Ron Savage wrote:
OK. So we're saying we should be able to store multiple versions of any
assertion, since - at least at first - it's all 'just data'.
Then, we should be able to report on various matters regarding this
data, such as the error/warning (of type 'inconsistent') when someone
says today (Aug 6th) is a Sunday, when in fact it's a Saturday, since
they may have meant Aug 7th.
To me this suggests that all the data be preserved indefinitely, even
after the user (or the program as the user's behest) has somehow chosen
one version to be promoted as the definitive version (not necessarily
making it correct), even though future research could change the status
of that version. This in turn suggests the need for a multi-valued flag
attachable to each version indicating the researchers degree of faith in
the veracity of that version.
Yes, much as you say, though details can vary.
Do any current programs support such labelling of assertions?
None that I know of, but I am aware that some savvy genealogists retain all of
their sources even if contradictory.
-- Darren Duncan
