In responding to this message, I must first state that this subject is NOT appropriate for FreeBMD-Admins-L, which is *strictly* reservedfor transcribing queries and answers. Expressions of opinion as to how things are to be done differently must only take place on FreeBMD-Discuss-L. If people wish to discuss this issue further then they must do so ONLY on the apropriate list. Quoting Dick Bond <dick@bonds.plus.com>: > Sue has, for me, hit the nail on the head!!! As an humble transcriber I feel > that we are all being asked to be very pedantic in ways which are completely > unnecessary. Which implies that we are sitting here coming up with entirely arbitrary rules to make life difficult for transcribers. A ludicrous suggestion. > I understand that there may be good theoretical reasons for 100% TWYS ..... > but .... surely our ONLY (emphasis, not shouting!) objective is to produce a > very large index which is to be available on line. Our objective is to produce a large index which is entirely true to the original. > To this end SOME data > must be absolutely as in the GRO index - but other data is changed before it > is made available. If data is to be modified/improved before being made > available then how pedantically correct does it have to be in the original > transcription? The data in the underlying database is most emphatically *NOT* changed in any way. Any apparent changes are merely in the display of the data. the processing of the data is carried out using the raw unmolested data. > For instance, the discussion that Barry raises is basically whether or not > we transcribers are causing any problem by entering 'W.Derby', or 'W. Derby' > when we might see 'W Derby' (or vice versa). Now since there is an > 'aliasing' group which modifies these all into a single version for the > database (say, 'W. Derby') then where is the problem in relation to our main > objective? The aliasing group does NOT modify them. The entries are in the database EXACTLY as typed. The aliasing team simply links all the variants together to allow for them to be searched as a single district. > If, as in another recent discussion, 'St Olave, Bermondsey' is transcribed > as 'St Olave' then this is clearly wrong - but if the ONLY difference is a > space, period or comma (e.g. 'St.Olave Bermondsey') then surely there is an > aliasing process which modifies this before it is seen on the database. No, the aliasing process merely identifies that these are variants of a single district. It doesn't modify them. > The > aliasing process may be more complicated due to such variations, but I > submit that since all versions will appear at some time on the original GRO > index, then there will have to be an entry in alias conversion in any case. Yes, there are many thousands of aliases, but that is something that we can cope with very well. > I'm sure we can all appreciate that names and page numbers should be entered > 100% TWYS. Our friend Jonh Smith might indeed have been christened as that > and it is not up to a present day transcriber to modify the index. > > (Referring to previous discussions, however, I would apply similar logic re > changes between transcription and database if a 'dot/period' appeared after > ALL {please note the ALL} the 'John's in the list and these were transcribed > without the extra dot!! The database for each 'John Smith' will read 'John > Smith' even if the index said 'John. Smith'. i.e. a user of the database > will not look for the forename of 'John.' - but will look for 'John') I am by now *HEARTILY* sick of responding to this point! You are trying to second guess just how the database build and search programmes will work. You are suggesting that the data should be made to fit in with what you think people will search on. It simply isn't necessary. The search engine is designed so that it will find these entries. > I realise that in posting this I am laying myself open to a good trampling > upon by the FreeBMD organisers - but please listen to arguments like these > and I think there may be less misunderstandings and discussion. You have argued against the need for 100% faithful transcription, but your arguments ignore some of the major reasons for insisting upon it, whilst presenting problems with 100% accurate transcription that simply don't exist. -- Dave Mayall ---------------------------------------------- This mail sent through http://www.ukonline.net