Add combining interface

[ofer]

Allow the translation interface to translate a couple of segments at once, maybe by adding (add next segment) to translation interface.

I assume this will correct 99% of our parser badness.

The database can still contain the original segment with an order to get larger match.

example:
שלום, עולם.
שלום -> GN (get next marker)
שלום, עולם -> Hello, world.

[mail@…]

+1 to this. Many times the editor will separate sentences arbitrarily (say because there's a quote in the middle) which does not allow someone to translate the whole sentence correctly.

[datapharmer]

I was looking at this and a few related tickets and thought that perhaps a second pass could be done on individual words that aren't translated.

I found that google sometimes won't translate a phrase but will translate almost every word in the phrase individually, so it may be possible to do some sort of diff check and if the same word occurs in the translated version, run it through again. Not sure if this is the best approach to solving some of the untranslated issues, but thought I would throw the idea out there.

[ofer]

Thanks, Idea appreciated, will see how to implement when I come to implementation (Could also use next translation engine before breaking to words)