Arabic, Modern Standard

Joining forms

Because the Arabic script is 'cursive' (ie. joined-up) writing, letters tend to have different shapes depending on whether they join with adjacent letters or not (see cursive). In addition, vowels can be represented using different characters, depending on where in a word they appear.

In scripts such as Arabic, several characters have no left-joining form. In what follows we'll use the characters ي and د to illustrate shapes. The former can join on both sides, but the latter can only join on the right.

Left-joining glyphs are commonly called initial; dual-joining are called medial; and right-joining are called final. Glyphs that don't join on either side are called isolated. However, these glyph shapes can be found in various places within a single word.

Word-initial characters usually have initial glyph shapes (eg. 064A ). However, characters that only join to the right will use an isolated glyph shape (eg. 062F ). Furthermore, words beginning with a vowel are always preceded by a vowel carrier, which is normally ا (eg. 0627 06CC or 0627 064E ).

Word-medial characters will typically join on both sides (eg. 064A ) but those that only join to the right will use a final glyph (eg. 062F ). However, if either of those is preceded by another character that only joins to the right, the glyph shapes rendered will be initial (eg. 064A ) and isolated (eg. 062F ), respectively.

Word-final characters will typically use a final glyph shape (eg. 064A and 062F ). However, if the previous character joins only to the right, they will use isolated glyph shapes (eg.064A and 062F ).

In all this contextual glyph shaping the basic shapes used for a character can vary significantly in a script like Arabic. This also includes some characters that only have ijam dots in certain contexts.

Letters

Combining marks

Numbers

Punctuation

Symbols

Other

Vowel sounds

The diphthong aj is colloquially pronounced more like ej, however in this page we will continue to transcribe it phonetically per the official pronunciation.

Consonant sounds

	labial	dental	alveolar	post- alveolar	palatal	velar	uvular	pharyngeal	glottal
stops	p b	t d				k ɡ	q		ʔ
ejective		tˤ dˤ
affricates				d͡ʒ
fricatives	f v	θ ð	s z	ʃ ʒ		x ɣ	χ ʁ	ħ ʕ	h
ejective		ðˤ	sˤ zˤ
nasals	m		n
approximants	w		l ɫ		j
trills/flaps			r

Modern Standard Arabic covers many territories, most of which have their own dialects or languages, and these tend to influence the local pronunciation of Standard Arabic. In the chart above, we remain conservative, only mentioning variants that tend to apply to the standard pronunciation. For a slightly more detailed set of notes, see Wikipedia.

p and v are sometimes pronounced by some speakers for foreign words, such as باكستان فيروس

Sometimes alternative letters are used for such words (see foreign).wp

Although most dialects include it as a phoneme, ɡ is only used in Modern Standard Arabic as a marginal phoneme to pronounce some dialectal and loan words.wp

The sound ɫ occurs as a phoneme in a handful of loanwords, though not in all pronunciations. It also occurs in the name اللّٰه

The sound, d͡ʒ is used in Algerian, Hejazi, Najdi, Iraqi, and Gulf regions, whereas ʒ is used in Moroccan, Tunisian, Egyptian, Levantine, and Israeli regions. In both cases, the sound is written using ج.wp,#Local_variations_of_Modern_Standard_Arabic

ظ is pronounced zˤ in some regions, rather than ðˤ. 

Tone

Arabic is not a tonal language.

Structure

The following notes on structure are taken from Wikipedia.wp,#Phonotactics

[C1] [S1] V [S2] [C2 [C3]]

Arabic syllable structure consists of an optional syllable onset, consisting of one or two consonants; an obligatory syllable nucleus, consisting of a vowel optionally preceded by and/or followed by a semivowel; and an optional syllable coda, consisting of one or two consonants.

The following restrictions apply:

Onset

C1 can be any consonant, including a liquid (l r). The onset is composed only of one consonant; consonant clusters are only found in loanwords. Sometimes an epenthetic a is inserted between consonants.

Nucleus

Includes S1 V S2.

Coda

C2 and C3 can be any consonant.

Vowel summary table

This table summarises basic vowel to character assignments, in both vowelled and unvowelled forms.

Each table cell shows word-initial, word-medial, and word-final forms from right to left. The glyphs shown are illustrative; alternative shapes may occur (see joining_forms).

For additional details see vowel_mappings.

Because the sukun is also dropped in non-vocalised text, where a mater lectionis remains it only implies a vowel location, since it may either represent a consonant or a glide.

In word-initial position vowels are attached to ا. However, more often than not, a hamza is also attached to the alef to indicate the glottal stop. Although this actually constitutes a consonant plus vowel, in unvocalised text the alef (plus any hamza) signals the location of a vowel. The table therefore shows these maximal combinations.

The sounds eː, o, and oː are only used for transliterations of foreign words, and are spelled identically to iː u, and uː, respectively.

The letter ى, used as an alternative for a final aː, is a dedicated vowel character (see vletter).

Post-consonant vowels

Vowels are written using a mixture of combining marks and letters in vocalised text, but normally the diacritics are not used (and so it is difficult to accurately read the text unless you recognise the consonant patterns). However these diacritics and other phonetic information can be written where needed, and are regularly used for Qur'anic texts, dictionaries, educational materials, and where the pronunciation needs to be made clear.

In vowelled text, the Arabic language uses 3 basic vowel diacritics, but 4 more and 1 letter are occasionally also used. Long vowel locations are marked by matres lectionis (consonants indicating vowel locations).

Standalone vowels

Standalone vowels at the beginning of a word usually incorporate ا.

In principle, Arabic has very few true standalone vowels, since vowels are nearly always preceded by a glottal stop or other consonant, and in Arabic, unlike many other languages, the glottal stop is phonemic and a distinct letter of the alphabet.

Vowel sounds to characters

Consonant summary

This table summarises basic consonant to character assignments.

For additional details see consonant_mappings.

Basic consonant letters

The main Unicode Arabic block contains 153 letters, with 77 more in the extended blocks. As shown in the previous section, only a small subset of those are used to write a given language. The others represent special characters added to the repertoire for one or other of the many languages for which the Arabic script is used.

The vast majority of letters represent consonants. A few represent long vowels.

Click on each letter in the lists below for more details and for examples of usage.

Special letters

The following describe basic letters that require some more lengthy description.

Repertoire extensions

Letters for foreign sounds

The following characters are not part of the standard Arabic language set but are occasionally used to represent foreign sounds. Two of these are letters borrowed from Persian/Urdu.

Examples:

ڤيينا

پاريس

چاكارتا

Other letters

The following characters also have the general property of Letter, but are less commonly used for modern Arabic language text.

06A2 and 06A7 are alternative forms that are used in Northwest Africa. 08B2 is used for Berber.

0671 is described in the section hamza. Whereas many of the above letters with diacritics decompose in Unicode Normalization Form D (NFD), this letter does not.

0640 is used to stretch words for simple justification, or to make a word or phrase a particular width, or as a form of emphasis. For more information see justification.

Arabic definite article

The pronunciation of ال (alif followed by lām) varies when it represents the Arabic definite article.

The lām is not pronounced if it precedes one of the following characters, but instead the following sound is doubled, eg. السلام علیکم

These are called 'sun letters' in Arabic. The other letters are 'moon letters'.j,32

The alif is also not pronounced if the preceding word ends with a vowel or h. It is, however, written.j,32

Onsets

No special features are used for syllable onsets.

Finals

Final consonants in Arabic are simply written using ordinary consonant letters. No special features are used, other than the sukun in vowelled text (see clusters).

Consonant clusters

Consonant clusters in Arabic are simply written using a sequence of consonant letters.

When text is vowelled, 0652 can be used over a consonant to indicate that it is not followed by a vowel sound. Like other vowel diacritics, this is typically not used in modern text, unless it is necessary to clarify pronunciation, eg.

حمراء

أنف

Consonant length

The diacritic 0651 doubles the value of the consonant it is attached to, which is phonemically significant in Arabic, eg. تجّار

Like the short vowels, it, too, is not often used, although sometimes it appears when vowel signs don't.

When both shadda and kasra are attached to the same base consonant, a common, though not universal, practice is to display the kasra below the shadda, rather than below the base consonant, eg. مُمَثِّلْ Some fonts, such as Amiri, don't do this. (See also gpos.)

Consonant sounds to characters

Honorifics

Characters in the Arabic Presentation Forms blocks that are not just for compability use include the following. Click on the characters in the list for more information.

See also presentationForms.

Ijam symbols

Other characters in the Arabic Presentation Forms blocks that are not just for compability use include the following symbols that can be used for pedagogical purposes. In educational materials there is sometimes a need to show pictures of the dots and marks used to distinguish Arabic characters, particularly the ijam. These code points provide for that use case. They are never used as combining marks, nor in composition with Arabic letter forms, but are simply symbols.

See also presentationForms.

Ligatures

The combination 0644 0627 is always written as a ligature. The underlying code points are, however, preserved. The form of this ligature that joins to the right is ‍لاand unjoined it is لا

Observation: When diacritics are used with this ligature, they sometimes appear to be over the ALEF, rather than over the LAM, eg. قليلاً This would require a typing order that is different from the spoken sequence.

Other combinations of characters are likely to also ligate (see gsub). The number of ligatures in text typically depends on the font used, but ligation can also be used as a device to manage justification, in which case it needs some degree of manual control

Formatting characters

Modern Arabic text makes use of a relatively large set of invisible formatting characters, especially in plain text, many of which are used to manage text direction. Descriptions of these characters can be found in the following sections:

• shapingcontrols
• directioncontrols

Presentation Forms

The code points in the Unicode blocks Arabic Presentation Forms-A and Arabic Presentation Forms-B provide positional forms of Arabic letters and ligatures. They should not be used for ordinary text. Those code points are provided for compatibility with legacy code pages, and have (compatibility) character decomposition mappings. Normally, Arabic text should be written with code points from the main Arabic block and its extensions; positional forms are dealt with by the font and rendering algorithms.

However, there are some exceptions to this rule, which are listed here. These characters are not included in the Unicode repertoire for compatibility but may be used in Arabic texts, in their own right.u,398-400 They normally don't have character decomposition mappings. (See also Arabic ‘presentation form’ exceptions.)

The useful code points include the following:

• honorific word-ligature letters, see wordLigatures
• honorific word-ligature symbols, see honorifics
• two ornate parentheses, see phrase
• a set of ijam dots for pedagogical use, see ijamSymbols

Honorific combining marks

In addition to the honorifics described earlier, the basic Arabic block has a small number of corresponding combining marks. Click on the characters in the list below for more details.

Hamza & precomposed characters

Unicode support for the various uses of the hamza is complicated.u,384 In general, the Unicode Standard recommends to use 0654 in combination with a base character. However, there are a few exceptions to consider.

Codepoint sequences

When typing and in storage, combining marks always follow the base character they are associated with.

Digits

See type samples.

A set of arabic-indic digits are typically used in Middle Eastern and Gulf countries, whereas North African countries tend to use European digits. In neither area, however, is one digit style used exclusively.

The Unicode bidi_class property for these native digits is Arabic_Number, which makes them behave differently from ASCII digits, and differently from the set of extended digits used for Persian, Urdu, etc. For more information, see expressions.

Arabic script has its own number separators, which are used in Arabic language text when the non-European digits are used. They are ٫ and ٬.

Arabic also has its own characters for ٪, ؉. The ASCII % and ‰ are also used.

The CLDR standard-decimal pattern is #,##0.###. The standard-percent pattern is #,##0% or #,##0٪.c

See also expressions about directional implications for handling expressions or sequences of numbers.

Extended-Arabic digits. Still in the basic Unicode Arabic block, as mentioned, there is a second set of digits in Unicode for use in languages such as Persian and Urdu.

۰␣۱␣۲␣۳␣۴␣۵␣۶␣۷␣۸␣۹

The glyph shapes are typically different for 3 of the digits (although not always the same 3 digits) in Persian, Urdu and Sindhi.

Arabic
Persian
Urdu
Sindi

Currency

Unicode has a character for the rial: ﷼.

Managing text direction

Unicode provides a set of 10 formatting characters that can be used to control the direction of text when displayed. These characters have no visual form in the rendered text, however text editing applications may have a way to show their location.

202B (RLE), 202A (LRE), and 202C (PDF) are in widespread use to set the base direction of a range of characters. RLE/LRE comes at the start, and PDF at the end of a range of characters for which the base direction is to be set.

In Unicode 6.1, the Unicode Standard added a set of characters which do the same thing but also isolate the content from surrounding characters, in order to avoid spillover effects. They are 2067 (RLI), 2066 (LRI), and 2066 (PDI). The Unicode Standard recommends that these be used instead.

There is also 2068 (FSI), used initially to set the base direction according to the first recognised strongly-directional character.

061C (ALM) is used to produce correct sequencing of numeric data. Click on the character name, and see also expressions for details.

200F (RLM) and 200E (LRM) are invisible characters with strong directional properties that are also sometimes used to produce the correct ordering of text.

For more information about how to use these formatting characters see How to use Unicode controls for bidi text. Note, however, that when writing HTML you should generally use markup rather than these control codes. For information about that, see Creating HTML Pages in Arabic, Hebrew and Other Right-to-left Scripts.

Expressions & sequences

This section is about sequences of numbers, rather than a sequence of digits. Sequences of numbers are sets of numbers separated by punctuation or spaces, such as 10–12–2022. Sequences of digits, such as 123, in Arabic text run LTR automatically.

A sequence of numbers used to express a range of values generally runs right to left in the Arabic language (and languages using the Thaana or Syriac scripts), whereas for Persian language text (and in Hebrew, N’Ko or Adlam scripts) it runs left to right.

This also tends to apply to expressions such as 1 + 2 = 3.

fig_range_ar shows Arabic text which is right-to-left overall, containing an ASCII-digit numeric range that is also ordered RTL, ie. it starts with 10 on the right and ends with 12 on the left.

In Persian, however, the sequence would generally run LTR, so 10 would be on the left, and 12 on the right. The underlying order of the characters that make up the expression, and the order in which they are typed, remain the same. (Click on each figure to see the underlying character sequences.)

However, the preferred order for a sequence of numbers may also depend on the context. For ISBN numbers, telephone numbers, and so forth, a left-to-right sequencing is likely to be preferred.

The default direction for a sequence in an application that implements Unicode fully will depend on:

1. the digits used (ASCII, Arabic or Extended Arabic),
2. whether or not the sequence is preceded by Arabic script text, and
3. the separators used.

Writing styles

Arabic orthographies can be grouped into a number of writing styles, some of which are more common for particular languages, while others can be used interchangeably for the same language. Sometimes the variations are adapted to usage, for example book text vs. inscriptions; sometimes the variants reflect regional, cultural or stylistic calligraphic preferences.

The different styles include Naskh, Nasta'liq, Ruq'a, Thuluth, Taʻlīq, Kufi, Diwani, Maghribi, Kano. The examples in this page use a naskh writing style. For a brief introduction to font styles, with examples, see Text layout requirements for the Arabic script.

The naskh writing style is the most prominent style for the Arabic language, and has become the default form of Arabic language content in most contexts. It has clearly distinguished letters, which make it easy to read, and can be written in small sizes.

The ruq’ah writing style was designed for use in education, in official documents, and for every-day writing. It is known for its clipped letters composed of short, straight lines and simple curves, as well as its straight and even lines of text. It is a functional style of writing that is quick to write and easy to read. It also doesn’t extend baselines, like a naskh font does. In 2010's rebranding of Amman a ruq'ah font family was created to serve as an italic face. Monotype has an interesting article on the development of ruq'ah.

The nasta’liq writing style is the standard way of writing Urdu and Kashmiri, and is also often a preferred style for Persian text. Key features include a sloping baseline for joined letters, and overall complex shaping and positioning for base letters and diacritics alike. There are also distinctive shapes for many glyphs and ligatures.

The kano writing style is a common way of writing Hausa in Nigeria in the ajami script, and like other East African writing it is based on Warsh (Warš) forms, which incorporate Maghribi characteristics. Some sources describe an alternative Hafs (Ḥafṣ) orthography, used with hand-written adaptations for the newspaper Al-Fijir.

The kufi writing style is the original style used for the Koran, but is not used for newspapers or official content today. However, it is used in modern content for logos and other stylised applications.

Cursive script

See type samples.

Arabic script is always cursive, ie. letters in a word are joined up. Fonts need to produce the appropriate joining form for a letter, according to its visual context, but the code point used doesn't change. This results in four different shapes for most letters (including an isolated shape). Ligated forms also join with characters alongside them.

The highlights in the example below show the same letter, ع, with three different joining forms.

Most Arabic script letters join on both sides. A few only join on the right-hand side: this involves 4 basic shapes for Modern Standard Arabic.

ء doesn't join on either side.

Context-based shaping & positioning

Context-based shaping

See just above for shaping related to cursive joining.

In all but the most basic fonts, glyph shapes are highly variable for Arabic letters. For example, fig_mishafi_ka shows a wide variety of shapes produced by default in the Mishafi font for ك when followed by various letters.

Teeth letters

A good font will constantly change the shape of glyphs slightly so as to create a more aesthetically pleasing, and in some cases a more easily readable, flow.

teeth_letters shows an example where the same word is displayed using different fonts.a,#h_teeth_letters The font on the left applies rules to distinguish the letter bases clearly. Note, in particular, that although there are 3 letters which are repeated, none of those letters uses the same shape twice.

Special joining forms

In more traditional fonts, you will also often see the join between certain characters above the baseline. Compare the highlighted character joins in fig_raised_joining, showing the same sequence of letters but with joins above vs. along the baseline. (The first font is Mishafi, and the second Scheherazade New.)

But actually a good font will typically have a range of shapes and placements for a given letter, depending on the adjoining letter. This is illustrated in noon_joining_forms.a,#multi-context-joining

Characters within a word may also combine vertically in certain groupings. See the example in vertical_joining.

Context-based positioning

When vowel or shadda diacritics are used they can be placed in different positions, according to the context.

When both shadda and vowel signs are combined with a base character, a more complicated set of rules may be applied. Depending on the font used, some vowel diacritics may be placed relative the shadda diacritic, rather than relative to the base character.

Letterform slopes, weights, & italics

See type samples.

Case & other character transforms

Arabic has no case distinction.

However, as mentioned in numbers, Arabic sometimes uses ASCII digit glyphs and other times uses local digit glyphs. Some fonts and authoring applications allow you to choose which glyph shapes to use for the same underlying characters.

Arabic fonts may also have alternative shapes for glyphs, which can be turned on in certain circumstances. For example, some fonts have a set of swash forms for certain characters, which can be used for justification, or just for effect.

Word boundaries

Words are separated by spaces.

In Arabic, small words like 'and' (و) are written alongside the following word with no intervening space (eg. الجامعات والكليات means 'universities and colleges', but there is only one space). Such small words are handled typographically as part of the word they are attached to.

Graphemes

Larger typographic units

One potential complication is that fonts often render sequences of characters as ligated forms. The ligated forms are a font-specific feature, whereas grapheme clusters are based on code point sequences: some fonts may display the same sequence of characters without a ligated form. Most applications tend to move character by character through the text, producing situations like the cursor position in fig_gc_ligature_cursor.

This approach allows for easy deletion or insertion of any component of a ligated form.

Many Brahmi-derived scripts are segmented by units that incorporate more than one grapheme cluster, for operations such as in-word line-breaking, justification, letter-spacing, and initial letter highlighting. It is not clear whether such typographic units are needed for Arabic language text, since there is usually no hyphenation, and no initial letter highlighting, and letter-spacing and justification follow quite different rules, extending the baseline. It may be worth looking at the very rare examples of vertically-set lines with upright Arabic letters to check whether ligatures like lam-alif or others are kept together. (My expectation is that lam-alif is not split, but the others may be.)

Apart from lam-alif, where one can expect a rule to apply consistently, another issue is that an application that wants to keep ligatures together as a single unit would have to be aware of the rendering behaviour of the particular font in use, since some fonts have ligatures for a given code point sequence and others don't. There is no way of deriving this information from the code point sequence itself, since that is always exactly the same.

Phrase & section boundaries

Arabic language uses a mixture of ASCII and Arabic punctuation. Other languages using the Arabic script may use different punctuation, such as the full stop in Urdu.

Arabic language text also uses ‐ U+2010 HYPHEN, –, and —.

Mirrored characters

The words 'left' and 'right' in the Unicode names for parentheses, brackets, and other paired characters should be ignored. LEFT should be read as if it said START, and RIGHT as END. The direction in which the glyphs point will be automatically determined according to the base direction of the text.

The number of characters that are mirrored in this way is around 550, most of which are mathematical symbols. Some are single characters, rather than pairs. The following are some of the more common ones.

(␣)␣<␣>␣[␣]␣{␣}␣«␣»␣‹␣›

Quotations & citations

See type samples.

Two different styles of quotation mark can be found in Arabic language texts. When quoted text appears within quoted text different characters are used, though usually of the same type. (Of course, depending on ease of input, quotations may also be surrounded by ASCII double and single quote marks.) Spacing inside the marks is optional.wqm

Because they are mirrored, when using the guillemets, LEFT should be read as if it said START, and RIGHT as END. The guillemet shapes are typically rounded, as shown in fig_quotation_marks.

Unlike the guillemets, these quote marks are not mirrored during display. As a result, LEFT means use on the left, and RIGHT means use on the right.

Sometimes these styles are mixed in the same text. The example in fig_quotation_marks uses both in a single sentence.

Emphasis

Emphasis can sometimes be expressed by stretching the baseline of one or more words. See the section on justification below for more information about baseline stretching.

Abbreviation, ellipsis & repetition

Arabic uses ….

Text decoration & other inline features

Underlines & overlines

Underlines and overlines in Arabic text are usually further from the baseline than they are in Latin text. This is because the Arabic letters extend further from the baseline, and because there are also sometimes diacritics beyond those long extensions. Typically, the line will be drawn so that it is further from the baseline than any other glyphs reach.^§

In some cases, however, while still keeping the line further from the baseline than in Latin text, typographers don't clear the glyphs. In this case, the lines usually skip the ink of the other glyphs.

When skipping ink it is important to avoid leaving very short remnants of the line between glyphs, since these may look like dots or diacritics.

The Qur'an tends to use overlines, rather than underlines.

Line breaking & hyphenation

Lines are normally broken at word boundaries.

They are not broken at the small gaps that appear where a character doesn't join on the left.

Text alignment & justification

See type samples.

Arabic script justification can be implemented using a number of different techniques, which ideally are applied in combination. These include:

• Adjustment of inter-word spaces.
• Micro-adjustments to the intra-word spaces that occur where characters don't join with the following letter.
• Use of wider glyph forms or swash forms for certain characters.
• Stretching the baseline connectors between letters, which we'll refer to here as kashida.
• Application of ligated forms to reduce the width of certain words.

(In hand-written manuscripts it is also possible to find instances where the letters the would appear at the end of the line are squeezed above the last word in the line, or hang into the margin.)

The application of the various techniques is generally subject to rules governing the frequency and location of use of particular methods. Rules can differ by writing style – for example, elongation is not normally used at all for ruq'a fonts. Where baseline stretching is applied, the rules for what can be stretched, and how much, are complicated. Unlike space-based width adjustments, baseline extension is not a question of simply adding equal-length extensions across the line. The rules tend to differ across orthographies, and eminent typographers of the past also had their own preferred or idiosyncratic rules.

The baseline extension character ـ is sometimes suggested as a way of producing justification by extending the baseline, however when a browser window is resized, or when new text is added near the start of a paragraph, lines wrap differently and all the places where tatweel would be needed have to be recalibrated. Thus tatweels only work for static text with fixed dimensions.

Better quality justification systems stretch glyphs, rather than adding baseline extensions. This dynamic stretching of glyphs is often called 'kashida'. In some typesetting systems, such as InDesign, the stretching can be produced automatically without the need for tatweel characters. InDesign has controls to vary the preferred length of the extensions.

Note that the result of the automatic justification in fig_kashida_justification is different from that in the newspaper clipping. For one thing, the kashida effect is only applied once per word (but is applied to most words). The rules determining which combinations of characters receive baseline stretching, and the extent of that stretching also differ.

InDesign also allows fonts to substitute long swash variants for certain characters, which soak more some of the horizontal space.

Well justified text would apply a mixture of swash characters, space stretching, kashidas and ligatures to achieve a visually appealing and effective justification. Also, the baseline stretching in fig_kashida_justification is flat. A more advanced system would instead produce elegantly curved kashidas more like handwritten text.

Text spacing

See type samples.

Spaces are not added between characters, with the exception of micro-spacing during justification, which is applied to word-medial letters that don't join to the left. On the other hand, the baseline within words is often stretched.

It is very common to see baseline stretching in modern Arabic text where a word or phrase is stretched to fill a particular space, eg. the Arabic tag line (الابداع المتجدد Creativity renewed) below the word Lexus in fig_lexus is stretched to be the same width.

The amount of stretch is determined by the area that needs to be filled, but the distribution of the baseline elongations and their length may be affected by choices about which letter pairs best support elongation. Unlike general justification, the fit-to-width nature of this text spacing means that window width ajustments, author edits, etc. are less likely to impact the stretched text. The text boundaries would need to change with any font size changes, though, in order to maintain the effect.

Flat line baseline elongations are very common, which raises the likelihood of success by using tatweel characters (unlike full justication). However, if curved baseline elongations are wanted, like in fig_kashida_curvilinear, some more sophisticated mechanism is needed.

Observation: Text that is stretched in this manner very often has multiple kashidas per word. This is perhaps understandable, given that usually only a small number of words are involved.

In some cases, it may be that elongation of words is driven by stretching the distance between letters rather than matching an external template, for example to express emphasis or prolonged sound. However, as for justification, this is not normally based on an even amount of stretch between all letters, as letter-spacing tends to be in other scripts.

Baselines, line height, etc.

The alphabetic baseline is a strong feature of Arabic script on the whole, since characters tend to join there. This is not always the case: for example, some adjacent pairs or ligatures have joins above the baseline, and initial letters in some fonts may start slightly above the baseline, but for most cases it remains a strong feature.

The nastaliq writing style, on the other hand, uses arrangements of joined glyphs that cascade downwards from right to left, and ressemble a strongly sloping baseline.

However, even writing styles with an ostensibly flat baseline may, in good quality fonts, draw words on a slightly slanted baseline, or multiple baselines, as shown in fig_word_baselines.a,#multi-level-baselines

Characters within a word may also combine vertically in certain groupings, as mentioned in the previous section.

Line height and multi-script positioning. Even without the deviations from the baseline described above, the ascenders and descenders of Arabic letters tend to travel further from the baseline that is usual in Latin script text. Allowances for this need to be made for line height settings on a page, but also it can be problematic when combining Latin and Arabic text on the same line using different fonts for each.

If the Arabic font supports the needed Latin letters, the font design will already take into account the relative sizes of the letters, and their placement relative to the baselines of each script. If different fonts are used, though, it's important to match the baselines and harmonise the font sizes used.

Arabic letters have ascenders and descenders that tend to be longer than the Latin ones. fig_baselines shows ascenders and descenders for Arabic letters in the Scheherazade New font. With the addition of diacritics above and below the letters, the line height needs to be significantly higher than for Latin script text.

Counters, lists, etc.

You can experiment with counter styles using the Counter styles converter. Patterns for using these styles in CSS can be found in Ready-made Counter Styles, and we use the names of those patterns here to refer to the various styles.

The Arabic language uses 1 numeric and 2 fixed styles. Wikipedia lists 2 more styles: an old maghrebi sequence and the hijai sequence.

Styling initials

It is possible to find cases where Arabic enlarges and styles the first character at the beginning of a paragraph, but it is quite rare.

Observation: It is not clear whether it is appropriate to maintain the joining forms of the initial letter and the following letter. A good proportion of the examples seen have the initial letter in a box, in which case it appears to be in isolated form. For further discussion see this thread and this one.

General page layout & progression

Arabic books, magazines, etc., are bound on the right-hand side, and pages progress from right to left.

Columns are vertical but run right-to-left across the page.

Grids & tables

Tables, grids, and other 2-dimensional arrangements progress from right to left across a page.

Forms & user interaction

Form controls should display Arabic text from right to left, starting at the right side of the input field. Form controls should also usually be arranged from right to left.

fig_form shows some form fields from an Arabic language web page. Note the position of the labels relative to the input fields and the checkbox, mirror-imaging a similar page in English. Note also that the input text in the first field appears to the right of the box.

The position of a scrollbar should depend on the user's environment, not on the content of a page. A non-Arab user viewing a web page in Arabic shouldn't have to look for the scroll bar on the left side of the window. In a system that is set up for an Arab user, however, the scrollbar can appear on the left.