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

Re: hidden composites

Berthold K.P. Horn wrote:
>    I wrote (thinking about extracting information from PFA/PFB files):
>    probably go for the composite character entries).  I'd be really interested
>    if someone who knows could enlighten me, just for curiosity's sake.
> PFM's are rather anaemic.  They do not contain character bounding boxes.
> They contain no information on unencoded characters.  They do not have

Which means kerns for unencoded characters are definitively lost...
(Examples are fi, fl on a PC)

>    (Similarly, I've never really understood why Adobe introduced PFMs and
>    MMM files when they had AFMs and AMFMs, it just seems to add needless
>    complexity and incompatibility to font issues.)
> For the same reason TeX uses TFM's rather than say PL's: speed and
> compactness.  For many fonts the AFM file is almost as large as the
and for some half-hearted interface to make life easier for M$-Win:
store win face names and little confused attribute bits ;-)
The essential message would be here that PFM's do not contain the height of the
individual characters in a font, which means that M$-win *cannot* access
character heights and consequently uses the *font* bounding box as an ersatz.
Given that is true *no* software getting metric information only from the M$-win 
"fontsystem" and not directly from the fonts (AFM or PFB) will be able to
produce reasonable typography. (Most people will know the line-spacing 
problems in M$-word - when a font which has only a single "large" character 
comes into play). Simply a design fault ;-)

Another version are OFM-files: That's what Os/2 uses to compile from
AFM-files. I'd be curious if there is a specification available
somewhere. Especially interesting would be the question of completeness
also for this case.

> As for AMFM's, I have never seen one and am beginning to doubt whether
> they really exist.  

They exist! The news is that with the level 4 ATM-compatible mutiple
masters you will need besides the AFMs also access to the PFB itself for
blending (at least in the general case). That means that from this point
the metric files no longer contain the *complete* metric information of a font.

> I always make AFM or TFM files for my Multiple
> Master instances directly from the installed font.

Berthold, there is a little problem with that, however! This method
doesn't work for the composites in multiple master fonts (in contrast to
PFAtoAFM) - or do I miss here something?

Hilmar Schlegel
mailto:hshlgaii@mailszrz.zrz.TU-Berlin.DE?Subject=Mail response