Saturday, September 23, 2006

READER'S FEEDBACK: Extraction Program

READER'S FEEDBACK: Extraction Program

From: Charlotte
DearMYRTLE,
I sincerely hope that the [data derived from the] extraction program is more
reliable than the Ancestral File [database of patron entries]. I was in
charge of one of the *data entry* sites. There were 32 of them. We were not
allowed to make corrections to what the submitter sent in even when it was
obvious that a mistake was made, nor were we allowed to let the submitter
know he/she had made a mistake.

I know the extraction program is reading from the microfilm and that as
minimum of two people have to agree as to what it says, but I worked in that
too and often if the second reader saw that the first reader was someone
he/she respected as lot, if there was a question he/she always took what the
first person had put.

Charlotte in southern Utah
p.s.
Data entry site was in Texas

DearCHARLOTTE,
Your Name Extraction Program experience differs from ours in the early 1990s
in Bradenton, Florida. We were extracting names from a Catholic parish in
Mexico City. We received the microfilm, and photocopies from the film to
distribute to volunteer participants. The UDE software provided for person
"A" to fill in the blanks for the individual's name, parents where listed,
date, place; as he interpreted it from the original text. Once the packet
was returned, we distributed the packet to another volunteer. The software
for person "B" provided the blind double-entry I described in yesterday's
article. Person "B" was specifically not to know who had previously
processed the packet of pages from the original

When there were difficulties in reading the text, it was possible for either
person "A" or person "B" to come back to the two of us coordinating the
project, so that we could look at the microfilm, zoom in/out, and reprint a
page positive/negative and such. We merely voted on our best guestimate.
However, we had been trained to look at the handwriting elsewhere on the
page to compare letter for letter, allowing for irregularities in
handwriting.

The computer software would compare both "A" and "B" entries for
differences, which I theorize were usually just typos.

THERE WAS AN EARLIER MANUAL PROGRAM for extracting information WITHOUT A
COMPUTER. This meant that the extractor deciphered the handwriting from the
original document from a photocopy of the microfilm. He reported his
findings by hand writing the names and dates on little cards. Others typed
from those cards into a computer. This sounds like the *data entry* you
participated in, Charlotte.

THANKFULLY, THE ADVENT OF HOME COMPUTERS has facilitated the part "A" and
part "B" indexing of original records.

Somehow having two or more sets of eye looking at the cryptic handwriting in
an old church record seems a lot more reliable than trusting my
inexperienced, untrained eye. However, no process is without it's faults.
For that reason, Ol' Myrt continues to recommend that we obtain copies of
the original document (usually from microfilm) rather than rely on even an
extracted entry.

Happy family tree climbing!
Myrt :)
227 Bellevue Way NE PMB544
Bellevue, WA 98004
Myrt@DearMYRTLE.com
http://www.DearMYRTLE.com

No comments:

Post a Comment