Hi Herbie,

In upgrading  pdfIdentifier/AcroButler, you might
consider the following two supplements. 

(1) I have just modified Acrobat Reader 4.0 (=AR)
to accept files of all types -- by replacing the
BNDL resource and adding a FREF resource #133 for
type ****.

The old and new versions of the resources of these
two types are supplied. as AR4.0patch_DIR.cpt.hqx
-- just 1294 bytes posted at

ftp://topo.math.u-psud.fr/pub/lcs/fb/DragAndDrop/

Installation and deinstallation currently require
Resedit, but AcroButler could do that as an 
optional 'extra'.

Incidentally, since AR does the header check 
before beginning to process a PDF file, 
there is no risk whatever in making this 
modification.  AR even posts a polite alert
on aborting.

(2)  Someone in the Adobe front office may have
thought it clever to accept text files.   
DANGER!!!

In fact, TEXT files are terrible hazard
(unfortunately known mostly to experts).  A PDF
file that has been transferred as a text file and
thus subjected to change of line endings is very
often unusable to Acrobat Reader.  For one thing,
because offset tables are thrown off, and for
another thing because binary blocks get damaged.

(The ONLY significant case where the PDF survives
intersystem text file transfer undamaged is when
there are no binary blocks, *and* DOS/Win is *not*
involved.  It is true that AR will attempt to
repair the offset tables, but often that is 
just not possible.)

It would be hard to have a *rigourous* test for the
above sort of damage, but type TEXT is a strong
danger signal.

REMEDY:
With this in mind, pdfIdentifier/AcroButler might
do well to mark as possibly damaged those PDF files
that originally have type TEXT.  This can be done
gracefully by inventing a new type and icon, the
icon being an Adobe PDF icon with a crack and a
question mark.

Cheers

Larry S