laser-lovers@uw-beaver (06/05/85)
From: Cal Jackson <CALVIN@CIT-20.arpa>
Recently, we acquired a LaserWriter. Also, we acquired a
dvi-to-postscript translator. It is important to note that we
received a translator and supporting data that was clearly marked as
preliminary. Our other printers (and quantity) are: Versatec (1),
Imagen (2), and HP 2688A (2). The impetus was: PostScript appeared to
be a desirable intermediate language for the existing and developing
"homegrown" graphic languages. We use TeX, almost exclusively, as the
composing/formatting tool. Much of the research work here involves
techniques for designing VLSI components, therefore, many homegrown
graphic languages. Another excitement was the apparent access to
typefaces other than Computer Modern (CM).
Authors of graphic languages had, almost trivially, integrated their
work into dvi-to-impress translators provided by Imagen or local
dvi-to-device drivers. It turned out that this ease also occurred with
the dvi-to-postscript translator. Some authors feel it is easier to
implement their graphic primitives using PostScript (PS) than with the
other translators/drivers.
We had and are having some trouble setting text, both with the
Computer Modern and the additional fonts. The downloading of rasters
for the CM fonts is expensive and tranformations on rasters is not
appealing. For the other fonts, the metrics required by TeX were
incomplete or inaccurate. We felt some problems were due to the data
provided with the translator (e.g., the TeX tfm file for Allied Times was
close to published ADOBE widths, but included the TeX amr10 kerns).
We made some inquiries (week of May 27) to ADOBE and they
shared the following:
Kerning data: No kerning data has been released to date.
Data for the Allied Times family is expected to be released in about six
weeks. The ADOBE manual was printed using PostScript (PS) and without kerns.
Font metrics: The distributed metric data, for printer fonts, does not
permit one to determine the height and depth values for individual
characters. ADOBE supplied me a PS sequence that should generate those
values. Height and depth information will be part of the metric data
in October. (The supplied PS sequence works, but the tfm width data
values do not match those supplied in the ADOBE manual. This is
probably due to my inexperience in using the PS language.)
Automatic PS kerning: The PS 'show' operator, with a string operand,
will not enable kerning. A new operator will be provided, i.e., the
PS programmer must specifically enable kerning. This is important to
users of TeX: there is no way to differentiate in a dvi file
between a font kern and a user directed positioning.
Font dict (kerning data): It has not been decided where the kerning
data will reside. Current thinking is that it will be in some
structure other than the font dict.
Proprietary info: All font metric information is public. The
geometric description of a font, the encryption algorithm ,and the
encryption software is proprietary.
Volunteered info: Someone from Stanford is working with ADOBE to
generate ADOBE encoded Computer Modern Fonts. At least the symbol and
math extension font. These are expected in October. I interpreted the
comment to mean that the TeX folk at Stanford were being fully
cooperative.
END OF ADOBE INFO
Maybe somone at Stanford and someone at ADOBE can comment on the above?
cal
------