Re: design size in DVI

[Laurent Siebenmann <lcs@topo.math.u-psud.fr>]

Thanks to Berthold and Nelson for replies to my
question:

 > (2) Does/should anybody or anything currently pay
attention to
 > the design size entry?

I see it essentially as Berthold does.  Some extra
details:

-- DVIPS will complain if a font's "design_size" in
the DVI does not coincide with design size in the
TFM; many other drivers ignore it entirely.

-- The size of the print is entirely determined by
"at_size", and is directly proportional to it. More
precisely, a character's box width on paper is:

         "TFM width" * "at_size" * mag/1000

Similarly for all other dimensions dictated by the
TFM.

-- DVI drivers ignore "design_size" in the DVI and in
the TFM or VF, except for the mentioned coherence
check; ie drivers at most use "design_size" as an
extra checksum.

-- TeX invariably copies "design_size" from TFM to
DVI, which is the same behavior as for the checksum.

-- TeX however also uses "design_size" when the
following syntax is met:

     \font\fnt=myfont

and it responds by setting the "at_size" in the DVI
equal to the "design_size" in "myfont.tfm". It
similarly uses "design_size" in interpreting

     \font\fnt=myfont scaled <mils>

But TeX does *not* use "design_size" in interpreting

     \font\fnt=myfont at 14pt

indeed it simply sets the "at_size" in the DVI to
14pt.

IN SUMMARY: "design_size" is used in a way that
influences what appears on paper/screen only by TeX,
which,  merely for user syntax convenience,
sometimes uses it as a default "at_size" in
interpreting font definitions.  If the use of the
keyword "at" in TeX font definitions were
*obligatory* (thus "scaled" would be forbidden) then
"design_size" would be nothing more nor less than an
extra checksum, something that Carl Berry and
Berthold Horn badly want.

Near miss. Too bad!

Cheers

Larry S