Re: [Ringing-lib-discussion] Database updating

[Richard S. <ri...@ex...>]

Martin Bright wrote:

> The problem, as Don said ages ago, is to define what should be `the same'
> method.  We have an ID for each method, and I think these should be
> persistent.  Should the ID pertain to the name or to the place notation?  I
> think it should probably belong to the place notation, since I also think
> that we should potentially allow unnamed methods to appear (for example,
> the whole TDMM collection).  But then the whole problem of comparing place
> notations arises.

That problem is going to exist anyway -- people will need to
search by place notation -- so we might as well get it right
now.  I agree with you: the ID should pertain to the place
notation.  Names can be changed for a number of reasons --
reclassification, and avoiding clashes with extensions being
two.  Place notations should only ever get changed to
correct mistakes.

> We probably just need to make a slightly arbitrary decision and stick to it.

Agreed.  This will enevitably get the occasional thing
wrong, but either we can manually sort that out, or just
ignore it.  For instance, if we go with IDs pertaining to
place notations, this would have been broken when the CC
collection corrected the place notation for Reverse Tendring
Doubles a few months ago.  However, such corrections are
likely to be few and far between.

Richard