\centerline{\MFb\bf\ meets at Oxford}
\noindent
``Do you wish to succeed with \MF\ or even just get
it to work?'' ran the message from Charles Curran 
in \TeX hax. I was intrigued, and joined him, and a 
small band of other would-be \MF-ers at Oxford University
on March 20th. Part of the appeal of this meeting was
the presence of Richard Southall, well-known in
the \TeX\ world for his work with \MF. The bulk
of the morning was given over to a talk by Richard on
his approach to font design using \MF. Richard had
used mainly the `old' \MF\ (described in {\sl`\TeX\ and} \MF),
rather than the more recent \MF\ 
({\sl`The \MF\ book}).
Nevertheless he was clearly far more competent in the use
of either than the combined might of the others present.

Richard described the old \MF\ as using a pen metaphor, while
the new one adopts an outline fill approach. Nevertheless,
his comments as a someone trying to design `technically
satisfactory typefaces' were applicable to both. He noted that
there were few, if any, good books on the aesthetics of type, and
that type design is essentially a craft which utilises
unarticulated rules (in other words, its not easy, and it's difficult
to say why some designs are good and some aren't). There has been
a tradition of type design since the 1500's, and certain internal standards
have evolved which have formed opinion of what type ought to
look like: insofar as objectively defined criteria
of goodness of design could be established, Richard suggested the
following: letters should look the same weight, size and alignment, and should
appear evenly spaced. Even these apparently straightforward criteria
are hard to achieve over a single type face, let alone a family.

\MF\ was viewed by Richard as a draughting, rather than an original design tool,
and when he used it he tended also to use alternative approaches.
He began at the large scale, and then honed down the details.

One very relevant point was the impossibility to achieve
device independence --- most acutely at low resolutions. The
different `write white'/`write black' characteristics of the
Xerox and Canon laser printers mean that the \MF\ description
should take these factors into account. Similarly, Richard
queried the {\sl meta} nature of \MF, citing an example where
a description worked successfully over several ranges and styles,
but unexpectedly developed a wierd protuberance at one particular
set of parameters (within the previous parameter extremes). There
was no {\it a priori} way of demonstrating that this would occur.
In other words there appears to be no correctness or consistency
proof for a \MF\ description.

What \MF\ can do well (despite this last counter-example), is
to allow gradation of sizes, and to digitise an outline. What it
does not do (and he included Ikarus here too), was to allow
a designer to work directly on the appearance of the product.

It must be stressed that Richard has worked particularly on
low resolution fonts, especially those for vdu-type screens, and
subtitling for films. He felt that by the time you get to
1200 dpi, then shape and appearance become synonymous.

Given Knuth's oft-repeated recommendation that
\TeX\ people talk to designers, it was clearly of the
utmost importance to have this sort of discussion
led by someone of Richard's knowledge and experience,
especially someone who was aware of the deficiencies
of \MF, as well as its good points. There are many pitfalls present
in font design, and it would be wise to avoid them as far
as possible. The simple solution appears to be to become
apprenticed to a recognised type designer for a decade or so,
and then use \MF\ (or some other automated system). Otherwise
you will be in trouble.

That's not to say that \MF\ is not useful to the
lay user here and now. It is useful in order to produce
pixel files from existing descriptions, perhaps
at other resolutions, or for other point sizes. It will also
be useful to generate the other characters which
are currently unavailable. If you are willing to
sacrifice the {\sl meta} nature of \MF, the problems are
less acute. Sadly, we cannot afford to wait until
a competent designer picks up the financially rather unprofitable
exotic fonts like Sanskrit, hieroglyphic,
even IPA, to name a few (although I'm not sure whether
you could ever know whether a font like, say, Runic,
was well designed). These are fonts which are most likely to
be developed by enthusiastic amateurs.

In the afternoon we managed an almost hands on
session with pc\MF, thanks to \DW\
and the pc hardware might of Oxford University.
This reasured everyone that it was possible to
create characters (Dominik is working on a Sanskrit
font),
and gave us some feeling for the response time of the
process. A font will not be designed in a day.

Further \MF\ meetings like this one will be held,
although details are not fixed at present. Contact:
{\obeylines\parindent2em
Charles Curran
Oxford University Computing Service
13 Banbury Road
OX2 6NN
0865 273269
Janet:|charles@oxford.vax2|
UUCP:|ukc!ox-prg!charles|
}

\rightline{\sl Malcolm W Clark}
\vfil
\eject
\centerline{\bf Macros on disk}
\noindent
I can distribute two sets of macros on disk at present;
on 5{\frac1/4} inch disks for MS-DOS, and 3{\frac1/2} inch
diskettes for the Macintosh. The macros available
are: |manmac.tex| which are those listed on pages
413--425 of {\sl The \TeX book}, and include some double
columning and verbatim facilities: and |table.tex|,
together with its documentation file |tabman.tex|,
which are Ray Cowan's table macros.

It also includes a file of hyphenation exceptions, |hyphen.tex|, as
published from time to time in TUGboat. This particular
file will be updated as new exceptions to \TeX's algorithm
come to light. 

A disk containing these macros can be obtained from me
for the price of the medium, post \&\ packing,
and a small copying charge. For the 5{\frac1/4}
disks the charge is \quid2.00, and for the smaller
diskettes \quid2.50. Note that no
charge is made for the macros themselves.
I hope that other
macros may become available later.

\rightline{\sl Malcolm W Clark}
\vfil

\bigskip
\centerline{\bf Other \TeX ing}
\noindent
As \TeX\ becomes more and more widespread, material published
through it becomes more and more widespread, to the extent
that nobody bothers to mention the fact that it was
\TeX ed. I do think it is a pity that \TeX\ is
not acknowledged as the workhorse, but I suppose
the fact that it is taken for granted is encouraging.
It can be useful to show people who are interested in
using \TeX\ the sorts of output that are available.
If you can show them material that others have done,
they may be able to gain a broader view of its
capabilities.

From time to time I do note  publications that are
\TeX ed. Among them are the following periodicals:
ST-ECF Newsletter (ST-ECF is the Space Telescope ---
European Coordinating Facility); the Newsletter of
the International Association for Pattern Recognition,
(whose use of amr gave them away --- especially that little
blip on the `y', so characteristic of Almost Modern Roman);
and lastly the journal `Supercomputer', published by SARA (Amsterdam
Universities Computing Centre).

 There must be more:
where are they?

\rightline{\sl Malcolm W Clark}