[Scrollkeeper-devel] Better internationalization support

[Laszlo K. <las...@Su...>]

Hi,

I would like to implement this, but I think we still have some issues to
clarify. I think there are two problems here:

1. Using xml-i18n-tools to translate the category trees.
2. Being able to identify a doc and all of its translations easily.

1. depends on xml-i18n-tools that (I think) creates a message file from
the XML file and merges back all the translations into the original
file. This will be changed in the near future to create separate XML
files for each translation. It seems we will not use it until this is
not done so translation will be done by hand until then.

2. This is the long discussed document ID and this is what I would like
to address. We want an ID that is the same a document and all of its
translations and it allows to find the English doc if the translated one
is not there or fall back to other translations of the doc (like first
to German then to English if the French is not there. I think the main
problem is not how to implement the fallback mechanism, but how to
assign these IDs, how to make the docwriters, translators to request the
ID and how to send it to them. It should be very easy for anybody to
generate an ID like this. Then I think it should be down to the
docwriters, translators to make sure to use the same ID for a doc and
all of its translations. Uidgen was mentioned to be used to generate the
IDs. As this is not available on every platform a web based request
mechanism with a Linux server could be a good approach (like ask the
Linux server for a number). This should probably all be hidden inside a
Scrollkeeper script or binary that just retrieves an ID whenever run. 

Any thoughts?

Laszlo