Policy on transciption of full stops

"The simple position is that FreeBMD is a *transcript*, and should remain as true as humanly possible to the source that we are transcribing. That means that if there are full stops, they should be reproduced.

We can do all manner of clever things with the raw data, but ONLY if what we have is the raw data. As soon as somebody decides that they will "adjust" the data to make it easier to transcribe, or easier to search on, we are on a downhill slope.

Having said that, I can completely recognise that omitting/including a full stop is down at the VERY bottom of the errors that can occur in transcriptions. It is something that we can design the software to work around.

Understanding why it is important to get right something we can work around, requires a rather more in-depth understanding of the processes we use to match double keying, than one can acquire from simply observing the database.

At the simplest level, we can consider whether an entry is an exact match with another entry. That would be a binary yes/no decision, and very simple to implement.

FreeBMD uses much more intelligent algorithms to decide the extent to which entries match. We consider the concept of a perfect match, and many degrees of imperfect match. Thus if one transcriber has "David" and the other has "David.", we would regard it as an imperfect match, but a match all the same.

Likewise, if somebody transcribes "Stoke T" and somebody else transcribes "Stoke on Trent" (somebody isn't transcribing what they see), we regard it as an imperfect match. (a somewhat more imperfect match).

Overlay that with the fact that we also consider whether adjacent entries match,(using code borrowed from the Human Genome Project), and you can see that what we actually have is a degree of correlation rather than a binary matching.

Analysing the number of imperfect and failed matches by transcriber, helps us to work out which transcribers are getting it wrong.

So, omitting the full stops is wrong. It will flag as an imperfect match, and will (marginally) reduce the transcriber's accuracy rating. It will NOT however reduce the rating anywhere near enough to make it worthwhile going back and amending previous transcriptions. It is worth getting it right in the future though."

FreeBMD Main Page

Search engine, layout and database Copyright © 1998-2024 Free UK Genealogy CIO, a charity registered in England and Wales, Number 1167484. We make no warranty whatsoever as to the accuracy or completeness of the FreeBMD data. Use of the FreeBMD website is conditional upon acceptance of the Terms and Conditions