#11 Combination of symbols 'ij' in Dutch

[Kęstutis Biliūnas]

In the Dutch language, some words contain the 'ij'
combination. In the deu-nld, eng-nld, fra-nld, nld-deu,
nld-eng and nld-fra dictionary packages, this combination
is entered as the single character 'ij' which makes up
some inconvenience for the users. The Dutch users affirms
that substitution of all occurrences of 'ij' to 'ij' not
contradict to the official language rules (see Debian BTS
bug #381054).

In the Debian dict-freedict packages already made these
changes. I am sorry that I don't produce the patch for
these changes, because Debian not use in the meanwhile
the real source files (.tei files) for creating the
binary dict-freedict packages.

The maintainer of Debian dict-freedict packages -
Kęstutis Biliūnas

[Piotr Banski]

Just wondering: perhaps we should keep both versions for matching?

[Piotr Banski]

Not touching this today, sadly, merely noting the pattern to be introduced.

In e.g. fra-nld, in the entry for "Abyssin", the equivalent should probably just become:

<cit type="trans">
<quote>Abessijn</quote>
<quote>Abessijn</quote>
</cit>
(for searches within definitions)

And in nld-fra, the headword should become:

<entry>
<form>
<orth>Abessijn</orth>
<orth>Abessijn</orth>
</form>

[Sebastian Humenda]

Hi,

danke dafür. Ich bin nach wie vor unsicher, wie wir das lösen. Ich mag den
Vorschlag nicht die Einträge zu doppeln. Ich weiß aber auch nicht, wo ich
ansetzen müsste, damit ich die Regel überschreiben kann, die das
Standard-Stylesheet zum Umwandeln überschreibt. Außerdem müsste ich schauen,
wieviel man vom XSL ändern müsste und ob man teile davon quasi bei den
Wörterbüchern mit einfügen könnte.

Bin ich jetzt eigentlich der Bug-Eigentümer?

Viele Grüße
Sebastian
--
Web: http://www.crustulus.de | Blog: http://www.crustulus.de/cgi-bin/blog.cgi?de-0
Freies Latein-Deutsch-Wörterbuch: http://www.crustulus.de/freedict.de.html
FreeDict: Free multilingual dictionary databases - http://www.freedict.org

[Sebastian Humenda]

Hi,

well, the same in English, shorter and just for the records ;-).

I am not sure how to resolve this. On option would be to override the
transformation process in the makefile and to use a custom-tailored XSL style
sheet. On the other hand am I no expert in style sheet programming and I am not
sure even whether it's possible to just override parts of an included style
sheets.

Sebastian

[Michael Bunk]

I like Piotr's suggestion, but I'm fine with either option currently on the table.

Overriding rules in included stylesheets is a basic feature of XSLT. But the Makefiles currently don't allow this.

Another thing to consider is http://en.wikipedia.org/wiki/Unicode_equivalence. We could entirely leave it to the dict server to perform matching in the "canonical domain". But on the other hand I'm not aware that any dict server performs any canonicalization.

[Sebastian Humenda]

I like Piotr's suggestion, but I'm fine with either option currently on the table.
Doesn't piotr's suggestion force every future editor of the databasesto add two
orth's whwen adding a word containing ij? I think it could lead to
inconsistencies.

I thought about overriding the style sheets at the point where <orth> is parsed
and also the Makefile. Should pe possible and should throw away the need to
always add doubled entries.</orth>

http://en.wikipedia.org/wiki/Unicode
I am not sure what you want to point out. Do you mean that the dictd daemon
itself should remap the "ij" symbol to the string "ij" and vice versa, or do you
mean there are maybe more "ij"-code points?