Dorsey MicroFilm

Tom Leonard tleonard at prodigy.net
Mon Jan 5 18:28:52 UTC 2004


There are, indeed, services out there that will digitize the Dorsey
microfilms. The problem, thus far, has been the expense. Those service
companies aren't in it for the practice; the price comes real close to the
quarter per page you'd pay at your local public library.

Jim Duncan and I had access to a microfilm copier and we copied a great deal
of Dorsey's Osage slip file. However, the copier was old. The paper was
expensive, made terrible copies, and it was painfully slow (over 1.5 minutes
per page). The copier finally died and was sent to the junk heap.

Seems to me that scanning (to obtain image files) is the way to go. OCR, I
believe, is out of the question. OCR software sufficient to handle this task
would be cost prohibitive. I have a hunch the right scanner is now available
(at the right price) but I don't know enough about what would be required in
such equipment.

I bet there's a technical department in a university somewhere that could
tell us. Anyone have access to a "scanning techno-guru" that could shed some
light on the subject?

TML


----- Original Message -----
From: "Koontz John E" <John.Koontz at colorado.edu>
To: <siouan at lists.colorado.edu>
Sent: Monday, January 05, 2004 11:02 AM
Subject: Re: Dorsey MicroFilm (fwd)


> On Mon, 5 Jan 2004, Craig Kopris wrote:
> > Services go beyond just putting the image on cd;
> > companies will offer (for a greater fee) to clean up
> > the results as well.  I avoided that "help", on the
> > assumption that the cleaners wouldn't recognize old
> > Jesuit handwritten diacritics as other than smudges.
> > Indexing can be added, as well.
>
> What kind of cleaning up was proposed?  I assume that there was no
> question of OCR with handwriting, let alone old handwriting.  The Dorsey
> materials range through handwriting - not old Jesuit, but obscure enough
> at times - and typescript with interpolated handwritten characters and
> diacritics.  My understanding is that OCR even with pure typescript is
> also more or less infeasible, unless maybe with newer typefaces, and this
> was all done in the 1880s and 1890s.  Anyway, I figured that images would
> be the best that could be done, and the only question is, what quality of
> image is good enough to make out the necessary details?
>



More information about the Siouan mailing list