Re: bakoma tfms: Checksum mismatch

["Berthold K.P. Horn" <bkph@ai.mit.edu>]

Hi:

At 03:20 PM 3/27/98 +0000, Rebecca and Rowland wrote:

>>>1.  Is it possible that tex uses bakoma's tfm and xdvi cm's, or the other
>>>way round and this be the cause of the warnings?  ---

>>A dangerous situation when the DVI processor sees different
>>TFM files than TeX.  Which is why I prefer DVI processors
>>that do not need to refer to TFM files (after all, the font
>>knows its metrics).

>This is a bad idea because the virtual fount mechanism allows you to
>construct glyphs that are made of more than one glyph from the `real'

Real important for CM :-)!  

And even with `real' fonts, not of much interest to anyone but a `TeXpert'
Most users would rather be able to use fonts without going through
purgatory first :-)  and arguing about how many angels can dance
on the head of a pin (how's that fontinst manual going :-)?

>fount, and also to create special effects by telling TeX that the metrics
>for any given glyph are different to those in the `real' fount

Generally, a TFM file's metrics not matching the metrics of the font is an
invitation to disaster.  But I think I know what you mean.  You are talking
about something else.  If you are using virtual fonts, then sure enough
the metrics of the virtual fonts won't match that of any real font.
BUT: the metrics of the fonts from which the VF is constructed better
match exactly the metrics of the corresponding real font!  

An aside: a very annoying feature of VF files is that they unneccesarily 
force you to look at the TFM files because the advance width of the
characters is not specified!  A most annoying design flaw IMHO.
The world would be a much better place if the tools for constructing 
VF files made sure that after each character is typset, it advances
by the appropriate width!

>So the
>metrics in the `real' fount are not necessarily any use at all.  On top of
>that, I have some (shareware and free) founts which clearly have some sort
>of bug regarding metrics and do not display correctly when I use them with
>wysiwyg applications - glyphs overlap and things like that. 

So, don't use junk fonts.  Or make sure the metrics for them are correct.
Send me such a font.  I'll probably be able to tell you what you are doing
wrong :-)  Or what the font designer did wrong...I am serious.

> I have never met any such bug in the companion afm files, so I can 
>use the founts in TeX with no problem.

That is pretty bad.  If the advance widths in the font are wrong.
My suspicion is that your real problem lies elsewhere and you
are just papering it over by typesetting each letter individually.

Junk fonts typically don't even come with AFM files. And if they
have AFM files, you can't trust them anymore than the font itself.

In addition, most DVI processors will typset strings of characters
at once, under the assumption that the font works correctly.
This obviously doesn't work if the metrics in the font are wrong.
Just look at PS produced by DVIPS: does it position each letter
separately?  No, only when it hits kerning and interword spacing
does it break a PS string up.  Hence you are screwed if the
font doesn't know its own metrics. Right?

>And, of course, pk founts do not contain metrics - not all fount files
>contain metrics.

That is wrong.  Have you looked at the PK font file format?  DVIPSONE 
can use even PK fonts without needing to refer to any TFM files (of course, 
why you'd use PK fonts when others are available isn't clear :-).

You are a very good protector of the standard faith!  Keep it up!  

Regards, Berthold.