Eric wrote:

Martin 53N 1W wrote:

Eric wrote:

Martin 53N 1W wrote:

[...]

how does boink make it easy for developers?

They only need worry about their project application. All the mundane stuff for connecting across the internet to a project server, and security issues, are all wrapped up in Boinc and taken care of for them.

The program that does the actual
processing is now more complex. Updating the seti program (for example) to
add new science tests is no more easy under boink than under classic and in
fact its probably more difficult as now developers have to not only develop
the science code but also keep up with changes in boink

The "Boinc" + "Projects Clients" system overall is more complex, but then also there is a lot of new features and new flexibility added.

Boinc includes the functionality of a proxy and SetiQueue of the old classic, plus a few other important DC bits. The s@h project client is now /simpler/ than before.

Developers can now concentrate on their project clients and have a fixed interface into Boinc. New Boinc developments are usually transparent to the project code. That is, the Boinc to project interfacing remains the same.

In the not too distant past, each 'update' of s@h classic has caused
problems and frustrations for all.

how so? all you had to do was download the new version - whats so difficult
with that?

Users didn't. Finally, an upgrade would be forced upon the users (to many screams and howls of protest) by kludging the servers to refuse old client results.

With Boinc, this can now be managed automatically so that you could have a different client specific to each WU. This is effectively already being done for some projects.

Nothing in the search for ET requires boink (vs Seti Classic).

No, but then also Boinc makes life a lot easier and more efficient for helping to find ET across multiple projects. Science moves on.

... Did they test it well enough before release?
nope - not even close.

Indeed. We /are/ the 'rapid development' LIVE test.

The thing is... somebody is making money on boink,
not sure who or how, but why else would you offer some of your own
(Seti's)

I'm not sure that any money is involved either.

As for 'giving away' the resource. Yes very definitely. This is all in the true nature of Science to have a worldwide cooperative effort to speed up /all/ developments.

Berkeley are indeed 'giving away' time and resource. They also get far more than that effort back in return from all the positive interest and help they get from others 'on the outside'. This is a good example of where the development system for "Free Open Source Software" works well.

Boinc is still getting new functionality added. There will be a few bugs and system problems yet. Worse still, Berkeley are having to juggle some badly overloaded servers.


[...]

So why allow your WU crunchers to be lured away to other projects?

For the good of all science. Berkeley have also been so successful that they've had an 'embarrassment of riches' for some time.

And that is another important part of Boinc. s@h will not have enough
data to keep everyone always busy.

Oh really? No i dont think thats true. They can always increase the workload
by adding new tests and shortly they are supposed to have the ability to
gather far more data in a much shorter time (or so i have heard).

It is silly to 'make work' for the sake of it. Their new client for s@h ("Enhanced Client?") includes more processing to double the sensitivity of the present search. To do more will give diminishing returns for the compute effort and would be an unnecessary waste.

As the users' computers increase in power yet further, Berkeley will oncemore run out of data for us to crunch. This may well happen in the short term when the classic users switch over. Hopefully, the new data sources will come online soon!


Happy crunchin',
Martin