John Wood commented in the context of numbering complex site records...
"on our system anyway we are getting rid of all the sub-numbers and have
adopted a parent / child system instead. This entails giving a record a
'parent' where appropriate - it can be any other record. It allows much
more flexible linking between any records. Any record can have any number
of children (and if necessary grandchildren). Relationships can also be
changed if need be later without the need to renumber."
From a data standards point of view this is definitely the best approach.
MIDAS recommends that the Primary Reference Number attributed to an
inventory entry should be just that - a number or code without any other
significance other than to uniquely identify a site. 'Sub-numbers'as
referred to (e.g. a site numbered as 567 with 'child' records numbered
567.1, 567.2 etc) are not recommended as they are trying to do two things at
once, i.e. uniquely identify a site AND identify its relationship to another
entry. Incidentally, the same problem occurs with incorporating map sheet
numbers into PRNs: it is best not to, as any long or large site may well
straggle over more than one map sheet. We're moving away from this approach
in the NMR, replacing old style O.S. numbers (e.g. TQ 23 SW 123) with unique
computer generated numbers (Heritage Object Unique Identity Numbers or
HOBuids). It is these numbers that are used to relate monument records
together in Parent/Child relationships.
Look up the details of the standard on the MIDAS web-site at
http://www.rchme.gov.uk/midas/mds/nandr1.html
Edmund Lee
Data Services Unit
NMR
|