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

Re: fontinst 1.800 vs 1.801

At 06:34 PM 6/7/99 +0200, Walter Schmidt wrote:
>On Mon, 07 Jun 1999 12:18:09 -0400, Berthold K.P. Horn wrote:
> >Could you pass the two TFMs through TFtoPL to see what is different?
>the checksum only

Oh :-)  Nema Problema!

The checksum as implemented at the moment is basically useless.

The are 32 bits there that give you just *one* bit of information:
the checksum does not match or it does match.  And even that
one bit is not trustworthy --- if it doesn't match it isn't clear what 
it means.  Does it mean the checksum algorithm has changed? 
Can it be ignored?

A much better use for the checksum is to hide information about
the font, the font encoding in particular.   That way you can
later figure out why things don't work.  And instead of saying
"checksum mismatch" which says next to nothing ---
it can say things like: the DVI calls for this font with LY1 encoding, 
but the TFM file used by DVIPS is set up for 8r encoding".

Of course DVIPS shouldn't be referring to TFM files in the
first place, but that is the beginning of another sermon :-)

Regards, Berthold.

Berthold K.P. Horn		mailto:bkph@ai.mit.edu