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

Re: bakoma tfms: Checksum mismatch

By the way, there really is something odd going on:

Several of the TFMs for BaKoMa CM font metrics don't match the 
Knuthian TFM files.  Just to focus on the text fonts first, cmss9 and 
cmssi9 TFM files are different - yet have the same checksum! -
(so much for the utility of a widely used checksum algorithm!)

In CMSS9, the original TFMs have, for the digits 0-9, a height of 728.375,
while the BaKoMa TFM's give 654.321.  The actual height of the
glyphs is 654 for the flat-topped ones and 676 for the rounded ones.

Ditto for CMSSI9.  In addition, the dollar in CMSSI9 has different
TFM height and italic correction in the original and the BaKoMa TFMs!

What gives?  Any have any ideas?

When it comes to math fonts, some differences may be due to 
AMS changing from V2.1 to V2.2 for the extra bold math fonts
and the extra CMEX and extra CMCSC fonts (which are not
part of Knuth's 75 fonts).  

But that doesn't explain everything.
For example, there are differences in CMMI5, 6, 7, 8, 9, 10
and some of the CMSY* fonts.  Which are of course in the
original Knuthian 75.  Didn't have time to see what the
differences are.

Moral:  don't use the BaKoMa TFMs.  
You may get different typsetting from other people...

But why are the original TFMs `wrong' for CMSS9 and CMSSI9 digits?


t 03:20 PM 3/27/98 +0000, Rebecca and Rowland wrote:
>>At 04:31 PM 98/03/24 +0100, Stefano Frabetti wrote:

>>>I have bakoma installed, so I have checked the tfm files of bakoma
>>>against those of cm and the following were reported as differing by


>>Why would you use other than the TFMs you already have?
>>The Type 1 versions of the fonts supposedly have identical
>>metrics to the MetaFont/PK versions.

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