Thinking about this, there are a couple of elements.
- Detecting whether an input is English/French. This could be done via filename.
- Whether the data is translated to or from French to English then run, or the model is reconfigured to French. That decision will hinge on how many tools there are that reference text or part thereof. For instance, if there are only 2 filters and three formulas that reference text, then I would rebuild the workflow manually. If there are 150 references, different story.
- In this situation, the workflow would direct one way or the other depending on English/French. On Server, this is easier as you can have the entry workflow call the appropriate one via API.
If the workflow is to be re-built exactly the same, then I would use the yxmd as an xml document and make a lot of the direct conversations on the XML, then save that as a yxmd, open it, and then make the remaining changes (summarises etc). If you're not sure about what this means, open a yxmd in Notepad and you'll see that it's XML underneath.
The other option, translating all the data... remembering you only need to translate the fields you'll use... has a bunch of concerns. Not least of all, how do you know that it translates properly every time. So, it depends whether it has unknown text that will be used, or whether the only fields used are categorical fields that a lookup table can translate.. For instance, and excuse my French... a lookup table that relates Oui to Yes, Puis je to Please etc.
