It is not one of Wikipedia's policies or guidelines; rather, its purpose is to explain certain aspects of Wikipedia's norms or practices. It may reflect varying levels of consensus.
These are characters of the Arabic abjad: أبجدية عربية. If instead, you see a bunch of little boxes or question marks, then you need to install fonts with the appropriate characters
These are characters which appear in related languages:
The word العربيةal-ʿarabiyyah, "the Arabic [language]". From top to bottom:
1. Left-to-right (incorrect);
2. Right-to-left but not joined (incorrect);
3. Right-to-left and joined (correct).
The biggest problem for incorporating Arabic language text into the English languageWikipedia is that Arabic flows right-to-left while English and other languages written in Latin script flow left-to-right. Worse, the numerals shared by the two languages don't have as strong directionality as the letters, sometimes causing seemingly inexplicable glitches. (See Help:Hebrew for an example in Hebrew script.) This can be fixed by using the template {{lang}} (click on it to learn how to use it) or by the Unicode left-to-right mark (LRM) U+200E at the end of the Arabic text to signal that the following English text should be read left to right. The LRM can be placed using an HTML character identity of either the hexadecimal or decimal value: ‎ or ‎ or ‎. In some cases, it might be possible to just rephrase or move the text around so that the more strongly directed text follows the Arabic text. This avoids the need of the LRM altogether. Arabic script can be incorrectly rendered on a system not supporting Arabic.
Transliteration
For Wikipedia's guidelines on how to transliterate languages using Arabic and Persian script in articles, see Wikipedia:Manual of Style/Arabic.
One suggestion is to give the word written in Arabic the first time it appears in an article, followed immediately by one romanization, then using that romanization consistently through the rest of the article.
DIN 31635 standard is one of the widely used schemes, with one sign for each Arabic letter.
Another transliteration standard is ALA-LC, which uses fewer additional diacriticized letters but adds more digraphs.
Unless the article or section is about a local dialect or names, the pausal (without case endings إﻋﺮابʾiʿrab) pronunciation of the literary Arabic (الفصحىal-fuṣḥā) is preferred. Usually, the literary Arabic pronunciation is preferred when the article deals with something related to Islam. See Arabic and Islam.
The case ending may be appropriate when a name uses ʾIḍāfahإضافة (here: possession). Whenever ʾiʿrāb is used, make sure that you drop superfluous auxiliary constructions such as a in the article (a)l-, e.g. عيد الفطرʿīd al-fiṭr or ʿīdu l-fiṭr.
