Anne Shankland said: >I would like to see two different fields on the GMI entries, one > for the member who actually provided the information, and one for > the member who has the study registered. Not so easy. As others have pointed out, a single entry may contain information supplied by several individuals: the original GRO reference from one ONS, the spouse name from another ONS, the location from a Marriage Challenger - and all that's before we move into the realm of corrections: Tony Mingay's "Puzzle" of a few months ago comes to mind. As Anne said: > But what should we do when a record submitted by member A for a > study name registered by member B is updated or corrected by > member C? A correction, even the fact that something is recognised as being wrong, may depend on the surrounding entries, which also may be the result of information supplied by several contributors. Keeping track of all that would be a huge task, and getting quite far from the original purpose of the GMI: a database to allow the location of marriages to be found from GRO references and a means of connecting members' studies where interests coincide in a marriage. When the GMI started, the right-hand columns were labelled as giving the contributor. Years ago, that was changed to "further details". Now you can see why! The GMI was not intended to be a store of members' data. We now have other databases for that purpose. Anything going in to the GMI has been subjected to at least some verification: many typographical errors are detected and removed at this stage. Where there is sufficient information in a particular area, it is even possible to spot GRO reference errors, such as the incorrect page number that led to Tony Mingay's posting. Sets of Cardinal Points for a Registration District can be checked for errors and omissions. There are some simple rules for CPs. Using the notation in the CP recording template on the Guild web site (www.one-name.org/members/GMI/CardPointCollect.html), s for a CP at the start of a quarter, f (finish) for one at the end and b (both) if the CP is the only marriage in the quarter, the first rule is that s must be followed by f in the same location, and f or b must be followed by s or b (with a different location). The second rule: if f or b is on an even-numbered page (call it p), the following CP is on page p+1; if f or b is on an odd-numbered page, the next CP is on page p+2. Checking for this can be done manually, but it is very tedious and prone to error: a computer is far better at that sort of thing. The result is that the GMI is more accurate than information supplied by any one contributor. Entries supplied by one member may be corrected by information supplied by others. It really is more than the sum of its parts. Peter Alefounder.