Why should they create a whole new architecture to support when you can find changed articles between two dumps with a simple query? I'd rather load a big file into a database than maintain a firehose consumer.