.... much snipped............
Reading some previous posts on this thread, I sincerely hope Ancestry are not doing the Censuses on OCR, or we're all going to be in DS - how about the surname Russell" in the 19th century? Given that the "long s" was in use in those days, I imagine a lot of ancestors are going to end up as Rufsell or Rupell, depending on the handwriting...
The simple answer is yes!!, - It's already the case in the ScotlandsPeople censuses that there quite a number of folk with the surname ROFS and ROP ! Think about any such double ss combination in a surname, and you'll find some instances, - doesn't cost any credits to have a look at the number of such hits.
Similarly there's quite a few folk on SP with the surnames SMTIH and BRWON !, i.e. transposition of letters.
The point that I'm making is that it would be inordinately expensive to make such indexes totally error free.
Apart from anything else, that would require a true double entry system, i.e. two people independently key in the data. Any instance where the input is different is referred to a third person, and resolved at that level, with any unresolved entries being looked at by even more expert folk, and so on, with further levels of expertise above that.
In my recent experience, many so-called "double entry" systems ain't!, - instead they involve the situation where a second person just checks the entry made by the first person, which just ain't the true double entry system as defined above ..............
That written, my opinions on the Ancestry indexes are well known. I have been reassured from the highest level in Ancestry that OCR or similar or equivalent technology don't come into the equation; that their sub-contractors are given extensive training, and that their sub-contractors are supplied with every which look up list for surnames, occupations, county and paroch ("parish"
) names............ OK, at the end of an exhausting 8 hour day entering data based on peculiar hands of the enumerators involved, plus some ink fading taken into account, I can just about accept that the 900+ parish list isn't properly consulted, but the number and quite ridiculous nature of many of the errors that I've seen in terms of Scottish county names, - it's a list somewhere in the 30s, isn't it?, when Haddingtonshire and similar "alternative" county names are taken into account, - does tend to call into question the Ancestry quality control procedures involved, and the validity of claims that a
true double entry keying system is used, - see above. Such a short list must surely be easy to check?, or am I missing something
However, I've slowly and painfully at times coming round to the belief that I can just about buy into the opinion that we're better with the Ancestry indexes than without, if only on the basis of the extra search field possibilities that Ancestry provides compared to ScotlandsPeople, plus, critically on occasion, the ability to search on a given name only.
BTW, on SP there's always an implied wildcard on the given name, i.e. searching on "J" is the same as searching for "J*" ................
All that written, I can well appreciate that some folk out there believe that, to some extent, they are being "sold a pup" in terms of what Ancestry claim to offer, compared what turns out to be on offer after the subscription has been paid.
At least, with ScotlandsPeople, it's very much the case that credits are not only refunded but also some extra compensatory credits most often provided, when it is shown to be the case that mis-indexing etc. has led someone to wasting time and credits on a futile search.
ibi
