The most significant new features in LibreOffice 26.8 concern writing systems, and it is worth explaining what has been done and, above all, why it has been done, as there is an important decision behind the developers’ work.
What has been done
Writer now automatically detects the paragraph direction when opening or pasting documents or plain text, without requiring the user to set it manually.
Spaces at the end of a line are positioned according to the paragraph direction rather than that of neighbouring characters, which corrects a long-standing issue with how line breaks are handled in mixed Arabic and Latin texts.
Object resize handles now work correctly in right-to-left CJK documents and in vertical documents.
Bidirectional control characters are displayed alongside other formatting marks, so that invisible directional substitutions can be both viewed and edited.
The semantics of paragraph alignment have been revised. New documents use justified alignment by default rather than left-justified, and changing the paragraph direction no longer overrides the explicitly set alignment. This is a small change with a significant impact, as a document written in Arabic and one written in English can now share a template without conflicting.
In Calc, when typing text from right to left into an empty cell, the direction of that cell is now set automatically.
LibreOffice Math has gained functions and operators for the N’Ko and Adlam alphabets, along with left-pointing vector arrows and harpoons for right-to-left formulas, without which mathematical notation in these two alphabets cannot be composed correctly.
Finally, the names of fonts registered in languages other than the interface language are now recognised, so that a document created in a Japanese or Russian installation is displayed using the fonts chosen by the author rather than substitute ones.
Why this is unusual
N’Ko is a writing system devised in 1949 by Solomana Kanté for the Mandingo languages of West Africa. Adlam was devised in the late 1980s by two teenage brothers, Ibrahima and Abdoulaye Barry, for the Fulani language, which until then had been written using adapted Arabic or Latin scripts – an inadequate solution. Both scripts run from right to left, are used by populations numbering tens of millions of people, and have been added to Unicode over the last twenty years.
Neither language represents a market. The native speakers concerned do not, as a whole, constitute a customer base large enough to justify the time spent on development based on a commercial analysis. No tender process, anywhere in the world, has led to the completion of this work. No corporate client has requested it. No competitive pressure has made it necessary.
The contributor who added support for N’Ko and Adlam to LibreOffice Math is a volunteer.
Much of the work on bidirectional text was carried out by a developer employed by The Document Foundation, which is funded by donations.
The renaming of the CJK text grid option, so that it correctly referred to the genkō yōshi handwritten grid used in Japanese typesetting, was carried out by a member of TDF’s staff.
Native support for OpenType font variations – which is of enormous importance for alphabets whose typography cannot be approximated by modifying the dimensions of a Latin character – was provided by a type designer working independently of any company.
A fundamental difference
Every software vendor must decide which users are worth serving: if the software is a product, the decision is made on the basis of a business case, and languages that cannot pay for themselves fall by the wayside.
A foundation does not have this constraint, and does not need to ask itself whether an alphabet justifies its development costs, as there is no revenue stream against which to make this choice. The work is carried out because someone with the relevant expertise has deemed it worthwhile, and there are no reasons to prevent it.
This is the practical difference between software managed by a foundation and that managed by a vendor, and it is evident in the release notes. LibreOffice is available in 120 languages for this very reason: the number is a consequence of the governance model rather than a marketing objective.
This explains why the argument in favour of open-source software in public administration is not a question of cost, because a public body has obligations towards citizens who do not represent a commercial interest: users of minority languages, users of assistive technologies, and people living in regions that no vendor prioritises.
Software developed on the basis of a business case will only serve these citizens if doing so is profitable. Software developed without a business case can serve them, as a matter of course and in line with the principles of open-source software.
