Re: Run8 analysis planning

[Tom Banks <tbanks@socrates.berkeley.edu>]

> There is a problem with next weeks telconf, because Steve will leave to
> Puerto Rico and I will be in a plane. I will talk with Steve when we
> would have a chance to include him in a discussion. If we would
> have a complete plan for the UCB analysis now, we could run in
> during the next 2 weeks and have results when Steve comes back.
> But I don't think that's realistic. In the end, we might be able
> to make studies, but -in my opininon- there is no way to avoid
> another round of UCB processing through the skimmed data, to assure
> that your code really gets the same answers. This is essential
> and it is hard to find a substitute for that in a modified UCB
> analysis. It would compromise the truly important independence
> of your analyses, which has helped us many times on the long
> path to the light.

I was not suggesting that UCB avoid any future analysis passes over the
skimmed data.  My point is, given the wide range of lifetime values
observed by Steve and me, it seems quite possible that the lifetime is
sensitive to the ePC data processing parameters.  In that case, it is
difficult for Berkeley to "map out" the ePC/lifetime relationship because
of our long turnaround time on merlin.  We will, of course, be making
several more passes on the merlin cluster in the coming months, but I
worry about whether we'll be able to establish a convergence.  If we get a
different result each time, what does that tell us?

Tom