[tex-hyphen] dehyph-exptl v0.20 RC1 (experimental German patterns)

[Mojca Miklavec]

On Mon, Jun 15, 2009 at 20:44, Norbert Preining wrote:
> On Mo, 15 Jun 2009, Mojca Miklavec wrote:
>> I think the question wasn't about "experimental" patterns - we already
>> have them in hyph-utf8 since the beginning and they are loaded in
>> XeTeX/LuaTeX by default.
>
> Ah, ok. But not into latex or etex based formats, right?

They are loaded in xetex, xelatex, luatex, lualatex, but not in tex,
latex, pdftex and pdflatex.

>> Apart from that there's a special package that should enable switching
>> to timestamped German patterns (don't know the details).
>
> But that should be done on package level, and not on hyphenation, i.e.
> \language level, right?

Because there's no chance to load patterns on demand, one needs to
modify language.dat and then the package selects the proper \language.
(An ugly solution.)

>> For 8-bit we didn't dare to do the change (even though I don't believe
>
> Which change? From 8bit to UTF8?

No. We didn't dare to use these new patterns for (pdf)(la)tex, so for
those we still use the ancient ones. (I was afraid that Germans would
start complaining about different line breaks.) We are not talking
about the difference in traditional and reformed patterns, but the
difference in patterns generated ages ago (dehypht.tex, dehyphn.tex)
and the ones generated by the WL & friends last year
(hyph-de-19xx.tex). So we basically use 4 different patterns now +
those timestamped + Swiss ones.

>> means multiplication, not addition :) :) :)
>
> Oh hell, yes we German-speakers are horrible. I would vote for
> abolishing all old variants, but I would get slaughtered in some German
> regions (and from many German newspapers) sticking to the old form.

I don't have a problem keeping two sets of patterns (traditional and
reformed). I only find it a bit stupid to keep two variants of each,
but I don't dare changing anything unless somebody else asks for that.
(Last year we have promissed that the new architecture would not
change the behaviour.)

>> I was talking about adding or not adding Swiss German (pattern file
>> most probably bearing the name similar to hyph-de-ch-1901, right?).
>
> Ok, well, Swiss ... funny ... ;-)))))
>
> I would propose that we standardize "Weanarisch" as the standard German
> in all the German speaking countries.

I propose Windisch statt Weanarisch. Klingt doch besser :)

Mojca