Ingleside, it is time you accepted that World Community Grid is not a cookie-cutter BOINC project, and that this is perfectly fine.

Or would you like to ask David Anderson whether WCG is a BOINC project or not? I'm sure WCG would be quite happy to remove the BOINC logo from their site and software - but would David Anderson be happy about this?

I think not.

BOINC suffers from changing protocols with every version. There is no stability, and no documentation. Saying that all BOINC projects must use the official server build is a complete failure - it is the Berkeley OPEN infrastructure for network computing.

You would do well to remember this. World Community Grid would lose very little if they dropped support for multiple BOINC projects altogether.

That's my point, Saenger. This group is surprisingly small.

Of course, World Community Grid don't want to lose anyone.

BOINC was chosen, not so that other projects could be run, but because the old client did not support Linux or Macs.

Saenger, you have it all wrong.

BOINC doesn't have standards. If they did have proper published standards, then World Community Grid would be happy to comply. But they don't! They just have a mish-mash of out-dated documentation, and a buggy reference version that is the only thing they can point to when asked what the "standard" is supposed to be.

If you don't want to help humanitarian projects, then you are welcome to leave.

I am strongly in favour of standards, Saenger. However, World Community Grid cannot impose standards on the rest of the BOINC community.

Unfortunately, in the absence of standards from Berkeley, and the total lack of community involvement in the development of BOINC features, anarchy reigns.

This is bad for BOINC, bad for crunchers, and bad for World Community Grid.

Let me put it another way: if we pretend for a moment that the official BOINC server is a de facto standard - in this situation, the BOINC server should be carefully versioned, features should never be changed in ways that break backward compatibility, and new features should receive full consultation with the entire BOINC community.

None of this is true. Features are added and removed arbitrarily, compatibility is not given high priority, and consultation is in name only.

This is why the official BOINC server build is not and can not be a de facto standard. And in the absence of a stable and complete paper standard, this means that none of BOINC is governed by any sort of standard at all. The best anyone can do is try to interoperate with the many versions of the BOINC software that have been released into the wild.

This means not making assumptions. Not assuming that CPIDs can be used as primary keys. Not assuming that a particular server supports a particular feature. I could go on, but it would get boring.

This is harder than you think. BOINC itself fails in many regression tests.

The poor hapless alpha testers have to deal with the same bugs reintroduced again and again. It's not a task I envy them.

When World Community Grid conflicts in some way with BOINC clients, the techs treat it as a bug. When something just happens not to be the way some stats sites have come to expect, they treat it as a bug in the stats sites. And in this case, they are quite right to do so. Willy is wrong, and using the CPID the way he did was a mistake.