This paper will review progress being made on developing more defect selective forms of positron annihilation spectroscopy (PAS) at the University of Hong Kong. The first of these, positron deep-level transient spectroscopy (PDLTS). parallels conventional deep-level transient spectroscopy (DLTS) except that the positron is used as the probe, either to tell if the defects have vacancies attached to their microstructure (type I PDLTS) or as a simple electric field probe (type 11 PDLTS). It is shown the more important type I PDLTS has an intrinsic problem brought about by the high donor densities required to operate electrical trap filling. The problem-namely fast positron drift out of the active deep-level region into the reverse biased junction-is suggested as having two solutions. The first is to move to higher positron beam intensities and take spectra of 10 9 events. The second is that by using lower dopant densities (<10(15) cm(-3)) deep levels may be filled by inter-band optical excitation thus forming the workable technique positron optical (PO)-DLTS. Other techniques briefly considered in this paper are deconvoluted-coincidence Doppler broadening spectroscopy (CDBS) and Fourier transform (FT)-CDBS. Such are seen from a different perspective than most contemporary works, which tend to concentrate on the high momentum region. It is pointed out that the additional root-of-two improvement in hardware resolution and the factor of three improvement gained through deconvolution, can produce final effective resolutions similar to ACAR. Moreover, since in deconvolution, the natural space for regularized solutions is "real crystal space"-it is suggested that the autocorrelation function B-2gamma(r) be taken as the experimental CDBS data-not just because it provides easier visualization,-but because data in this form lies directly on the crystal lattice. (C) 2002 Elsevier Science B.V. All rights reserved.