Python documentation in DocBook

[Zoltan Sekeres]

Martin v. Loewis wrote:
> The problem that I still consider unsolved is how may translators can
> work on a single document, and how retranslation of the entire
> document in case of changes can be avoided. This can be solved
> theoretically, by assigning IDs to relevant elements (and then finding
> out which IDs need to change), but I'm not sure whether this would
> work in practice.

(It drifts into off-topic space, but ...)

That's what translation memories in general and XTranslate from Trados 
are good for. Translation memories save segments for re-use, either in 
the same document or in later versions. But the segments are usually 
stored without context, so the translator still has to check.

Additionally there is XTranslate which takes an old version of a 
document including translation and a new version without. XTranslate 
will then use the context to identify segments which were added or were 
changed. The translator has to check less.

Using IDs for this purpose didn't catch on AFAIK. Especially as 
segmenting, finding the correct units for those IDs, is a though job.

-- 
Zoltan