Initially we thought it could be useful when we introduce dependency on other evaluated features, though we observed no benefit from repeating the evaluation context into the evaluation result. For now we'll delete it, and revisit the topic when time demands it.