Re: Outdated page was: Debian Dictionary
Am Mittwoch, 29.09.04 um 20:11 Uhr schrieb Frank Lichtenheld:
makdict takes simple text files with entries delimited by "::". and
makes files for dictd and html pages for different languages out of
those files.
Why the special fileformat? Why not using something like gettext?
As far as I understand gettext is used to internationalize programs.
That could be in question for the messages from the program itself. But
I do not see, how it could help with dicts and linking between entries.
And in addition I cannot see, why any format would be easier to keep by
people helping collecting meanings of words and phrases than the
colon-delimited. As far as I understand gettext, it is just to
substitute words and phrases by those in a special language. And a
dictionary does more than that.
Here one example:
RC is Release Candidate
RC also is Release Critical
This are two lines in source dward.txt:
RC::Release Critical
RC::Release Candidate
On the page and dictd files this reads out to:
RC
Release Candidate, Release Critical
Both meanings can be translated and both of them can be in the
glossaries, but must not. For other examples, entries may have multiple
possibilities to translate to a certain language. For Debian specific
dictionaries, that are not too much, but for other dictionaries like my
photographic dict, there are more words with more than one meaning, if
translated from one language to another vice versa.
Dicts are not supposed to internationalize/localize, but to give
meanings and translations of words and Phrases. I you translate Debian
pages or programs, you may decide which is the best translation in a
special context, but dictionaries do not decide. They let the decision
to the reader searching for a meaning. Beside that the program still
simplifies things related to a real dictionary.
If you still think, using gettext would make things easier for those
editing the source dictionary files, please mail me and tell me, where
the advantage would be.
As I already was asked, why I do not use "utf8" module, here the
answer, too: The machine creating the dicts would have to keep all the
locales for the languages, the dicts are available for. And surely, I
do not want to install Punjabi or Bulgarian, as I cannot do anything
myself with it. The program can be used on any machine with perl and
having Unicode::Strings installed (which just can be copied from Perl
5.8 for machines with Perl 5.6 as Woody installed machines are).
cu
Jutta
--
http://www.witch.westfalen.de
http://witch.muensterland.org
Reply to: