S4 Search nearing completion - Einstein in Hibernation?

I agree it's not explicit, but as I read it, processing will go straight on into the S5 data they already have.

Failing that, maybe they could put up some old or test WUs and let us test an official optimised client?

It's not clear from that if he's saying mid summer for the s4-s5 transition, or that s5 won't be ready until mid summer regardless of what happens with s4. My first read was the latter, but I'll freely admit it's ambigious.

The interferometers have been collecting data since November '05, as they are now, admittedly with some problems. The issue is whether how much there is/will be of 'decent' data for signal analysis for us in the near future. The best type of data is 'triple lock' where all three LIGO's ( Hanford 2km, Hanford 4km, and Livingstone 4km ) are in optimal listening mode simultaneously. That co-incidence was about half of desired/designed goal in the period up to mid March. There is a poorly known noise source at low frequencies which seems to be related to time of day, 'traffic' if you like. They have on occasion had to reduce laser power to maintain lock but that reduces sensitivity though. My guess is that it is not clear to the project scientists right now as to what subset of the collected data is best to present to us for crunching.
Cheers, Mike.

I forgot to say that to a certain extent the S4 data analysis can be extended anyway. The central idea is that collected data is compared against expected signal profiles ( binary inspirals say ). This is done for certain time periods to sift and attempt to uncover patterns. The longer the period of analysis the better the chance of declaring the fainter disturbances. So you choose a data set, a signal template of interest, and then crunch - perhaps producing a probabilistic comment about a 'discovery'. The same soil can be tilled in many ways for results.
The current analysis is obviously looking for expected signals, but also can put upper bounds on various astronomical phenomena. In addition they have been testing the whole 'data pipeline' by injecting hardware signals ( actually 'jiggling' the detectors if you like ) and software injections ( sending out work units with 'constructed' data ) to see if the whole analysis system is accurate and robust. A calibration of the method really, to gain further understanding, refinements and confidence in this new adventure!
Cheers, Mike.