18 May
2008
18 May
'08
9:23 a.m.
Warren wrote:
I'm going to take a "devil's
advocate" stance here, and argue to keep the
existing "as" binary untouched.
As first priority, certainly. We want to preserve original
data for posterity.
As second priority, there may be several possibilities.
A best-guess reconstruction of an approximation of V1
is one such additional possibility. Interpolation based
on noisy data sources is a time honored practice -- as
long as one carefully differentiates the results from
pristinely clean data.
I don't see one kind of reconstruction as inherently better
or worse than another kind; it depends on what the goal is.
I think Apout, for instance, is a *really* nice variation
on the usual approach of a closed-world simulation.
And certainly there are sharp limits on what we have available
right *now*, but I would be quite cautious about saying "never".
Additional data sources may become available; who knows.
Doug
--
Professional Wild-eyed Visionary Member, Crusaders for a Better Tomorrow