Crop Attributes and Data Standard

In Permaculture practice, each Crop has attributes such as: deep roots, nitrogen user, ground cover-er, attracts ladybugs. If we could get these attributes, then we unlock a lot of potential. These attributes could be crowdsourced or imported from another source. Here is a resource we have found that has started something similar and they are interested in working with us!
http://sourceforge.net/p/kitchengarden/gitcode/ci/master/tree/resources/species.xml

If we collect attributes, I think this is a good approach: Users would not type-in the attributes, but rather choose them from a list of possible attributes so that the data is not defined by an arbitrary name, but a canonical concept that can be language translated and used as cleaner, more interoperable data - just like the naming of Crops and the prerequisites and growing instructions of Guides. Then, we can programmatically (automatically) list which Crops are companions and which are not, and for what reasons. ie: a deep rooted plant that uses nitrogen goes well with with a shallow rooted plant that fixates nitrogen from the atmosphere. Maybe an upvote/downvote mechanism could crowdsource whether the automated suggestion is correct or useful or not, and also the suggested relationship's popularity.

With this data, a user could perform a blank search and then filter the Crops by attributes to find the perfect one for the niche in their garden they are trying to fill!