Some thirty years after the discovery of GRBs a generic GRB model is beginning to emerge. The observations of isotropy, peak flux distribution and time dilation indicated that GRBs are cosmological. The measurement of a redshift provided a final confirmation for this idea. All cosmological models are based on the fireball mode. The discovery of the afterglow confirmed this general scheme.
The Fireball-Internal-External Shocks model seems to have the necessary ingredients to explain the observations. Relativistic motion, which is the key component of all Fireball Model provided the solution for the compactness problem. The existence of such motion was confirmed by the radio afterglow observations in GRB970508. Energy conversion via internal shocks can produce the observed highly variable light curves while the external shock model agrees, at least qualitatively, with afterglow observations. There are some indications that the two kind of shocks might combine and operate within the GRB itself producing two different components of the signal. This fireball model has some fascinating immediate implications on accompanying UCHER and high energy neutrino signals. An observations of these phenomena in coincidence with a GRB could provide a final confirmation of this model.
In spite of this progress we are still far from a complete solution. There are many open questions that has to be resolved. Within the internal-external shocks model there is a nagging efficiency problem in conversion of the initial kinetic energy to the observed radiation. If the overall efficiency is too low the initial energy required might be larger than 1053 ergs and it is difficult to imagine a source that could provide so much energy. Beaming might provide a solution to this energy crisis. However, so far there is no indication for the corresponding break in the afterglow light curve, which is essential in any relativistic beaming model when ~ -1. These last two facts might be consistent if GRB970508 took place within a very low density ISM - an issue that should be explored further. Afterglow observations agree qualitatively, but not quantitatively with the model. Better observations and more detailed theoretical modeling are needed. Another nagging open question is what determines the appearance of afterglow. Why there was no X-ray afterglow in the very strong 970111? Why was optical afterglow observed in GRB970228 and in GRB970508 but it was not seen in others (in particular in GRB970828 )? Finally we turn to the GRB itself and wonder why is the observed radiation always in the soft -ray band? Is there an observational bias? Are there other bursts that are not observed by current instruments? If there are none and we do observer all or most of the bursts why is the emitted radiation always in the soft gamma-ray range? Why it is insensitive to a likely variability in the Lorentz factors of the relativistic flow and to variability of other parameters in the model.
While there are many open questions concerning the fireball and the radiation emitting regions the first and foremost open question concerning GRBs is what are the inner engines that power GRBs? In spite of all the recent progress we still don't know what produces GRBs. My personal impression is that binary neutron mergers are the best candidates. But other models that are based on the formation of a compact object and release a significant amount of its binding energy on a short time scale are also viable. A nagging question in all these models is what produces the the "observed" ultra-relativistic flow? How are ~ 10-5 M of baryons accelerated to an ultra-relativistic velocity with ~ 100 or larger? Why is the baryonic load so low? Why isn't it lower? There is no simple model for that. An ingenious theoretical idea is clearly needed here.
However, I believe that theoretical reasoning won't be enough and only observations can provide a final resolution of the questions what is are the sources of GRBs? The binary neutron star merger model has one specific observational prediction: A coincidence between a (near by and therefore strong) GRB and a characteristic gravitational radiation signal. Luckily these events have a unique gravitational radiation signature. The detection of these gravitational radiation events is the prime target of three gravitational radiation detector that are being built now. Hopefully they will become operational within the next decade and their observations might confirm or rule out this model. Such predictions, of an independently observed phenomena are clearly needed for all other competing models.
GRBs seem to be the most relativistic phenomenon discovered so far. They involve a macroscopic relativistic motion not found elsewhere before. As cosmological objects they display numerous relativistic cosmological phenomena. According the the NS2M model they are associated with the best sources for gravitational radiation emission and more than that they signal, though in directly, the formation of a new black hole.
I thank E. Cohen, J. Granot, J. I. Katz, S. Kobayashi, R. Narayan, and R. Sari for many helpful discussions and D. Band and G. Blumenthal for helpful remarks. This work was supported by the US-Israel BSF grant 95-328 and by NASA grant NAG5-3516.