I was thinking about Carl\'s intention to provide shared memory access under boinc, the same as in cpdn classic but it occurs that cpdn doing its own thing is a bit against the point of boinc. Instead how about the following additions to a generic shared memory dll from boinc itself which could interface with the projects through a common api. Then people could easily create their own visualisations for any boinc project.

1. GetProjects(&list) - returns a list of all projects the user is currently running, with an id and type for each. Eg:

1 CPDN
2 CPDN
3 SETI
4 PREDICTOR

2. GetSharedData(ProjectID,&datalist) - returns a list of the data available from the project. Eg. For CPDN:

1 Timestep
2 Phase
3 Model Date&Time
4 Temperature
5 Precipitation
6 Pressure
7 Clouds
8 Snow&ice
9 Wind (pleeese)

3. GetDataDimensions(DataID, &x,&y,&z,&t) - returns the x,y,z dimensions of the dataset and the data type such as string, integer, double. Also the units to enable sensible display.

Eg for clouds: 73,96,3,float,%
Eg. For temp: 73,96,1,float,K

4. GetData(DataID,&mem) - fills the user allocated memory (mem) with the requested data

Using this, the user can ask boinc what data is available for which projects and request that boinc gets the data from the project and populates some user allocated memory with it. That\'s all that is needed to create our own visualisations.

Ideally this would all be implemented in such a way that a generic visualisation tool could immediately display a new type of project without recoding, effectively removing the need for new boinc projects to spend any time developing their own visualisations. Boinc has standardised the method of distributing workunits but why stop there?
<a href="http://www.users.globalnet.co.uk/~sykesm/cpdn.html"><img src="http://www.users.globalnet.co.uk/~sykesm/gfx/sig.jpg"></a>