[luatex] some bits of opentype math testing

[Reinhard Kotucha]

On 26 July 2010 Arthur Reutenauer wrote:

 > >                                                       Well, I know it
 > > is a major source of confusion, but the OpenType specification choose
 > > to make our life harder by supporting two contradictory file name
 > > extension schemes!
 > 
 >   I've seen a TrueType-flavoured OpenType font with extension .otf once,
 > but it's admittedly very rare.
 > 
 >   And the main difference between TrueType- and CFF
 > (PostScript)-flavoured OpenType fonts is more likely to be in the
 > hinting engine anyway, not the degree of the Bézier curves.

I'm convinced too that the hinting engine is the culprit.  I just used
the degree of the Bézier curves in order to distinguish between the
different font formats an OTF can contain.

I have the impression that Type1 hints are ignored by Microsoft's font
renderer and only anti-aliasing is applied.  But anti-aliasing can
never be a replacement for proper hinting.  The mathematical theory
anti-aliasing is based on is quite clear in this respect. 

I mention this because a few years ago there was a discussion on a
TeX-related mailing list where someone proposed not to apply hints
anymore.  He claimed that modern printers have a reasonably high
resolution and for screen preview anti-aliasing is sufficient.

I suppose that he was bothered by fonts with broken hints.  Sure, no
hinting is better than wrong hinting.  At this time I wasn't able to
provide an example which shows the difference between hinted fonts and
non-hinted fonts.  But now I can, at least for Windows users:

   Install a Type1 font (or an OTF containing CFF) on Windows,
   configure your web browser to use this font, and enjoy the result.

I'm convinced that it makes clear that proper hinting is still
unavoidable.  At least as long as manufacturers of computer screens
believe that a pitiful resolution of 96 dpi is enough for everybody.

Regards,
  Reinhard

-- 
----------------------------------------------------------------------------
Reinhard Kotucha			              Phone: +49-511-3373112
Marschnerstr. 25
D-30167 Hannover	                      mailto:reinhard.kotucha at web.de
----------------------------------------------------------------------------
Microsoft isn't the answer. Microsoft is the question, and the answer is NO.
----------------------------------------------------------------------------