[1] From: Jim Marchand <marchand@ux1.cso.uiuc.edu> (8)
Subject: Greek-Sanskrit isoglosses
[2] From: Jim Marchand <marchand@ux1.cso.uiuc.edu> (18)
Subject: edge-finding
--[1]------------------------------------------------------------------
Date: Sun, 19 Sep 1999 15:18:58 +0100
From: Jim Marchand <marchand@ux1.cso.uiuc.edu>
Subject: Greek-Sanskrit isoglosses
I am not quite sure what you mean by isoglosses, Chiara. If you mean
cognates, such things as Skt. pitar :: Gk. pater `father', a look into any
of the comparative Sanskrit grammars, e.g. Wackernagel, Thumb, would yield
hundreds of isoglosses. Ditto for any of the comparative Greek grammars,
e.g. C. D. Buck. An easier way to gather sets of isoglosses between Greek
and Sanskrit might be to look in: Carl Darling Buck, _A Dictionary of
Selected Synonyms in the Principal Indo-European Languages_ (University of
Chicago Press, 1949).
Jim Marchand.
--[2]------------------------------------------------------------------
Date: Sun, 19 Sep 1999 15:19:18 +0100
From: Jim Marchand <marchand@ux1.cso.uiuc.edu>
Subject: edge-finding
I think we may be having a terminological problem with "edge finding".
Determing the edge of areas in a digital representation is an old procedure,
and the algorithms for it are well known. Rafael C. Gonzalez and Paul Wintz,
_Digital Image Processing_, 2d ed. (Addison-Wesley, 1987): 334: "We define an
edge as the boundary between two regions with relatively distinct gray-level
properties." If finding an edge under this definition is what is sought,
then most commercial "photo" and "paint" programs do this. Old as it is,
Gonzalez-Wintz will tell you all about "edge detection", "edge linking",
"boundary detection", "detection of discontinuities", and there are many,
many newer discussions. Edge-finding and edge-linking are important in
working with manuscripts where the letters may be blurred, broken, flaked
off, you name it.
On the other hand, many of the tasks one encounters with shards, pieces
of papyrus, lead plates, etc. are more on the order of jig-saw puzzle
problems, and, whereas there are programs for this, none seem overly
successful. We need to ask the original poster precisely what he is looking
for.
!
Jim Marchand.
-------------------------------------------------------------------------
Humanist Discussion Group
Information at <http://www.kcl.ac.uk/humanities/cch/humanist/>
<http://www.princeton.edu/~mccarty/humanist/>
=========================================================================