OCR rendering has a minor consistent error throughout; each surname is doubled up

WillardKelchner

Full Text Search Keywords: "Wills Probate / Samuel Mears / Catawissa Twp, Columbia County / 1747 to 1813"

Full-Text Search Returned multiple hits, including

"Bloomsburg, Columbia, Pennsylvania, United States records," images, FamilySearch (https://www.familysearch.org/ark:/61903/3:1:3QS7-89SM-WN9Y?view=fullText : Jan 18, 2025), image 762 of 777; Pennsylvania. County Court (Columbia County). 005544891

Selecting this citation your softwares delivered Document Image 762 displayed correctly. However the OCR rendering has a minor consistent error throughout; each surname is doubled up. I provide the OCR rendering with the error highlighted.

I Samuel Mears Mears of Catawissa Township , County of Northumberland , have arrived to an advanced age and being low in health , but of as sound a mind and understanding as usual Do make and ordain this instrument of writing as my last will and Testament in the manner and form following . First I give to my wife Mary hears , the plantation whereon we now live , except a part hereafter disposed of otherways , for her to injoy the issues , rents and profits thereof during her life . I likewise give her all my personal property that shall be found after my decease , except such things as are hereunder mentioned and otherways bequeathed . Secondly after the decease of my wife I give and devise the aforesaid personal Estate , or such part or portion thereof as may then be found to my Daughters Sarah Hughes Hughes , Nancy Fleming Fleming , Mary Pancost Pancost , Martha Mears Mears and Elizabeth Ent Ent , to be equally divided among them , or their surviving children..

I have not seen this error elsewhere in the search hits.

Most Appreciatively, Will Kelchner

Mark McKenzie_1

@WillardKelchner This is something I too have pretty consistently observed when viewing transcriptions of image records. Seems to be confined to just surnames.

Some transcriptions are worse than others which could be explained by images processed 'early' in the life of Full Text Search as compared to later transcriptions generated with newer/improved algorithms…

Either way still pretty amazing as to what the algorithm can accomplish