Right to left in Rutba Office, and PDFs that keep Arabic, Urdu and Hebrew as text

In September our list of what Rutba Office could not do included right-to-left layout for Arabic and Hebrew. A document written in Arabic, Urdu, Persian or Hebrew would open, but its paragraphs ran from the left, which makes the page hard to read. A run of releases in October changed that, and the PDF was the hardest part.

In Documents

Since 1.29.4, a paragraph Word marks as right to left, or whose style does, runs from the right margin. Its indents, its hanging indent, its borders and its drop cap are mirrored. Home → Paragraph has left-to-right and right-to-left buttons to turn a paragraph either way, and aligning a right-to-left paragraph writes its alignment mirrored in the file, as Word does.

1.29.6 took it further. In a right-to-left paragraph the ruler and its tab stops count from the right margin. A table can run from the right, with its first column on the right-hand side, from Table Layout → Right to Left. A section with columns lays them out from the right.

In Worksheets and Presentations

• Page Layout → Sheet Right-to-Left (1.29.6) puts column A on the right, the way an Arabic or Hebrew workbook reads.
• A slide’s paragraph can read right to left (1.29.6), drawn from the box’s right edge with its indent and bullet on that side. Home → Paragraph turns one either way.
• View → View Direction (1.29.3) turns the whole Presentations window right to left.

A PDF that keeps the words as words

Getting right-to-left text onto the screen is half the job. Saving it as a PDF is harder, because the PDF has to carry the finished shapes. Arabic letters change shape with their neighbours, and a right-to-left line holding numbers or English words has to be put in order before it can be drawn.

Since 1.29.8, when a document’s PDF needs letters its built-in fonts cannot draw, Rutba Office uses a TrueType font on your computer whose licence allows embedding, and embeds only the letters the pages use. Arabic letters take their joined forms, including the ones Urdu and Persian add, and lam and alef become their ligature. Each right-to-left line is put in drawing order by the Unicode bidirectional algorithm, so numbers and Latin words inside it still read correctly and brackets face the right way. Every letter is mapped back to its character, so the words can be copied out of the PDF, searched and read aloud.

The same release lays a right-to-left paragraph from the right margin in the PDF, with its list number in the right-hand gutter, and draws bold or italic for a font that has no bold or italic face, the way Word does. 1.29.10 added a check that holds the PDFs of worksheets and presentations to the same standard for Arabic, Hebrew and Greek.

Marking the language

Review → Language (1.29.1) marks the selected words as a language, or as not to be checked at all, in the way Word keeps it. A right-to-left language fills its own slot in the file, so Urdu marked as Urdu is proofed as Urdu when the document is opened in Word. And the suite’s spelling check no longer sends a word with no Latin letter in it to the English dictionaries, which could only call it wrong.

What is not done

• The interface in Urdu came a few releases later. At 1.29 the interface was English only, with every word of it going through a message catalogue that counts by each language’s own plural rules. 1.38 used that catalogue to add Urdu, with the whole frame turned right to left, and 1.40 added Arabic, followed by Hijri dates and kashida in print; the translations still await a native speaker’s review. Rutba Office in Arabic and Hindi has the detail.
• Spelling is checked in English only. Words marked as another language, and words with no Latin letter in them, are passed over rather than called wrong.

For the general question of writing right to left in any office software, Word and LibreOffice included, see right-to-left documents in Arabic, Urdu and Hebrew on the Rutba Office site.