First, a historical perspective:
What would a conference on ``cosmology'' have been like in earlier decades? In the first half of the century the agenda would have been almost solely theoretical. The standard models date from the 1920s and 1930s; the Hubble expansion was first claimed in 1929, but not until the 1950s was there any prospect of discriminating among the various models, Indeed, there was even then little quantitative data on how closely any isotropic homogeneous model fitted the actual universe.
A cosmology meeting in the 1950s would have focussed on the question: is there evidence for evolution, or is the universe in a steady state? Key protagonists on the theoretical side would have included Hoyle and Bondi. Ryle would have been arguing that counts of radio sources - objects so powerful that many lay beyond the range of optical telescopes - already offered evidence for cosmic evolution; and Sandage would have advocated the potential of the Mount Palomar 200 inch telescope for extending the Hubble diagram far enough to probe the deceleration. Intimations from radio counts that the universe was indeed evolving, were strengthened after 1963 by the redshift data on quasars.
The modern era of physical cosmology of course began in 1965, when the discovery of the microwave background brought the early ``fireball phase'' into the realm of empirical science, and the basic physics of the `hot big bang' was worked out. (The far earlier contributions by Gamow's associates, Alpher and Herman, continued, however, to be under-appreciated). There was also substantial theoretical work on anisotropic models, etc. Throughout the 1970s this evidence for the ``hot big bang'' firmed up, as did the data on light elements, and their interpretation as primordial relics.
Theoretical advances in the 1980s gave momentum to the study of the ultra-early universe, and fostered the ``particle physics connection'': the sociological mix of cosmologists changed. There was intense discussion of inflationary models, non-baryonic matter, and so forth.
Here we are in the late 1990s with a still larger and more diverse community. The pace of cosmology has never been faster. We're witnessing a crescendo of discoveries that promises to continue into the next millennium. This is because of a confluence of developments:
1. The microwave background fluctuations: these are now being probed with enough sensitivity to provide crucial tests of inflation and discrimination among different models.
2. The high-redshift universe: the Hubble Space Telescope (HST) has fulfilled its potential; two Keck Telescopes have come on line, along with the first VLT telescope, Subaru, and the first Gemini telescope. These have opened up the study of ``ordinary'' galaxies right back to large redshifts, and to epochs when they were newly formed. In the coming year, three new X-ray telescopes will offer higher resolution and higher sensitivity for the study of distant galaxies and clusters.
3. Large scale clustering and dynamics: big surveys currently in progress are leading to far larger and more reliable samples, from which we will be able to infer the quantitative details of how galaxies of different types are clustered, and how their motions deviate from Hubble flow. Simultaneously with this progress, there have been dramatic advances in computer simulations. These now incorporate realistic gas dynamics as well as gravity.
4. Developments in fundamental physics offer important new speculative insights, which will certainly figure prominently in our discussions of the ultra-early universe.
It is something of a coincidence - of technology and funding - that the impetus on all these fronts has been more or less concurrent.
Max Planck claimed that theories are never abandoned until their proponents are all dead: that's too cynical, even in cosmology! Some debates have been settled; some earlier issues are no longer controversial; some of us change our minds (quite frequently, sometimes). And as the consensus advances, new questions which couldn't even have been posed in earlier decades are now being debated. This conference's agenda is therefore a ``snapshot'' of evolving knowledge, opinion and speculation.
Consider the following set of assertions - a typical utterance
of the r-m-s cosmologist whom you might encounter on a Cambridge
street.
Our universe is expanding
from a hot big bang
in which the light elements were synthesised.
There was a period of inflation,
which led to a ``flat'' universe today.
Structure was ``seeded'' by gaussian irregularities,
which are the relics of quantum fluctuations,
and the large-scale dynamics is dominated by ``cold'' dark matter.
but 
(or quintessence) is
dynamically dominant.
I've written it like nine lines of ``free verse'' to highlight how some claims are now quite firm, but others are fragile and tentative. Line one is now a quite ancient belief: it would have rated 99 percent confidence for several decades. Line two represents more recent history, but would now be believed almost equally strongly. So, probably, would line three, owing to the improved observations of abundances, together with refinements in the theory of cosmic nucleosynthesis. The concept of inflation is now 20 years old; most cosmologists suspect that it is was indeed a crucial formative process in the ultra-early universe, and this conference testifies to the intense and sophisticated theorising that it still stimulates.
Lower down in my list of statements, the confidence level drops below 50 percent. The ``stock'' in some items - for instance CDM models, which have had several ``deaths'' and as many ``resurrections'' - is volatile, fluctuating year by year! The most spectacular ``growth stock'' now (but for how long?) is the cosmological constant, lambda.