Mac Acrobat bugs as function of font encoding

["Y&Y, Inc." <support@yandy.com>]

Hi:

Melissa O'Neill was kind enough to put up some sample PDF files
done with Helvetica under different font encodings in

	http://www.cs.sfu.ca/~oneill/personal/encodings

These preview fine in Windows, and on Unix, but not on the Mac.  
She wrote:

> A significant reason not to use LY1 is that ... PDF files using the
> LY1 encoding don't view properly on the Macintosh Acrobug Reader

I saw no problems with these files on an old Mac with an old ATM.
I now had another look on a new machine and reproduced the bugs
she mentioned.  

Interestingly the error messages affect *all* encodings tested other 
than Macintosh Standard Roman encoding, so are not specific to LY1.
See attached.

I then reran the PS file through the Windows Distiller and discovered
this did not change anything.  This is not too surprising since the
Distiller is relatively bug free, while the Reader generally is not.

The problem is clearly in these cases that Helvetica is locally
available and Acrobat Reader is unable to reencode the font to
make the extra characters accessible.

The same problem does not occur if you use fonts that are not locally
installed.  I reran the tests with LucidaSans-BoldItalic instead of
Helvetica.  See


http://www.YandY.com/louis/lucida/texnansiencoding-sample.pdf

and 

MacRomanEncoding-Sample.pdf
NextStepEncoding-Sample.pdf
PDFDocEncoding-Sample.pdf
PDFDocEncodingPlus-Sample.pdf
TeXBase1Encoding-Sample.pdf
TeXnANSIEncoding-Sample.pdf
WinAnsiEncoding-Sample.pdf

(except for some reason when transferring the files from Unix the
file names all got lowercased...)

I was not able to reproduce any problems with the ``Modified shuffle
buffers are finally efficient'' test file when using LY1 encoding and Acrobat
Reader on any platform.  Since this is a real text example with ligatures
and quotation marks this is a more relaistic test than the above.
A version of this test may also be found on the above site:

	shuffle.pdf

This is on OS 8.0 where I understand a Type 1 rasterizer if built in.
I don't know how to disable this, or what version of ATM this corresponds to.

Clearly the Mac Acrobat Reader is still not ready for prime time.
Still, PDF documents using LY1 encoding generally work unless 
(i) you use characters not accessible on the Macintosh and 
(ii) usefonts  installed locally on the Mac.

Regards, Louis.

Here is what I got with Melissa's test files:

MacEncoding			no complaints
NextStepEncoding	"Helvetica 144" could not be displayed or printed ...
PDFDocEncoding		"Helvetica 138" could not be displayed or printed ...
PDFDocPlusEncoding	ditto
TeXBase1Encoding	"Helvetica 6" could not be displayed or printed ...
TeXnANSIEncoding	"Helvetica 128" could not be displayed or printed ...
WinANSIEncoding		"Helvetica 138" could not be displayed or printed ...

Note that:

144 in NextStepEncoding is Eth
138 in PDFDocEncoding is minus
6 in TeXBase1Encoding is Lslash
128 in TeXnANSIEncoding is Lslash
138 in WinANSIEncoding is Scaron

these are amongst the 21 characters not accessible on the Macintosh.
So it seems that this version of ATM Reader, on this particular Mac
is unable to reencode fonts, using a remapping of the Mac layout instead,
which fails when a character is called for that is not in Mac encoding.

Note that even PDFDocEncoding has this problem!

It only works without complaint for Macintosh standard roman encoding!

The problem goes way if you don't use fonts that are locally installed.
It also goes away if you don't use any of the 21 characters that are not
accessible on the Mac.  Either way...