[Trennmuster] Trennstellkategorien

Guenter Milde milde at users.sf.net
Do Mär 22 00:15:49 CET 2012


On 21.03.12, Werner LEMBERG wrote:

> > So in general - is used for hyphenation within stems or word, and =
> > is used for hyphenation between stems and words in compounds.

> And we also want to use `|' to mark word prefixes.

> > Aditionally we will add some <a|b> before the = the assist looking
> > up stems when they are conjugated. There only a will be used and <
> > and |b> can be discareded. This might even remove the need of a
> > two-stage format.

> Since `our' use of `|' can never occur in `your' use within
> `<...|...>', I don't see a problem here.

However, if the construct <a|b> is intended for alternatives, I suggest
using the forward slash / (solidus) instead because the "regexp
alternatives character" '|' is used as hyphenation marker in many 
dictionaries and hence easily misunderstood.

The solidus is used to mark alternatives in German text (like
"Vater/Mutter" for "father or mother"). This is why it is used for
alternatives in the German wordlist, too: 

#   Ab-fa{ll/ll-l}a-ger

(Abfallager becomes Abfall-lager if split at the word boundary.)

#   Dru{ck/k·k}er

(Drucker becomes Druk-ker if split between c and k.)

#   er[b-/-b]recht

(Erb-recht and er-brecht are both valid hyphenations (depending on
the semantic of the word). Because of this, automatic hyphenation is
suppressed at this point.)


> > As discussed with Werner, for now, we will use ~ for hyphens that
> > are always available in words (hyphenated or not).  However, this is
> > not always the preferred place for hyphenation.

> OK.

> > What would you recommend to indicate that hyphenation on = is in the
> > following example preferred over hyphenation of ~
> >
> >   signaal-ruisverhouding;sig-naal~ruis=ver-.houd-ing
> >
> > Simply by the order of:
> >   1: =
> >   2: ~
> >   3: -
> >   4: -.

> Including `our' use of `|', I would write it as

>   signaal-ruisverhouding;sig-naal~ruis=ver|.houd-ing

In our syntax, the marker for "inauspicious" hyphenation points '.' is
orthogonal to the categories, i.e. it can be combined with all of them.

This is why I would write:

  sig-naal~.ruis=ver|houd-ing
  
and weight the hyphenation points e.g.

  =   3
  |   2  
  -   1
  ~.  1  (downgraded one point from the default for ~: 2)


Günter



Mehr Informationen über die Mailingliste Trennmuster