Hyphenation of special ligatures in historic German texts

[AlbrechtHilmes]

This is the translation of the German thread “Trennung von speziellen Ligaturen in historischen Texten”
It was extended:

In the uploaded document Hypenation of special ligatures_03.odt (35.7 KB) the font AntiquaHistoricGerman-Normal.ttf.odt (29.5 KB) is used (rename the file “AntiquaHistoricGerman.ttf.odt” to “AntiquaHistoricGerman.ttf”)

This font includes the following OpenType features:

calt: long-s (ſ) / round-s (s) automatic
liga: ligatures double-s and ck

This font reflects features of historical German texts. The so-called broken fonts - or blackletter fonts - (Schwabacher, Gothic and Fraktur) have the same peculiarities.
The rough rule for that is: at the end of a word a round-s (s) has to be used and at the beginning or in the middle of the word a long-s (ſ) has to be taken.

Unfortunately, Libre Office has problems separating double-ſ and ck:

“ſſ” is usually hyphenated (e.g. Wasser (water) vs. Waſ-ſer) and the ck becomes a k-k when hyphenated (e.g. Drucker (printer) vs Druk-ker).

Libre Office’s automatic hyphenation separates these words incorrectly.
Examples and more details in the attached document “Hyphenation of special ligatures_03.odt”.

Questions

Is there a way to make the above hyphenations automatically in Libre Office implemented (if possible by adapting the OpenType programming)?
So is there a special glyph that can be used in the OpenType programming in the lookup tables to detect the hyphenation?

Herzliche Grüße
Albrecht Hilmes

[mikekaganski]

Hyphenation rules are independent of fonts. What you need is a special hyphenation “dictionary”, for another language (because the language you use is different from what is modern German); and that’s up to a contributor interested in that to happen.

[AlbrechtHilmes]

Dear Mike Kaganski,

thank you very much for your plausible answer.

It should be noted that there is already a corresponding dictionary with hyphenation aids:
https://extensions.libreoffice.org/en/extensions/show/german-de-de-1901-old-spelling-dictionaries

Unfortunately, it uses the long s (ſ) as Unicode character #383 in Fraktur mode. This character is not on the keyboard, which makes typing very complicated. The document is not easily searchable.

The long s/round s automation with OpenType solves all these problems.

Is there no way to intercept a hyphenation in LibreOffice Writer via OpenType programming?

Best regards
Albrecht

[mikekaganski]

No - OpenType features (and fonts in general) are orthogonal to the hyphenation, as I said already.

[AlbrechtHilmes]

Dear Mike,
thank you for your reply.
Is it true that you have now changed the word “orthogonal” to “independent” in the first answer? My problem in understanding this was the term orthogonal, which is used in geometry and vector calculus. In vector algebra, orthogonality means the linear independence of two vectors or linear equations. Thank you for that clarification.
In which order becomes “Wasser” then “Waſſer” and finally “Was-ſer”? Obviously the hyphenation in the last step influences the OpenType programming again.
What should the entry in the hyphenation dictionary look like for the two words “Wasser” and “Drucker”?

Best Regards
Albrecht

[mikekaganski]

OpenType could maybe know how to handle “s” after hyphen; that is up to a font maker. Indeed, then the three components - the hyphenator (splitting a word), the text layout language (telling the line limits to the hyphenator, and putting resulting parts to different lines), and the OpenType (which should get the information from the upper levels, that it handles a part of the word split between these characters) - would need to work together.

But it would be much simpler to avoid this complexity, and make the hyphenation dictionary to produce the correct parts after the split, taking the rules into account. One doesn’t need to type such a character; typing “Wasser” at the end of a line should produce the “Was-ſer” inside the hyphenator, with the outer world still seeing the word as “Wasser” for the purposes like search.