Oh, i see.
What about a spoken transcription? When the enumerators were filling in the census in the 1900's, if they heard SMIFF they would surely have written SMITH? knowing that the illiterate householder just mis-pronounced their own surname?
I imagine that varied according to the enumerators at the time. I get the impression some of them were more literate than others. And yes..if they knew what the name really was they probably wrote it.
But this can be an issue with for example H names. Eg someone might say their surname is ORTON and the enumerator thinks , “Oh, from the way you speak, you must mean HORTON”. But actually, they really are ORTON (of course this is the sort of name that changes anyway over the generations for that very reason, but you understand what I mean I hope)
Anyway…all this is to show - always look at the original image if you possibly can (because the other thing that transcribers have is a certain number of fields available to fill. And sometimes there is information on the image that isn’t in the transcription.
The other problem is fast typing. It is very easy to write for example 1689 instead of 1698.
And I’ve done transcribing where the thing I am transcribing clearly has an error, but because it is there in the original document It has to be perpetuated.
It would be very boring if it was all straight forward. There wouldn’t be the satisfaction of solving something.