|
When did you design your first typeface? What was it
called?
My first
typeface design is long buried: very few people have seen it. I started in
1993, but it took several designs before I felt confident enough to make a
typeface public, and even that one (Manticore) has been withdrawn now
because I’m not satisfied with it. I still get requests for it from time to
time, though, and I tell people ‘Sorry, it isn’t available’.
Where does the idea for a new typeface come from? Is
it an inspiration, or the result of a lot of study and research?
Almost all
the work I do is custom work, for a client. So there is usually a fairly
specific design brief, some set of requirements for what they need. A lot of
the requirements are likely to be technical ones, e.g. they need a font to
work in a particular environment or at a particular range of sizes. That
gets me thinking about possible solutions, which leads to the design.
If I’m
designing a typeface for a script I have not worked with before, then a lot
of study is involved. I also try to find an experienced typographer who is
either a native user of that script, or who has lots of experience with it,
and who can review my work. This is very valuable. It must be an experienced
typographer, though, not just any native reader of the script. A lot of
people can read a script every day and never know anything about typography,
or about the different styles and their internal consistency. The biggest
mistake I see people making when designing typefaces for scripts that they
are not very familiar with is mixing design details from different styles
without understanding what is particular to those styles. Imagine someone
trying to design an Arabic typeface that takes bits and pieces from naskh,
nas’taliq and ruq’ah styles: it would be a mess. So you need to do a lot of
research and you need to find people who really know the script well who can
advise you.
What are the stages of your font design?
If I am
making a font that supports more than one script, I make a point of working
on them at the same time. I don’t do the Latin first, for example, and then
the other script or scripts. If the goal is to make the different scripts
harmonise, even if this only means harmonise in terms of overall typographic
‘colour’ on the page, it is important that they be designed side-by-side.
I usually
start by identifying the key letters in a script that determine the basic
weight and relative proportions. For the Latin script, this would be
something like i, n, o, H, O. But then it is important to quickly move on to
some of the more distinctive letters that will determine the individual
character of the typeface much more strongly, such as a and g.
What software programs do you use to make a font?
I use a lot
of different software. I use FontLab for glyph design and general font
production. I use MS VOLT to add OpenType Layout data for glyph substitution
and positioning to a font. I sometimes use TTX, which is a tool to dump and
edit individual tables in a font to an XML text format; this is very useful
if you need to change some setting in a font without completely remastering
from FontLab, passing the font through VOLT, etc.
One piece of
software I use in font development almost every day is the spreadsheet
program Microsoft Excel. Because I am often working on very large fonts with
lots of characters and multiple glyph variants, it is important to plan and
manage the glyph sets carefully. Almost every font development project
begins with a spreadsheet, in which I list and order the glyphs that will be
included, what Unicode characters they should be mapped to, which OpenType
Layout features they are included in, etc.
Which of your typeface designs are you most proud of?
I try not to
be proud, which is certainly easy when I look at some of my older typeface
designs! I do think I’m getting better, though, and I’m pretty pleased with
what I have done recently. I’m particularly happy with the typefaces I am
making for the Society of Biblical Literature, which is an academic
organisation of scholars of Hebrew, Greek and other Biblical languages. I’ve
made Hebrew and Greek fonts for them, and a big multiscript font for
publishing critical texts and other scholarly publishing. This is some of
the best work I’ve done.
Font formats have evolved, from older small formats to
the large Unicode and OpenType fonts. For which phases of this evolution
have you made fonts?
We made
PostScript Type 1 fonts in the mid-90s, but I never liked that format very
much. Having multiple files for each font, which didn’t work on different
platforms, was very frustrating. We started doing more TrueType font
development as the custom font side of our business developed. We’ve been
making OpenType fonts since 1998: almost ten years now.
It seems simpler to create a Latin typeface than an
Arabic one. Why is this? Is the Arabic script naturally more complex?
I’m not sure
that I consider a Latin font simpler than an Arabic one. Because Arabic
re-uses so many basic shapes, differentiated by different numbers and
arrangements of dots, the glyph set can actually be quite small. With Latin,
you have both uppercase and lowercase, and you have the requirements for
italic, smallcaps and other stylistic variants that are very important to
Latin typography.
I think the
reason why there are so many more Latin fonts in the world than Arabic fonts
has more to do with technology and the market than the characteristics of
the scripts themselves.
We know that Microsoft and Adobe had developed the
Unicode character set to include all the world’s languages, including
Arabic, without consulting the owners. Do you believe that Unicode satisfies
the needs of Arabic fonts? Is there any weakness and what are your
suggestions to reactivate them?
Microsoft
and Adobe did not develop Unicode. Unicode is an international standard,
developed by a consortium of industry. There is also a parallel ISO
standard, 10646, which is identical to Unicode in terms of character set,
and through this the national standards bodies have a say in the development
of the standard. [I am a member of the Canadian standards committee for
coded character sets, which is involved in the ISO working group for
10646/Unicode.]
Ironically,
in the case of Arabic, the large number of redundant and completely
unnecessary ligature encodings in the Unicode ‘Arabic Presentation Forms’
ranges, are there precisely because one of the national standards bodies
(Egypt, I believe) insisted that they be included. So it is the ‘owners’ of
this script that have caused the complications of multiple ways to encode
the same Arabic typeform, and introduced a lot of stuff that is contrary to
the Unicode character/glyph distinction.
On the
whole, I think the Unicode encoding of Arabic is pretty good, if one ignores
the unnecessary presentation forms and uses the basic Arabic character
encoding, allowing the font and shaping engine to handle the positional
forms and/or ligatures at the glyph level.
One thing I
would like to see added to Unicode Arabic is
encoding of all basic underlying letters
without dots, and a means to represent the
undifferentiated vowel markings of early
manuscripts. Early Arabic texts are
written without dots or with only generic markings indicating the presence
of a vowel without necessarily indicating which vowel. For scholarly
purposes, it should be possible to transcribe these manuscripts accurately
and without making assumptions about the identity of the letters. Let’s say
there is an initial tooth letterform in such a manuscript; it could be beh,
teh, tteh, nun, or it might be an alif maksura. There are no dots in the
manuscript, so you can’t be sure what the writer intended. If you decide
that the letter is beh, and encode it as such in an computer transcription
of the document, then you are making an editorial decision about the content
of that document, about what it says, rather than transcribing what you see
on the manuscript page. Similarly, if you see a generic mark, such as a red
circle above a letter in a manuscript, and you decide it represents fatah,
you are making an editorial decision about the meaning of the text in the
encoding of the text. This is not a good way to conduct textual scholarship:
the editorial commentary on what an ancient text means should be separate
from the transcription of that text. So that is something I would like to
see possible with Unicode, but I realise that this is a pretty obscure and
specialist requirement. For most day-to-day Arabic text encoding, Unicode is
very good.
Some consider the advantages of Open Type as wish for
fonts to reach the quality of calligraphy. Is it possible for OpenType to
achieve that, to produce something close to Calligraphy? Could OpenType do
that Arabic?
I think
OpenType can take one a lot closer to the calligraphic styles than most
previous font technologies, if that is what one wanted to do. [It should be
noted however, that this doesn’t necessarily mean including hundreds or
thousands of ligatures in a font; in fact, that will almost inevitably break
calligraphic emulation at some point, because you will never have enough
ligatures for every possible combination. Monotype made a nas’taliq font for
phototypesetting many years ago, which had more than 20,000 ligatures in it,
some of them for whole words, and there were still combinations of letters
occurring in text that broke the system.]
I think one
can go a long way toward calligraphic emulation with OpenType, but no, you
can’t go all the way. There are elements of all the traditional calligraphic
styles that are beyond what OpenType is capable of. Well, this shouldn’t
surprise us: it’s a font format, not a scribe. My response to this is to ask
‘If I were a scribe, and I set myself the task of devising a new
calligraphic style within the limitations imposed by this OpenType layout
model, what would this style look like?’ That is, I wouldn’t set out to try
to do a perfect nas’taliq or even a perfect traditional naskh in OpenType,
but I would try to create something new that takes advantage of what
OpenType is good at.
It is an
important font technology, and there should be good new Arabic type designs
that take advantage of it. But if your goal is to really emulate the
traditional calligraphic styles in all their complexity, then you need
something like the Tasmeem layout model that Thomas Milo has developed with
WinSoft.
We know that VOLT, a free program from Microsoft, has
played an impressive role in the development of OpenType fonts. But do you
think that such a program needs to be developed and improved? What are your
suggestions?
VOLT is
being developed and improved. I know the developer at Microsoft, Sergey
Malkin, very well, and although he has lots of other responsibilities he is
quite responsive to requests from font makers for new functionality or
improvements. I attended a meeting at Microsoft a couple of years ago, along
with some other VOLT ‘power users’, and we gave Sergey a lot of feedback
about things we would like to see in future versions. Some of these he was
able to build into the last public version, but some will need to wait until
a more significant overhaul of the program code.
Just this
week, Sergey sent me a new test version of VOLT in which it is possible to
make comments in the individual lookups. This is a very simple feature, but
one that will be very helpful, especially in big and complex projects.
What is the right difference between TTF, TTF1 and
Postscript font, and where each of them used?
[Note: By
TTF1 I think perhaps you mean Type 1, which is a kind of PostScript font.]
Type 1 =
Legacy PS font format, developed by Adobe. No longer made by Adobe.
Supported via either Adobe Type Manager or system rasterisers in a wide
range of current and legacy operating systems. Not supported in Microsoft’s
new Windows Presentation Format (a layout and rendering system with MS
Vista).
Multiple
Master = An extension of the Type 1 format incorporating axis-based
interpolation between two, four, six or eight different designs.
Discontinued by Adobe. Support dwindling.
TrueType =
Originally developed by Apple as a challenge to Adobe’s Type 1 format in the
early 1990s. Licensed by Microsoft, who considerably extended the format
specification. Notable features include a Unicode character map, an
extensible table structure, and advanced ‘hinting’ instruction set for
improved low resolution legibility.
TrueType
Open = Fore-runner of OpenType, developed by Microsoft in the mid-1990s as
an extension of TrueType. One of the goals was to provide support for
complex script layout, and some of the first applications of the format were
to Arabic.
OpenType = A
further extension of the TrueType format, co-developed by Microsoft and
Adobe, building on the work MS had done with TrueType Open. In addition to
the optional OpenType Layout tables for glyph substitution and positioning,
the format may include either a TrueType glyf table or a PostScript CFF
table, meaning that there are different flavours of OpenType fonts.
There are
other font formats, both legacy and current, but this represents the main
development over the past couple of decades.
We hear about the type of fonts called ClearType. What
it ClearType and where is it used?
ClearType is
not a kind of font. It is a screen rendering system for type, developed by
Microsoft. You can view any TrueType or TT-flavour OpenType font within
ClearType, and this rendering system is used by default in Windows Vista,
the new MS operating system. It is designed to enhance the look and
readability of type on LCD monitors of the kind used in laptops and other
flat-screen displays (as distinct from the television-like CRT monitors). It
uses individual red, green and blue sub-pixels to triple the virtual
horizontal resolution of the rendering grid, allowing for better fidelity to
the design of the individual typeface. At the same time, it applies clever
colour filtering algorithms to maintain the density of the letterforms, so
that they retain a strong contrast and remain legible.
Apple and
Adobe also have sub-pixel rendering technologies, but ClearType is sharper
and clearer, with greater consistency of letter density.
Although
ClearType is a rendering system that can be used to display any TrueType
flavour font, the rendering model has certain characteristics which led the
Advanced Reading Technologies team at Microsoft to think that it would be
possible to design typefaces that would take particular advantage of
ClearType. So a few years ago they hired a number of type designers to
produce fonts optimised for ClearType. These are what one sometimes hears
referred to as ‘ClearType fonts’. My Constantia typeface is one of these.
These CT fonts ship with MS Office 2007 and Windows Vista.
Type 1, TTF, then OpenType… What do you believe the
next step future?
OpenType
does a great job for a large amount of things. I think it, or extensions of
it, will be around for quite a while. What is interesting now are the
typographic corners where OpenType cannot reach, the situations in which it
isn’t as good as some other option might be. OpenType is a general font
format, designed to support as many different writing systems as possible,
and it does its job very well in most cases. What I think we will see,
though, is the development of specialised typesetting tools and font
formats/extensions for the non-general cases. Tom Milo’s Tasmeem Arabic
engine, addresses the specific tasks of very high quality Arabic book
publishing, is a great example of this kind of development. It is also very
expensive, which tends to be a characteristic of specialised software.
We have seen designers who were awarded prizes for
Arabic fonts in competitions like TDC2, e.g. Mamoun Sakkal, with whom you
worked on the Arabic Typesetting font for Microsoft. Is this evidence that
the pace of Arabic font development is increasing and becoming more widely
recognised?
I think
there is a lot of interest in Arabic type design these days. Part of this is
due to a general increase in internationalisation, globalised business, and
multilingual communications. Part of it is doubtless due to the political
situation and conflicts in the Middle East: people in North America are a
lot more aware of the Arab world than they used to be, even if they don’t
understand it any better. And part of it is due to the availability of
technologies for working with the Arabic script in widely available software
like MS Word, and the availability of free or reasonably priced programs to
enable people to make Arabic fonts.
The TDC
competition is an interesting case, because they maintain a growing advisory
committee of non-Latin type experts. This is something they had to start a
few years ago, because the number of non-Latin typefaces being submitted
suddenly increased dramatically. They realised that their judges were not
always qualified to make decisions about these entries, so they started this
ad hoc advisory panel, which was a very sensible idea.
It is known that the Roman script has high fixed and
regular ascenders descenders, unlike Arabic font, in which there is much
more variety among the letters. Could this rule of the Latin script be
changed?
You mean
could we make the ascenders and descenders of Latin letters vary more than
they usually do? There are certainly historical Latin writing styles in
which this happens, and also styles in which the ascenders are not straight.
But we have these typographic norms which have been with us for more than
500 years now. You can go against the norms in a fancy display face, sure,
but in a text face it is more difficult because people have expectations and
generally you don’t want to be confusing the reader or drawing attention to
the typeface: you want him to read the text.
When establishing making a bold
font, do you design the regular font first and then design the bold on the
basis of it? Or do you design them together, or use multiple master
interpolation?
Usually I
design the regular font first, and then the bold. I generally use
interpolation in circumstances where, for instance, the client is unsure
just how heavy he wants the different weights. We were involved in such a
project some years ago, in which the client and their design agency wanted
to conduct tests of a range of different weights, so interpolation was used
between a quite light regular and a quite heavy bold, and then they settled
on two intermediate weights.
The general
rule of starting with the regular and then doing the bold doesn’t always
apply, though, and it may vary by script. Our associate designer Tim
Holloway (who has done some very good Arabic types) prefers to start with
the bold when he is working on some scripts, e.g. the Indian Devanagari
script.
What is hinting, and what is your approach to it? Do
you prefer to do it manually, or do you use FontLab’s autohinting?
There are
two different hinting models: one for PostScript fonts and one for TrueType
fonts. Both are concerned with improving the display of type in low
resolution environments. TrueType hinting is more complex and also more
advanced in terms of the level of control that it makes available to the
font maker. But it is important to understand the level of hinting that is
appropriate to different rendering systems, and to determine what level of
hinting is appropriate for a particular font.
It is
possible to set up the autohinting in FontLab to do a really very good job
for certain rendering systems. For example, the ClearType rendering system I
discussed earlier often needs only relatively light hinting, and autohinting
can do a good job for this. But black and white bitmap rendering requires
much more control over how the outline is mapped to individual pixels, and
autohinting is not really sufficient.
Many Arabic calligraphers think that fonts are an
abuse of the Arabic script that ignore the sound foundations of calligraphy.
What is to blame for this view? Is it because there are many bad fonts made
by amateurs?
I think
there is a sad lack of dialogue and cooperation between calligraphers,
experts in the script, and typographers and people making fonts.
Unfortunately, I don’t see this changing, despite things like the recent
Kitabat conference in Dubai, because the graphic design culture in the Arab
world is very much attuned to western design education and practice, and
doesn’t take seriously the challenges of typographic representation of the
Arabic script. The calligraphers and other experts in the script see this,
and they are not very interested in typography as a result.
This
isn’t simply a problem of ‘amateurs’ making bad fonts. There is a problem
with professionals too, many of whom seem to view Arabic typography in
context of Latin typography. There are projects like ‘Typographic
Matchmaking’, being organized by the Khatt foundation, in which Arab type
designers are matched with Dutch type designers to do what? — to make Arab
versions of Dutch typefaces, i.e. to make Arab types based on the
characteristics of Latin types. In itself, there are some interesting
aspects to this project — the Dutch types in question are good ones, and the
Arab and Dutch designers are apparently collaborating —, and I look forward
to seeing the results. But in a big way it seems to me wrong-headed, because
there are really basic issues about the typographic representation of Arabic
letterforms that are not being addressed by starting from Latin typefaces.
And these issues are not going to be addressed so long as Arabic design
culture and education is so completely oriented toward western design
culture. At the Kitabat conference in Dubai, there were presentations by
people like Uğur
Derman, Nabil Safwat and Mohamed Zakariya—real experts in the Arabic script
who understand that different historical styles and their place in the
development of Arabic writing—but what they had to say seems to be taken
only as history: it isn’t seen as relevant to modern Arabic typography.
Ironically,
if a western type designer sets out to produce fonts that mimic Arabic
letterforms, he is likely to be accused of Orientalism and exoticism, or of
cultural appropriation. It seems to me that contemporary Arabic design
suffers from Occidentalism; it is entranced by the apparent sophistication
of western design.
What new projects are you working on? Do you have
ideas for new Arabic type designs?
I have some
ideas for new Arabic designs, but nothing in development yet. I’m very busy
most of the time, so unless I have a client for something, it tends not to
progress. But this isn’t necessarily a bad thing: ideas can percolate in the
mind for a long time, and may improve as a result. One thing I want very
much to do is to make an Arabic OpenType font that does not use any ligature
glyphs, only contextual variants. There are a couple of fonts already out
there, such as SIL’s
Scheherazade
and Lateef types in the linear neo-naskh style, but I would like to apply
the concept to a different kind of design. Unfortunately, the Cursive
Connections OpenType Layout feature that I would also want to use is not
very widely supported yet — non in InDesign ME, for example — so I think I’m
going to wait a while before working on this.
There is a tendency in the Arab world towards the
production of fonts that simulate the shape of Latin font. Do you think this
trend will damage Arabic typography? In what direction should Arabic
typography be travelling?
See above.
Arab type designers, whether they acknowledge it or not, are the inheritors
of one of the world’s great calligraphic traditions, and also of great
literary and philosophical traditions that are intimately linked to that
calligraphic tradition. It seems to me perverse to ignore that tradition and
go and try to squeeze your writing system into the alien dress of another
culture. And let me be very clear: I don't think this means that Arabic type
design should be trying to emulate classical calligraphic styles. There has
to be forward movement, and there have to be new developments and new ideas
about the script. But they should come from within the culture and not be
cut off from the tradition, and they certainly should not be dependent on
someone else’s typographic and design culture.
This is not,
by the way, an observation I would limit to Arabic. There is a great allure
to western graphic design, because of its high production values — the
glossy magazines and the endless series of design annuals, competitions,
etc. —, its established educational programmes, and its celebrity designers.
In many respects, graphic design is a western discipline, and people seem to
have trouble adapting the critical and analytical aspects of this
discipline, which I think are the most valuable parts, to their own
cultures, and instead adopt the superficial visual aspects of western design
in place of their own aesthetic traditions.
What
surprises me is that I, a non-Arab and unable to read Arabic, look at the
classical calligraphic styles or at vernacular Arab lettering, and I see all
sorts of possibilities for new development. But the range of type design
being done seems to be very limited: most of it is display or headline
types, and almost every single Arab text type is a variation on the
simplified, linear neo-naskh style. There is very little else, and insofar
as people are trying to do something new they are looking to Latin type for
inspiration, not to the Arabic lettering tradition.
Who do you think are the good designers of Arabic
fonts? Are there promising designers, and what is your advice to them?
To be
honest, I have not seen a lot that inspires or impresses me. I think the
best people working in the field are people of an older generation, like
Mamoun Sakkal and Tim Holloway. I don’t think the younger generation of
Arabic type designers, whose enthusiasm is certainly admirable, have found
their way yet. A lot of the types I see are variants on ideas that have
already been explored in the past fifty years: the headline faces with the
very heavy baseline strokes, for example. And very few people are trying to
design text faces for continuous reading. This is really the biggest
problem. You can’t build a typography out of display faces: they are only
the façade, there has to be a foundation which is text typography.
But I’m
always, always happy to be shown to be wrong about such things. Maybe there
are designers out there with whose work I am not familiar, and who are
addressing these important issues of Arabic typography. I was impressed by a
student project by Titus Nemeth, an Austrian studying type design in
England, who made a strong Arabic text face.
There is a lot of font piracy in the Arab world. What
are your thoughts on this issue, and how do you think it can be overcome?
Font piracy
is a global problem, but obviously it is worse in some places than in
others. Lack of easy access to legitimate licensing sources can be a
problem, yes, but it doesn’t justify piracy. Global font pricing, which does
not take into account local economies, is also a problem, but again it
doesn’t justify piracy. These are significant problems, and the people who
make and sell fonts and the people who use and buy fonts need to work
together to solve these problems, to improve accessibility to legitimate
sources and to find ways to price fonts for local and regional markets.
For me,
these issues are a bit more theoretical than they are for many of my
colleagues, because I’m not involved in retail licensing. The fonts are make
are mainly distributed by my clients, and they are responsible for
protecting the intellectual property.
There is an
aspect of ‘fonts cheating’ that should be carefully considered, and which
relates to my earlier comments. This is unauthorised localisations of a font
for a particular market, sometimes involving adding support for another
script. So, for example, taking a Latin typeface from a North American font
company like Emigre and creating an Arabic version of it — or a Greek
version, or a Thai version, whatever — without permission from the original
designer and publisher. Why does this happen? In large part it is because
the local graphic design culture has manufactured a demand for such
typefaces, and local clients have come to see these sorts of types as
progressive and desirable. So it seems to me that the solution to the
problem of unauthorised localisations is directly related to the issue of
westernisation. And I think it is disingenuous of some of the people who
make such unauthorised fonts to say ‘Oh, but it is what the customers want,
and if we don’t do it someone else will,’ because the demand is being
created by the graphic design culture of which those people are a part.
If customers
do want localised versions of western fonts, then font makers need to go to
the western companies who own the intellectual property and/or licensing
rights, and ask for permission, enter into contracts, and do it
legitimately. And history has shown that this is good for the people who
make such fonts and for the local design environment. If you look at
companies like ParaType in Moscow or Cannibal in Athens, making authorised
Cyrillic and Greek versions of Latin types, you can see how not only has
this been good for those companies but also how it has improved the
environment in which fonts are licensed and used in those countries
 |
