The second LIGO science-data taking (dubbed S2) ran from February 14 to April 14 ealier this year. During S2, all 3 LIGO detectors ran simultaneouslly for 312 hours. Moreover, both 4km-long detectors, operating at not too disparate sensitivities, logged over 400 hours of coincident running, giving us the first serious opportunity at applying a coherent search for gravitational-wave signals from inspiraling neutron star binaries. This talk will present the algorithm and the codes implemented in such a search. It will also discuss the lessons learnt from a coherent analysis of the `playground' data, which comprise a fraction of the complete S2 data set aside for tuning the inspiral detection pipeline for that run.