# Re: fontinst 1.801 -- still worse

```Walter Schmidt wrote:
>On Mon, 7 Jun 1999 23:45:03 +0200 (CEST), Ulrik Vieth wrote:
>
>>I guess this means we'll all have to be carefull about putting
>>\installrawfont commands after every \transformfont into mathpple,
>>mathptm and mathptmx, so as to avoid checksum problems.  Sigh!
>>
>>Is this really true?
>
>I agree:  Doing so should theoretically fix the problem.
>
>checksums with less effort:
>
>The fontinst script for mathptmx starts with:
>
>\transformfont{ptmr8r} {\reencodefont{8r}{\fromafm{ptmr8a}}}
>\transformfont{ptmri8r}{\reencodefont{8r}{\fromafm{ptmri8a}}}
>\transformfont{psyro}  {\slantfont{167}{\fromafm{psyr}}}
>
>Instead of creating ptmr8r, ptmri8r and psyr again from the
>afm's, we could take the tfm's found under fonts/tfm/...,
>convert them back to .pl, and use these.  This makes sure
>that the faked math fonts always refer to the same "real" fonts
>as TeX does.

I don't think that would be such a good idea as it might sound, since there
are rounding errors introduced in all these conversions. A better way would
be to create the ligful PLs from the same MTXes as was used for the
non-ligful ones; this should ensure that the character metrics really are
the same.

Of course, I'm not quite sure at the moment how the TFMs in fonts/tfm/...
were made, so there might not be any MTXes to start from.

As for the checksums in VFs, I think that subject was debated here some
time ago and that it ended with "that's just the way things are", but
unless my memory fails me, Berthold will most likely volunteer to tell you