[Rd] Translation commentaries

From: Fernando Henrique Ferraz P. da Rosa <feferraz_at_ime.usp.br>
Date: Sun 20 Feb 2005 - 04:54:49 EST

        I have just finished translating R.pot to pt_BR and now I'll start the process of revision and will merge it with the latest SVN version (33261), as I started the translation on 02/06 (SVN 33048).

        There are some observations I'd like to make regarding the translation process:

  1. Sometimes it is not clear in the message whether a given term is referring to an actual function argument or to a more general concept. For example:

#: src/modules/lapack/Lapack.c:49 src/modules/lapack/Lapack.c:128
msgid "method must be a character object"

        Generally I look into the respective source code and try to figure out from context. When it is a function argument I'm keeping the original word untranslated, when it's just a reference to the concept I translate it. It would be interesting to standardize this: whenever an actual argument of function is being referred to, it could be enclosed in simple quotes:

msgid "'method' must be a character object".

        This already happens for some msgs, but not for all of them.

        2) Some messages are essentially the same, with minor variations. I think they should be merged into just one. For instance:

#: src/main/subassign.c:396 src/main/subassign.c:669
msgid "number of items to replace is not a multiple of replacement length"

#: src/main/subassign.c:1099

msgid "no of items to replace is not a multiple of replacement length"

        3) There's at least once the assumption that the plural form of
a noun will be composed by adding an s:

#: src/main/util.c:806

#, c-format
msgid "%d argument%s passed to \"%s\" which requires %d"

        Where the relevant source code reads:

   error(_("%d argument%s passed to \"%s\" which requires %d"),

              length(args), (length(args) == 1 ? "" : "s"),

        While it may hold true for English (and incidentally also Portuguese), in German for example the plural form for Argument would be Argumente.

        Unfortunately I can't provide more comprehensive lists for 1) and 2) right now because I haven't started the revision and updating process yet, but as soon as I finish it I'll send it to r-devel.

Fernando Henrique Ferraz P. da Rosa

R-devel@stat.math.ethz.ch mailing list
Received on Sun Feb 20 04:01:37 2005

This archive was generated by hypermail 2.1.8 : Sun 20 Feb 2005 - 04:29:44 EST