Urdu is the first language of over 70 million speakers and a second language to over 160 million people, mostly in Pakistan and India. Due to both design and engineering challenges, there were only a few typefaces available for the Urdu language in the preferred Nasta'liq writing style of the Arabic writing system. The scarcity of Nasta'liq typefaces was so dire that some newspapers and other printed materials were written by hand. Urdu speakers still sometimes write emails and messages in Romanized Urdu (the transliteration of Urdu in the Latin alphabet). 

Why are there so few Urdu fonts?

If there is such a large population of Urdu speakers, you might wonder why there are so few typefaces for this population. One reason is that technology companies have not seen Urdu-speaking communities as important markets and have not invested resources to develop Nasta’liq typefaces. 


Let’s go back in history and review some features of the Urdu language, the Nasta'liq script, and technological problems to find some more reasons.


Writing the Urdu language


To a great extent, the Urdu language is similar to Hindi. Speakers of both languages can understand each other. Yet the writing systems used for each language are very different. Hindi is written from left to right in the Devanagari writing system, while Urdu is written from right to left in the Nasta'liq style of the Arabic writing system. When calligraphers used to write Nasta’liq by hand, they traditionally used a reed pen. 


Container with reed pens

Reed pens used for Nasta’liq calligraphy, Photo credit: Borna Izadpanah


Nasta'liq was initially developed to represent the Persian language but has also been used to write other languages, including Ottoman Turkish, Kurdish, Pashto, Sindhi, Persian, and Uyghur. Today, most Arabic typefaces used for those languages in printed books, newspapers and digital platforms use a simplified form of the Naskh style. However, most Urdu books and newspapers are still printed with Nasta'liq type, which takes up more vertical space than simplified Naskh and has a more cursive style. 


Two lines of Urdu text in black

1. Nasta'liq

2. Naskh
Transliteration: Ab apne aap se bhi chup gai hai An Urdu poetry verse (“Now she is hiding from herself too”) by the poet Zehra Nigah, set in the Gulzar Nasta’liq typeface (top) and in a simplified Naskh typeface Markazi (bottom). 

Urdu cascades


Latin and many other writing systems can be written with unconnected letters and are more horizontally aligned on the baseline. In contrast, "most Arabic script writing styles are written in a cascading format in which characters are joined on multiple horizontal levels.” explained Borna Izadpanah, the designer of the Gulzar typeface. 


Typesetting technologies have been primarily designed for non-joining and horizontally aligned scripts like Latin. The cascading nature makes Nasta'liq typeface design, production, and composition more software reliant than Naskh. 


Designers have to find a solution for kerning the cascading forms and the precise positioning of diacritic dots for vowel marks to avoid clashes and overlaps.


Two words in black, vowel marks in red, arrow showing vowel marks moving downwards

The word “computer” کمپیوٹر in the Gulzar typeface. From right to left:
1. The diacritical dots used in the word are shown in their default positions.

2. Since this word has two dotted letters consecutively, the leftmost set of dots must be moved downwards to avoid a collision.


Cursor moving right to left as characters change


From right to left. This animated GIF shows the composition of characters with kerning and also how dots and marks move around to avoid clashes and overlaps. Characters change shapes depending on the letters that come before and after. (Set in the Gulzar typeface.)
Translation: “How to type in Urdu on your computer”.
Transliteration: Apne computer pe Urdu mein kaise type karein

The bari-ye vowel

Designers have to pay special attention to the bari-ye vowel (ے, /eː/ /ɛː/) which frequently appears in Urdu. Even though Urdu is written from right to left, the bari-ye sweeps backwards (left to right) and underneath preceding letters, especially impacting dotted letters which precede it. 

Circled numbers above text in black

From right to left:
1. Unconnected form of bari-ye
2-5: Sequence of characters in words ending with bari-ye. The arrows show the downward shift of diacritical marks to avoid clashes and overlaps with the bari-ye.
(Set in the Gulzar typeface.)


Shapeshifting

In Nasta'liq, letters change their form based on their position in the word (initial, medial, final, and unconnected) and depending on the neighboring letters before and after. 

Urdu letters in black on beige background

From right to left: The letter پ (“P” sound) in initial position (circled and in red) taking different shapes based on the next letter. (Set in the Gulzar typeface.)


Latin has something similar to shapeshifting. In some Latin writing styles, letters change depending on their position and connection to the letters around them. 


Lowercase combinations of o and n, o and r, and o and x


The angles of the connecting strokes of the lowercase “o” change depending on the following letter. Set in Style Script by Robert Leuschke.


One of the main differences between the Latin and Arabic writing systems is that Latin can be written in both connected cursive and unconnected forms, but the Arabic letters (apart from a few exceptions) must be written in a connected form. Additionally, most Arabic letters alter their shape depending on the preceding and following letters. 


Technical difficulties

Urdu Nasta'liq is poorly supported on digital platforms because the writing system requires software-intensive support. Some text-rendering systems/software are not designed to accommodate it. 


For a long time Urdu speakers have had to use existing simplified Naskh typefaces with distorted Urdu characters or find a way to hack operating systems to install their preferred Nasta'liq typefaces. However, this is changing. 


اردو ورژن


Making of Gulzar


To give Urdu speakers more typeface choices, in July 2022, Google Fonts added Gulzar, a new Nasta’liq Urdu typeface.

Four lines of white text on a black background

1. Gulzar Nasta’liq

2. Gulzar Latin 


All human beings are born free and equal in terms of rights and dignity.” in Urdu (Gulzar Nasta’liq) and transliterated in Latin (Gulzar Latin)


​​Simon Cozens, Dr. Borna Izadpanah, and Dr. Fiona Ross conducted their own research and consulted with Urdu language specialists in Pakistan and the United Kingdom to create the Gulzar Urdu Nasta’liq typeface project. Gulzar means “flower meadow” in Urdu. 
Izadpanah is a native Persian speaker from Tehran, Iran. He learned Nasta’liq as a model for Persian handwriting in primary school.

“Designing a digital Nasta’liq typeface was my long-held dream,” Izadpanah stated. As the principal Gulzar designer, he conducted the preliminary research and drew the glyphs

Ross was familiar with the Urdu language and type design from her language studies and earlier work on two Nasta’liq typefaces, Sheeraz and Qalmi, for which Linotype acquired a patent.

To make a modern digital font based on the Urdu flavor of the Nasta'liq style, Izadpanah studied the proportions, stroke modulation, and character features in calligraphy manuals and a collection of lettering specimens.

“The outlines in Gulzar are carefully drawn. Small details like terminals are designed to look clear and consistent in different sizes. We also made several revisions to improve the curves and achieve smooth connections,” explained Izadpanah. 

Two Urdu letters in white on black backgroundSharply cut terminals (in green). 
Letters (from right to left): س - Seen,
Letters (from right to left): س - Seen,
ح - Hey

Five versions of the letter Ain
 

From right to left: The team started by studying various calligraphic forms of the letter ع to understand better the particularities of the Urdu Nasta'liq and its conventions (a calligraphic reference on the right). The design went through four stages of development until the team agreed on the final version of the unconnected "ain" letterform on the left. The Ain letter makes the /ɑː/, /oː/, /eː/, /ʔ/, /ʕ/, /∅/ sounds. 

Engineering challenges.


To make it easier to design Gulzar, font engineer Simon Cozens created two open source tools. Simon Cozens was the Gulzar font engineer.  Since he didn’t speak or read Urdu, he embarked on self-study of the written language. Urdu is written from right to left. Like other languages using the Arabic writing system, the same letter in Urdu has different forms depending on its position in the word (initial, medial, final and unconnected).

“I gained familiarity with the script by carefully examining Urdu poetry, religious texts, and other materials. First, I learned to recognize the rhythm and overall feel of the design. Then I associated individual letters with one another, until finally I was able to recognize them in their different contextual forms (initial, medial, final, and unconnected),” explained Cozens.


Cozens worked with the designers to set up a method for creating a set of glyphs for the font. In the Nasta’liq style, each joining letter must connect smoothly to the one on its left and/or right with the correct stroke thickness and direction, following the calligraphy rules for the reed pen, the conventional writing tool used for Nasta’liq.


Three sets of Urdu letters with numbers and text labels


These are combinations of letters, not words.


Basic shapes of the letters without dots (from right to left):

ب  Beh in (green)

 ھ , Heh Doachasmee (white)

س, Seen (orange)


1: The default form of initial Beh (green) ends with an upward stroke when connecting to the letter Seen (orange).

2: When a calligrapher writes the sequence Beh + Heh (green + white), the pen stroke does not turn upwards (as shown in the middle) but must continue to flow smoothly downwards into the middle stroke of the Heh (white).

3. Final result for the Beh + Heh combination: The glyph used for Beh (green) must be changed to a different variant that reflects the pen’s movement into the Heh (white).



In Gulzar, there are 20 different possible variations for the initial form of the letter Beh, depending on the letter to its left. Each of those 20 variations of the letter must connect smoothly to the next letter.



Two Urdu letter combinations with numbers

                                                

Basic letter shapes without dots: 

1. ک, Kaaf (green)

2. ب,  Beh (orange) 
3. ھ,  Heh (white) 
4. ں, Noon ghunna (purple)

Both sequences start with the same two letters (Kaaf #1 in green and Beh #2 in orange). Yet both the Kaaf and Beh letterforms are different in each sequence. Different Beh variants were selected based on the letter to its left. Since the Beh variants have different stroke thicknesses, appropriate variants of the Kaaf are chosen to match the Beh’s stroke thickness.


In Gulzar, there are 9686 potential letter combinations. There are various initial or medial forms of 29 letters of the Urdu alphabet that could potentially connect to one of 334 different variants to the left of the letter. 

In some of the previous Urdu fonts, designers had to calculate and draw all the possible thousands of letter combinations. This was an arduous and costly task. Few font foundries wanted to take on such a large financial undertaking requiring a large staff and budget. 


To facilitate the Gulzar design process, font engineer Simon Cozens created two open source tools. To avoid the need to hand draw the letter combinations and to simplify the management of letterform variants and combinations, Cozens created the Nasta’liq Connection Editor, a Glyphs font editor extension using OpenType technology. He drew on the innovative patented work that Ross and her colleagues, Michael Fellows and Timothy Hollway, had done to make the Sheeraz and Qalmi fonts without the need for such extensive character sets.  


The Nasta’liq Connection Editor extension allows the type designer to easily choose the intended form of each letter from a spreadsheet (left) and see how the two characters looked in the preview panel (right). (The extension can also be used for any Arabic writing style, not just Nasta’liq.) 


Spreadsheet with six columns and 13 rows

1. The preview panel shows the two selected letters (Ain ع and Fe ف ) in one of their medial variants.
2. Selection of Ain ع variant AINm4 and Fe ف variant FEm9.
The left and top rows have names assigned to each letter followed by their position (i = initial), (m = medial), (f = final) and (u = unconnected). The number 4  in the columns is a code for a glyph variant.


“OpenType technology provides the ability to ‘sew’ together different glyphs so that the top of one glyph becomes the new baseline for the next. The designer puts a mark, or ‘anchor’, in each glyph where the previous and next glyphs should be attached,” explained Cozens.

Four Urdu letters with numbers1. ج “Jim”  2. م “Meem” 3. Exit anchor  4. Entry anchor

From the right: An initial form of the letter ج “Jeem” (red) has an exit anchor which is connected to the entry anchor on the medial form of the letter م “Meem” (green). The composed form of the two characters is shown on the left.

Cozens also created Fez the (“Font Engineering made eaZy”) system for handling typographically complex scripts such as Nasta’liq Urdu. The font engineer describes the script's behavior using either the Fez or Python programming language, and the system determines the necessary rules.

For example, Cozens used Fez to find the many variations that are needed for the vowel bari-ye as it frequently appears in Urdu words. Even though Urdu is written from right to left, the bari-ye sweeps backwards (left to right) and underneath preceding letters. For more information about the bari-ye, see “Why there are so few Urdu fonts?”

“I wrote a plugin which measures the length of the tail, measures the width of each glyph, and then produces a set of combinations,” explained Cozens. By having the computer do the work of evaluating the different possible letter sequences, Cozens found a way to manage the complexity that comes with engineering a Nasta’liq font. 






This artwork shows the letter Jim ج from Gulzar and the terminology and mathematical method used in Urdu to describe the different parts of the letter and its conventional proportions.

There used to be few Urdu typography choices. Now, with more Urdu fonts to use as reference models and new open source tools, Urdu type design is democratizing and diversifying as more people can make Urdu fonts to meet readers’ requirements.

Gulzar is open source and has a GitHub project where people can find the font files and post issues and comments to contribute. Gulzar is open source and has a GitHub project where people can find the font files and post issues and comments to contribute. 


Cozens, Izadpanah, and Ross see their work as a contribution not just to Nasta’liq type design, but also to Urdu language publishing and culture. “The Gulzar project aims to encourage type designers to make more Arabic typefaces that address the linguistic requirements and stylistic preferences of various languages that use the Arabic script.” said Izadpanah and Ross. 





Gulzar banner 


Gulzar is available on Google Fonts. 

Currently, Gulzar fully supports the Arabic, Persian, and Urdu languages. Gulzar is the first Urdu Nasta'liq typeface that includes an original Latin counterpart. The Latin typeface, designed by Dr. Alice Savoie, matches the typographic color of the Nasta’liq Urdu design. The Gulzar Latin character set contains sufficient characters with diacritic marks to transliterate Urdu into Latin.

Bios: 


Simon Cozens is a font engineer based in Gloucester, UK. He specializes in OpenType layout of complex scripts, and operates Corvel Software, which enables type designers to support the world’s languages.

​​Borna Izadpanah is a Lecturer in Typography at the University of Reading, UK, where he received a PhD, and an MA in Typeface Design. His doctoral research explored the history of the early typographic representation of the Persian language. Izadpanah has received numerous prestigious awards for his research and typeface design, including the Grand Prize and the First Prize for Arabic Text Typeface in the Granshan Type Design Competition, a TDC Certificate of Typographic Excellence, and the Symposia Iranica Prize for the best paper in Art History.



Dr. Fiona Ross specializes in type design and typography primarily for Arabic, South Asian, and Thai scripts, She has a background in languages and Ph.D. in Indian Palaeography from the School of Oriental and African Studies (SOAS), where she had studied Sanskrit with Pali. She is Professor in Type Design at the Department of Typography and Graphic Communication (University of Reading, UK). She works as a consultant, type designer, author, and lecturer and has co-curated seven international exhibitions. Ross received the SoTA Typography Award (2014) and the Type Director’s Club Medal (2018). She leads the Women in Type research project.


Resources:

Gulzar website (more information on the methodology and technology behind Gulzar)

Gulzar GitHub directory

Fez program

Nasta’liq Connection Editor

Linotype patent for the Qalmi and Sheeraz fonts (Fellows, Holloway & Ross in 1989) Posted by the Google Fonts team