The current prose of this section is wooly, and lacking in normative conformance requirements (the invokation of a successful refresh command has none). We should model these commands after the more algorithmic prose of the DOM spec:

http://dom.spec.whatwg.org/#queuing-a-mutation-record 

For a majority of them, however, our algorithm can simply be a wrapper around references to the HTML spec[1,2,3], with our own exceptions, page load strategies and responses sprinkled on top,

[1]: http://www.w3.org/html/wg/drafts/html/master/browsers.html#read-html
[2]: http://www.w3.org/html/wg/drafts/html/master/browsers.html#history
[3]: http://www.w3.org/html/wg/drafts/html/master/browsers.html#dom-location-reload