[Bug-gs] Type 1 and TrueType Issues with GhostScript 7.0?

[Ray Johnston]

Dear Mr. Lucas,

> Damian Lucas wrote:
> Are there any known issues with Type 1 and TrueType Issues with GhostScript 7.0?
> 
> Regards,
> Damian Lucas
------
Ray wrote:
> Yes.
------

I couldn't resist the chance to answer such an open ended question with
such a terse reply. Hopefully, you have a sense of humor equal to my
mischievousness. ;-)

Now to a real answer:

There is a document distributed with Ghostscript 7.00 that contains known
issues. When we learn of things that we may never be able to address, they
will generally be documented in doc/Issues.htm. Please check there in 7.00
and in all future releases.

Also, please check the SourceForge bug tracker for the ghostscript project.
Artifex and artofcode jointly maintain and improve Ghostscript and we use
the SourceForge bug tracker system to allow anyone to submit, review and
respond to Ghostscript bug reports. Bug reports that come in via bug-gs@
ghostscript.com and/or support@artifex.com are added to the SourceForge
bug tracker once the Artifex Tech Support staff verifies that the bug is
unique and has not already been fixed. Please look at:

   http://sourceforge.net/projects/ghostscript/

			for the project links, and

   http://sourceforge.net/tracker/?group_id=1897

			for the tracker services, and

   http://sourceforge.net/tracker/?group_id=1897&atid=101897

			for bug browsing.

As far as Type 1 issues, there aren't many. The main area that we know of and
have invested resources in improving is the Type 1 hinting and dropout control.
The appearance of Type 1 fonts is the issue, particularly at low (display)
resolution ~72-90 dpi. For displays, the use of anti-aliasing enabled with the
-dTextAlphaBits=4 command line parameter can bring Ghostscript to near Adobe
ATM screen quality.

As far as TrueType, the main issues are hinting and bytecode interpretation,
and the primary issue is not technical, but legal. Apple and Microsoft have
many patents in this arena and despite over four years of effort, we have
not been able to break this log-jam. You can live without hinting on the
screen by using -dTextAlphaBits=4 (see above) or printing at 300 dpi or
greater, but some Asian CIDFonts that use TrueType font bytecodes rely on
this for positioning of sub-elements of a glyph. Without a legal bytecode
interpreter, this is not solvable.

The FreeType project is working on an auto-hinter for TrueType (they had
developed a fully hinted TrueType renderer prior to the patent being
awarded and enforced, possibly in the false belief that Apple/Microsoft
would not enforce the patent in order to promote the TT font usage).
Artifex is working on a 'plug-in' interface to font renderers that will
allow users to hook to FreeType and other third party font renderers
(some of which require payment, but *DO* have licenses for the patents).

There is one bug that I just fixed that can cause the graphics library
layer to crash when using TT fonts. We don't have any report of this
affecting any PostScript/PDF users. We at Artifex discovered this when
using the Ghostscript 7.00 graphics library with our PCL/PXL interpreters.
Since we haven't yet released PCL under AFPL and even then it will be
using the 6.50 graphics library layer, this does not appear to be a
serious issue.

In case your question stems from your considering Ghostscript for use in a
commercial product, I am also copying this message to our marketing folks.
This type of synopsis of where Type 1 and TrueType stands is useful for
talking to other potential customers, even if you are not in that category.
If you do have an interest in licensing Ghostscript for commercial use,
please contact info@artifex.com.

Regards,
Ray Johnston
Director of Engineering
Artifex Software Inc.
============================================================================