A blind cross-read
Old handwriting is where a history most easily goes wrong, and goes wrong confidently. A clerk's hurried loop can pass for one letter or another, and a reader who half-expects a name will find it there. So the project keeps a firm rule for any value a claim will rest on that has to be read from a hand: a name, a street, a house number, a date, an act number, a signature. One reading is not enough to register it. Two independent readers of different makes read the tight crop blind: the project's own reader, and at least one from a different family outside it. Neither sees the other's answer, and neither is handed a candidate to confirm. Each is asked cold, because offering an answer is how a wrong one gets confirmed. Only then are the two readings weighed against the scan itself and against whatever else the record holds, and the project's own reader settles the value.
The rule caught a mistake that had already been made, which is why this note exists. On two Rotterdam birth acts of 1922 and 1923, the home street of the painter Gerrit van Duffelen's family, written by hand, had been read as Eendrachtsstraat and recorded that way. Read again blind, the crop came back close but not clean. Both readers landed on the -hoorn- street, but on the 1923 act one of them slipped on the middle letters and returned Eerhoornstraat, a cursive trip between an n and an r. Checked against the scan, the address books and the city's own street index, the reading was fixed as Eenhoornstraat, a small street near the Leuvehaven, and the record now carries that value. The address books list the painter himself at no. 4a there in 1923 and 1924.
The 1923 act brought a second correction. Its declaration line and the family's household card name the son Kees, where the Open Archieven index had read Elis. That correction belongs to the record too, but the street name is the worked example of a blind cross-read here.
The 1922 act carries the corrected street. In the clerk's set closing formula it also gives the elder son's own naming line, the same construction that on the sibling act turned Elis into Kees:
Today, 4 August 1922, appeared before me … No. 7402. Gerrit van Duffelen, thirty-two years old, painter, residing here, who declared that on 3 August of this year … here, at his home at Eenhoornstraat number 4a, a child of the male sex was born to Geertruida Kaleveld, without occupation … which child shall be named Frank. The ellipses mark passages omitted from the Dutch extract; they are not unacknowledged translation omissions.
The rule decides which readings may be trusted, and when. A single reading of handwriting, however confident and whatever produced it, is only a finding aid to be checked, never the registered value. The registered value is the one kept after two blind readings and a comparison with the document. For the street name, the scan comes with a note of who read it and what each reading returned, so the decision can be re-examined later instead of taken on faith.
The discarded reading is not erased either. Eendrachtsstraat stays on the books as a reading the project has since dropped, kept so that no later draft can wander back into it. Under the current method, no new handwritten value that a claim rests on becomes accepted knowledge until it has passed this check. Older records remain subject to correction where their reading provenance does not meet that rule.