[Trennmuster] »Allgegenwart«

Guenter Milde milde at users.sf.net
Fr Apr 27 16:56:49 CEST 2012


On 27.04.12, Werner LEMBERG wrote:

> >> Die Bindungsstärke der Trenn-Operatoren ist (von schwach nach stark):
> >> 
> >>   == || -- = |  -
> > 
> > We must make an unambiguous list or priorities for these
> > operators. At the moment I have
> > http://xml.resource.org/cgi-bin/xml2rfc-dev.cgi?mode=html&url=https://raw.github.com/PanderMusubi/hyphenation-patterns/master/hyphenation-patterns.xml#rfc.section.3

> Uh, this doesn't represent what we have in German.  There are two
> different concepts which you've incorrectly mixed up.

>   (1) The hyphenation weight.  The higher the weight, the more
>       preferable the hyphenation point is.  Currently, we don't
>       directly support that.  Instead, we are going to categorize the
>       hyphenation points (from a grammatical point of view), and this
>       is in most cases identical to assigning a weight.  In general,
>       the categories `=' and `|' have a higher weight than `-', and
>       `==' and `||' have a higher weight than `=' and `|', etc.

>       Uncatogorized hyphenation points `·' have the same weight as
>       `-'.

>   (2) Demerits for hyphenation points.  The more `.' are appended to a
>       hyphenation point, the worse the hyphenation point is, from
>       various points of view.

> > Feedback on priority is welcome.

We have, however a concept of operator precedence and associativity in order
to express binding differences in multiple-part words like

 Fachbuchverlag = Fachbuch + Verlag = (Fach + Buch) + (ver + lag)

The following table summarises the operator precedences from highest
precedence (most binding) to lowest precedence (least binding).

  - and .  normal hyphenation point, discarded hyphenation point
  | 	   hyphenation after prefix       (un|klar)
  =        hyphenation between sub-words  (Rat=schlag)
  ||       hyphenation after prefix       (Ur||groß=vater)
  ==       hyphenation between sub-words  (Arm=band==uhr)

The | operator groups left to right.
(BTW:I would like to use the + character instead of |.)

The above example becomes:

 Fach=buch==ver|lag

Currently, the use of double (triple etc) hyphenation markers is optional
and rare as this information cannot be used with TeX.

Günter



Mehr Informationen über die Mailingliste Trennmuster