This raises an interesting question... Isn't a model of (potentially) limited accuracy better then no model at all? I say potentially because, until plans emerge (if they ever do), there's no grounds to criticize a model for inaccuracies. We already live with those all the time with our models, even some models from the classically "best" manufacturers such as Neptun. After all, the model is a replica of a real thing, not the thing itself and by definition will not be "true" in every dimension.