowledge from your project, e.g. by comparing the actual weather to your forecasts, analysing the differences, and perhaps helping to improve the modelling science & software.

But there could be a can of worms here. What would happen if lots of local weather forecasters wanted to use your BOINC/WCG-modified-WRF setup? It could become an admin nightmare for WCG if they were to run many of these efforts.

And could it be argued that since WCG states that all research results are public-domain, it should put all the source-code used to achieve the results, including the modified WRF code, into the public domain too, so that the methods used could (hypothetically) be verified by independent researchers?

Hi!

If results are made public, but not the software that generated them, then how can any independent researchers verify them? And, obviously, the documents about output data structure must be made public too.


Dmitrio