Johan Plane wrote:
[...]

Anyway: Yes, I like the project as such, otherwise I wouldn't have been a participant since
1999. I am however more interested in the science behind it and I was lead to beleive that
S@H was a legitimate science project under a well reputed US university. This thread has,
however, made me doubt that this is the case, and your reply doesn't give the straight
answer I would have expected.

Reading carefully, there is nothing in this thread saying that the team at Berkeley or s@h are not "legitimate" in any way. Please check their website and try a little Googling to check their credentials and their science.


[...]

you can't run a local serverpark with local centralized caching of workunits as you could
with the old software and an addon, like SetiQ. However, the workaround is to set the

You can run all your connections through a web proxy if you wish. That way, your farm of machines can connect via a secured proxy on just one server connected to the internet. Not quite the same as SetiQ but just as secure.

For the really paranoid, you can set your firewall to only accept connections to and from Berkeley. This is very easily done in Linux for example.

connect intervals to 10 days, fill up the cache and then revert to 1 day. Also I dislike
that each of my computers have to have direct contact with the internet to upload the

At present, s@h have a 14-day deadline for returning results. (Any longer and the screams and howls and sheer pain about 'pending credit' would become unbearable!!!) If you have a 10-day cache, that only gives you 4 days or less to get the results uploaded if Berkeley or yourself are having connectivity problems.

Hence, the optimal max cache should be 7-days.

Better still is to use a small cache (_1_ day or less for fast turn-around time for the project) and attach to a second project to share or as a 'backup project'. For example, I run with 0.5 days for the cache which appears to be optimal to minimise the overheads on the Berkeley servers.

results, as opposed to the SetiQ system. One can only pray that there will be a solution to
this, however I doubt it due to the complexity of the way the client communicates with the
S@H servers - the communication protocol as a whole seems to be unneccessary complicated.

This is all open source so some sort of Boinc-dedicated proxy could be written by anyone interested.

However, the lack of a SetiQ should not be a problem. The caching is just done differently so as to make cheating harder, and there are "add-ons" to let you centrally log your stats and results. As mentioned, you can also use a web-proxy if you wish.

The protocols are built on top of http/https and a readily available program library called libcurl. This follows existing 'standards' and avoids a lot of tcp/ip rework!

However, the software seems to be doing what it's supposed to and I haven't had any of the
problems that others seem to have had, so I'm content with it, which, as mentioned above,
doesn't necessarily mean I like it.

Other than SetiQ and 'the look', what is it you would like to see (instead) and how?

Or is it just the upheaval of changing from the SetiQ way of working? (Yes, I used to use that very good program.)

BTW It was Wayne Brown that commented on the software, not me!

??? I must have missed or confused one of the previous posts, sorry.

I agree with you that it's sad when the US Congress stops funding for political reasons,
however the way university science projects are funded in the US is alien to me since, here

It can be just as complicated here in the UK also!

Hence my 'non-direct' answer, because I simply don't know the exact detail. However, from what I've seen of s@h, they are running as good an experiment for this science as is possible for this scenario. It is also science that should survive peer review well.

The USA does not seem in reality to have the ideal freedoms that are claimed... Unfortunately, the UK appears to be fast slipping down a similar slippery slope into restrictions as opposed to freedoms... A balance is needed but at what costs?


Keep searchin',
Martin