proposal: slight change in use of entities

[Matthew East]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

At the moment, when generating translation templates, and translated
docbook, we use the "-e" function of xml2po, which basically means that
the entities we use get expanded into the template, so the translated
docbook doesn't contain the entities any more, but rather has the entity
text directly in it.

This causes a slight problem if we want to change something in an
entity, because it means that the translators have to retranslate the
string if we just change an entity.

For example, to change the version entity from 6.06 to 6.06.1, we'd have
to get the translators to re-translate every string where that appears.

We can get around this by changing how we use entities - instead of
expanding them into the translation template, we can keep them separate.

However, this requires that we don't use any entities that might change
when translated into a different language. So an entity which reads
"Ubuntu" would not be satisfactory, because "Ubuntu" gets changed in
many languages depending on the context (e.g. "in Ubuntu", might require
a different word to "from Ubuntu").

So I'd propose that what we do is limit the use of entities to version
numbers and urls, and switch the translation practise as described above.

I don't think too many entities would have to be removed, to be honest.

Do people have any thoughts on this?

Matt
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFE6uB6tSaF0w5rBv8RAsVEAJ9TMhHdTOKVh5eAigalgc+Mm3Gh8QCcCWCP
lsb3pKhF7P4ei0x6jBWuwQ4=
=r2AS
-----END PGP SIGNATURE-----