[freedict] Re: German dictionary

  • From: Einhard Leichtfuß <alguien@xxxxxxxxxxxxx>
  • To: freedict@xxxxxxxxxxxxx
  • Date: Wed, 11 Aug 2021 17:44:15 +0200

Hi Julius,

On 10/08/2021 22:29, Julius Hamilton wrote:

Did you check the latest version of the eng-deu/deu-eng dictionaries? The
current version is significantly enhanced, compared with the version that we
had for years.

I believe so. It says it has about 400,000 headwords. Is that professional
grade?

The dictionary is far from complete. However, it should cover almost all common words and a large share of less common ones. I use it as my primary dictionary and only rarely have to resort to others. Common commercial dictionaries are generally more comprehensive, though.

Please have a look at the TEI header of each dictionary that you are
interested in. The sources are at <https://freedict.org/downloads/>.

Thanks. What is the TEI header? I saw dictionaries for download at that
link. I guess I should open the dictionary directory and read a README
file, for example, or the XML file?

The main FreeDict format is TEI XML. All other formats are generated from that. The XML files therefore contain the raw TEI header.

You seem to be interested in deu-eng/eng-deu. This is imported from the
DING
project, see <http://dict.tu-chemnitz.de/> (just looked it up in the
header).

I saw the Beolingus homepage here but I couldn't see much information about
the source of the German-English freedict there.

DING is independent from FreeDict. The Ding itself is handwritten, mostly by Frank Richter, AFAIK. The translation program (ding2tei) can be found in our `tools' git repository [0], under `importers/ding2tei' [1].

I did email Einhard, but all in all, I suppose I could use a little more
understanding of the German-English dictionary, where its homepage is for
example.

FreeDict's translation of the Ding does not have its own homepage; all available information can be found in `tools:importers/ding2tei'. See in particular `doc/thesis.pdf' [2], the Bachelor's thesis I wrote on the translation program.


In my experience the German-English dictionary (which I believe to be the
largest in the project) was not comprehensive enough for my needs when I
used it in a CAT tool for a translation project.

I assume you only used the dictionary to translate single words and the problem were missing headwords. Correct?

If these were rather common words, the cause is most likely a shortcoming/bug/todo of the translation program: It retains most auxiliary words such as `etw.' and `sth.' as present in the Ding source.
 - E.g.: "jdn./etw. in etw. verwandeln {vt}"

Do the dictionary
designers have a plan to make the dictionaries potentially comprehensive?
How so? Where will they get enough data from, to do so?

There are two ways to improve the deu-eng dictionary.
 1. Improve the Ding dictionary.  I believe it allows for on-line
    submissions of new data.
 2. Improve the translation program.  See the `todo/' folder and the
    aforementioned thesis [2].



[0] https://github.com/freedict/tools/
[1] webview: https://github.com/freedict/tools/tree/master/importers/ding2tei
[2] https://github.com/freedict/tools/raw/master/importers/ding2tei/doc/thesis.pdf
--
FreeDict - Free And Open Dictionaries
Manage your subscription at https://www.freelists.org/list/freedict
Wiki: https://github.com/freedict/fd-dictionaries/wiki
Web: http://freedict.org

Other related posts: