[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Best Encoding for PDF (was: `limitations' of OzTeX)

Hilmar Schlegel writes:
> Since the discussion has the target to produce reliable (with respect
> to Acrobat bugs) and complete PDF files the deficiencies of the
> individual platforms and especially the previewers is completely
> irrelevant here.  (PDF provides all available characters on all
> platforms...)
> It is not the question which system Tex is running on but the optimal
> way for making Tex output a fine cross-platform communication format.

This viewpoint in large part echos my own views, but it doesn't seem to
be the view of those creating LY1 or 8r, given that the first design
goal of 8r is listed as:

% Character code assignments were made as follows:
% (1) the Windows ANSI characters are almost all in their Windows ANSI
% positions, because some Windows users cannot easily reencode the
% fonts, and it makes no difference on other systems. The only Windows
% ANSI characters not available are those that make no sense for
% typesetting -- rubout (127 decimal), nobreakspace (160), softhyphen
% (173). quotesingle and grave are moved just because it's such an
% irritation not having them in TeX positions.
% (2) [...]

... and in LY1, similar decisions were made:

% Character code assignments were made as follows:
% (1) The character layout largely matches `ASCII' in the 32 -- 126 range,
% except for `circumflex' in 94 and `tilde' in 126, to match `TeX text'
% (`asciicircumflex' and `asciitilde' appear in 158 and 142 instead).
% (2) The character layout matches `Windows ANSI' in almost all places,
% except for `quoteright' in 39 and `quoteleft' in 96 to match ASCII
% (`quotesingle' and `grave' appear in 129 and 18 instead).
% (3) [...]

For best PDF, however, trying to be compatible with Windows ANSI is not
necessary and so certainly should not be the primary design goal.

Probably, for optimal indexing and searching of PDF files, it may be
that the closer the TeX encoding is to the PDFDocEncoding, the better.
This is certainly the approach I take in the (unreleased) encoding I



P.S. My encodings are used both as a back end encoding for dvips and a
front end encoding for TeX, so using the vanilla PDFDocEncoding would
painful. So, I juggled the quote characters around to match the
sensitivities of TeX, and added some additional ligatures and dashes.
This encoding works with Acrobat Exchange and ATM Deluxe 4.0 on my
friend's Macintosh, but I have no data on whether it works well with
other configurations. I can provide the encoding and some test files
if anyone is interested.