Hi everybody, In Human Inference we have an idea/requirement to make it possible to remotely explore metadata and fire queries towards a DataContext that is bound to a specific server machine. The idea is that it should be possible for clients to also access this DataContext, but obviously since the context is bound to the server, it would have to be via some web service or other remoting technology. While I was sketching out what was needed to do this I kinda realized that this could actually be something we make available as a general MetaModel module instead that we build a one-off for our specific application.
So first question I guess is: Do you think such a feature would be a good addition to MetaModel? Onto the design of it ... I imagine that in the MetaModel perspective we could make abstract servlets, controllers or something like that, which the application developers could then quite easily adapt and expose the DataContext that they wish from. I would personally prefer if they are REST based. So we could even consider if we can make something JAX-RS based or something like that. I imagine that for read-only access there would be at least two web services: 1) a schema web service and 2) a query web service. On the server side this would expose data for the client to implement the getSchemas() and executeQuery(...) methods on a client-side DataContext implementation. Of course there are always security concerns with something like this too. My feeling is that if we make it in deed abstract then the application designers can judge what should be possible security-wise and so on. Please let me know what you think. I can be more specific if needed, but I thought to first just present the general idea. In my opinion this could be a pretty cool way to broaden the scope of MetaModel and open up for more data federation type use cases. Best regards, Kasper
