Many important points emerged during the meeting and several of these
were emphasised in the concluding discussion. Among these are the
following.
On the practice of H0 determination:
Hipparcos parallax work on a series of star clusters indicates that
main sequences can differ by up to 0.4 mag. Main-sequence fitting
itself does not necessarily figure in the distance-ladder method, but
the differences do indicate a worrying lack of understanding of
main-sequence luminosities - something we all thought was tied down.
One has to keep in mind that the number of Cepheid variables with
parallax distances is critically small.
There is the question of whether enough parameters are being used to
describe the correlations. For example, in Tully-Fisher relations,
would one be better off characterising the line-shape with more than
just one width? There is certainly point in introducing more
parameters if there is an underlying physical model (which is always
worth striving to achieve), and Bayesian analysis will tell you whether
the reduction in scatter is worth the extra parameter.
Ever present is the risk that something that doesn't fit the
correlation - or expectation - will be classified as something else
and thus excluded. A related point is that if, having done a programme of
work, one gets an extreme or unpalatable value of H0, one should
still report it to avoid introducing biases.
On the present position on the value of H0:
The distance-ladder community seems to be arguing about 50-65 km s-1
Mpc-1, rather about than factors of two. But it is certainly still
arguing: "we have moved away from being bimodal" claimed one
participant, yet "the bimodality has narrowed" was how another put it.
Looked at from another perspective, it is both remarkable and
reassuring that the physical methods (Sunyaev-Zel'dovich plus X-ray
and gravitational-lens time delay) and the distance ladder methods
agree at all, given they have such different bases.
And on the point of it all:
What you get from primordial microwave background anisotropy and
structure formation depends on total density - and on H0. What you get from
nucleosynthesis depends on the baryon density - and on H0 too. So to
find out what went on and what's going on, you need H0 - certainly to
a few percent. So we'd better make the effort...