Update the scroll example ascii and Java docs, so it is more clear when to consume the scroll documents. Before this change the user could loose the first results if one uses copy & paste.
With this new structure supported APIs are ordered and grouped in the same way as they are in the Elasticsearch reference docs