I'm interested to know why the configuration of full text indexes (by which I mean the indexed labels and properties) can't be updated after they've been created.
I'm using a single full-text index to index pretty much every text property in my db to provide a "search everything" function in my app. As the product grows in complexity, I expect I'll want to add new properties and labels to the index. My reading of the documentation is that the idea of adding a property or label to an index is not a possibility. Instead, I'll need to drop the index and re-create it from scratch. Which, you know, seems less than ideal, especially with a growing amount of data that constantly needs re-indexing from scratch. 
So, what's the reason for only allowing create and drop, but not update?
Is updating of full text indexes something that could be added in the future?
Alternatively...
Have I got the use of full-text indexes wrong?
Is there some best practice I haven't found for efficiently text searching across the whole graph?
Maybe creating a separate full text index for each property and then somehow querying all the indexes in parallel?
(Or just managing an external index?)