[Transferred from FreeBMD-Admins-L@rootsweb.com] Dave Mayall <david.mayall@ukonline.co.uk> wrote to FreeBMD-Admins-L@rootsweb.com on Fri, 12 Sep 2003: >On Thu, 11 Sep 2003 19:47:24 +0200, you wrote: > >>Marriages Dec 1889 >> >>FERRIS Fanny Gloucester 6a [56]44 >> >>Ferris Fanny Gloucester 6a 544 >> >>Question: is there some clever software which in time will recognize >> these duplications and merge them into a singe (bold) entry? Because >>in the meantime they are distorting the number of "unique" records >>which FreeBMD claims in its home page.... >No, ultimately all unmatched entries will be investigated and resolved >by experienced transcribers. Who will presumably be able to determine whether the page should be entered as 544, 644, or even [56]44, and that at least one of the transcribers had made an error. Now suppose that both transcribers had entered 544, or 644, or [56]44. What is the probability in each case that not just one, but both, have made an error? And would that error be recognised? -- Iain Archer

On Wed, 24 Sep 2003 20:31:50 +0100, you wrote: >[Transferred from FreeBMD-Admins-L@rootsweb.com] > >Dave Mayall <david.mayall@ukonline.co.uk> wrote to >FreeBMD-Admins-L@rootsweb.com on Fri, 12 Sep 2003: > >>On Thu, 11 Sep 2003 19:47:24 +0200, you wrote: >> >>>Marriages Dec 1889 >>> > >>FERRIS Fanny Gloucester 6a [56]44 > >> > >>Ferris Fanny Gloucester 6a 544 > >> > >>>Question: is there some clever software which in time will recognize >>> these duplications and merge them into a singe (bold) entry? Because >>>in the meantime they are distorting the number of "unique" records >>>which FreeBMD claims in its home page.... > >>No, ultimately all unmatched entries will be investigated and resolved >>by experienced transcribers. > >Who will presumably be able to determine whether the page should be >entered as 544, 644, or even [56]44, and that at least one >of the transcribers had made an error. > >Now suppose that both transcribers had entered 544, or 644, or [56]44. >What is the probability in each case that not just one, but both, have >made an error? And would that error be recognised? At some point in the future, I will complete coding of the module which analyses the data to find entries which are out of place. In the example you give, it is unlikely that Gloucester extends to page 644, so it would be caught that way. Alternatively, it will show up due to an incorrect number of entries on a page. -- Dave Mayall