Sorry, when I said a vertex whose name was medicine, I meant it had a name 
property. The class name would be topic where medicine is an instance of 
topic. I didn't intend to use the class name. 

Yes I think that that a vertex for each localisation of the concept name is 
the way to go. Thanks for your help.

On Wednesday, March 9, 2016 at 4:59:57 PM UTC, scott molinari wrote:
>
> I'd say, you'd want to avoid using class names to determine language 
> choice.
>
> There are two types of text that can be translated. Content (user input or 
> application data) and application text (text added by the programmer). From 
> my experience, the application text is usually centralized in a single 
> class. So, if "medicine" as a term is being used in the application (added 
> by the programmer), but the German version is needed, then there would be 
> an extra query to call up the translation for the term "medicine". 
>
> Now, with content about the medicine, like the name of a specific medicine 
> or other properties of a particular medicine, then I'd put the translations 
> in separate documents. This could also be handled well over a "language" 
> edge, which links medicine documents to other translated medicine 
> documents. This would allow, for instance, to go from German to English and 
> vice versa, if needed. If you don't need this, then I'd just have a 
> language property in each document and filter on that property. 
>
> Scott
>

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to