Hi,
this a RFC for adding a new mode to bibupload that we might call
--merge/-m, and that is very oriented to WebSubmit use case.
Imagine a submission that might allow a user to modify records that
might have not been created by the same submission (say they are of
other types, or they have been imported on the fly from another
repository).
It is an unrealistic dream that the submission will handle any possible
metadata that might already be existing in the record. (e.g. say it
nicely handles the author field as composed by a name and affiliation.
What if the incoming record being modified has a dedicated subfield to
store the email?)
I would like to propose the addition of a new mode that should handle in
the smartest way these cases (i.e. existing subfields not foreseen by
the submission), with the following algorithm: lets call the original
record A, and the incoming modification B
* for every field in A that is not existing in B this is left
untouched (as in bibupload --correct).
* for every field in A that has a subfield $8 (i.e. origin), which
happens to be mentioned in a corresponding field in B, the field
is removed from A and fields from B are taken (this is because
if they have the same origin, they are most probably managed in
a known way)
* for all the other fields they are taken in order of appearance
(as submission interface are usually gentle enough to keep the
order of fields stable):
* for every subfield in the field from A also existing in
the corresponding field from B, the value from the
subfield in B is taken and the value from A is
discarded.
* any other subfield in A that is not considered in B is
kept (this is to not loose additional information).
* if there are more fields in B than in A for a given tag,
these are also added to the record.
* If there are more fields in A than in B, these are not
copied over. (this is the only destructive action)
What do you think?
Cheers,
Sam
--
Samuele Kaplun
Invenio Developer ** <http://invenio-software.org/>
