I think one must differentiate (at least in biological sciences)
between data supporting a finished and published research project.  I am
a systematic entomologist and biographer, and like most of my colleagues
require several years to complete complex monographs--sometimes just 3
years, sometimes a decade.  Right now many systematists have primary
research data written on paper and representing tentative ideas or data
acquired by long and tedious observations--raw data not yet interpreted.  I
certainly would NEVER expect a colleague to be obliged to show me their
unpublished data.  This is a simple matter of ethics--I have no right to
examine data for an unpublished study.  Certainly if someone walked into my
own office and demanded to examine data for an unpublished study, I would
absolutely refuse and might suspect the person of wanting to use my data
in their own study.
        Just because data are in electronic form does not mean they should be
public or made available to someone who asks.
        The situation for data supporting published studies is so
different.  Once a study is published, colleagues expect to be able to view
data--even raw data not listed in the publication--so that they can attempt
to verify or repeat experiments or observations used in deriving the
published conclusions or hypotheses.
        I think we must be careful to avoid feeling that electronic data
somehow are so different than paper data that everyone should have access.
For example, people would be outraged if someone opened and read their regular
snail mail.  However, I have heard some people state that electronic mail
from 1 person to another (as opposed to mail sent to forums) is not to
be considered as private.  In this age of increasing computer data banks its
important to think hard about what data should be private and what should
be released.  Policies will have to consider the wide variety of data, and
(at least for biological data) the potential harm that publication of data
might cause.