Hello all,
I am currently working on redesigning a database structure for online
translation utility called Entrans [1].
Entrans is quite nice piece of software, but unfortunately it has one
major design flaw: it only supports languages which have only two plural
forms and to fix this issue, a database structure has to be redesigned.
The way Entrans works is simple: it parses the PO-catalogs [2] and inserts
the entries found in the file into the database. The main issue is how do
I store these entries, because there are two different types:
a) a singular entry
b) a plural entry
With singular entries it's easy, we following items of data:
flags
comments
msgid
msgstr
Now, the problem is storing the plural entries, which have four "static"
items, followed by the n > 1 of dynamic entries:
flags
comments
msgid
msgid_plural
msgstr[0]
...
msgstr[N]
At first, I tried to do it like this:
msgid_table:
id
msgid
msgcontext_table:
id
msgid_id (FK for msgid_table.id)
is_fuzzy (flag)
comments
file_id (FK for file_table.id, not presented here...)
msgid_plural (present if msgid_plural exists)
plural_count (number of plural forms)
msgstr_table:
id
msgcontext_id (FK for msgcontext_table.id)
user_suggestion
index (to which of these msgstr[n] fields it will be put, 0 if string
doesn't have any plural forms)
is_valid (is it valid translation suggestion)
Now, as you can see, this approach doesn't work at all... because it's
quite a hell designing a query based on msgid_table.id ->
msgcontext_table.mgsid_id -> msgcontext_table.id ->
msgstr_table.msgcontext_id.
What I'm thinking is that the msgcontext_table.id should be made a
superkey, but this does not help when I need to figure out which of the
suggestions are valid for current msgid.
Your ideas are very welcome ;)
[1] http://entrans.sourceforge.net/
[2] http://www.gnu.org/software/gettext/manual/html_node/gettext_9.html
Best regards,
Priit
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
