Better computing power meant more particles = more accuracy. Gas and gas physics (hydrodynamics) were introduced to the simulation in order to follow the effects of star formation. Faster computers also meant that not a single astronomical meeting was to go by where you didn't see a tidal feature movie and there were numerous papers discussing the uses of galaxy interactions as probes of dark matter.
It is also possible that for over a range of impact parameters and velocities the two galaxies can become bound and merge. However,
``Collision of two cars doesn't make a new kind of car'' - Unknown astronomerso if mergers are common why is there no signature in the galaxy population. Or maybe the whole Hubble sequence is the consequence of mergers at early epochs?
Soon the theoretical work outpaced the observational work (always a dangerous time where imagination runs unbridled) since the simulations with sufficient number of particles could predict velocity fields and regions of star formation.
Observations Strike Back (1990's):
The limitation on observations with respect to interacting galaxies is that velocity information is obtained in the form of slit spectroscopy. Only a single slice of the pair can be observed at a time, and in a straight line whereas most tidal features are curved.
This continued until the development of superior Fabry-Perot instruments on optical telescopes. Now there can be a detailed comparison of not only the structure of an interaction, but also the velocity field and the regions of star formation (traced by the H-alpha line feature from ionized gas by young, hot stars). This led to an understanding of bright IR galaxies discovered in the early 1980's as merger products.