24063 2013-12-11 17:04:18 +0000 Don't mandate augmenting the the WebDriver interface 2014-07-01 16:19:35 +0000 1 1 1 Unclassified Browser Test/Tools WG WebDriver unspecified PC Linux RESOLVED INVALID P2 normal --- 20860 1 ato dburns dburns fisherii mike simon.m.stewart public-browser-tools-testing oldest_to_newest 97456 0 ato 2013-12-11 17:04:18 +0000 The spec currently says: If this capability is supported, local end MUST add the TakesScreenshot interface to the WebDriver interface. We shouldn't mandate that the WebDriver interface is augmented, because it's perfectly legal to use a bare-bones RemoteWebDriver type against a backend that has more capabilities. 101380 1 dburns 2014-02-25 19:27:42 +0000 the MUST should be SHOULD 102147 2 fisherii 2014-03-10 18:00:11 +0000 Is it even useful to talk about augmenting in this spec? It seems to me that this is a local-end detail that could be handled in a variety of ways depending on the language, and the augmenting pattern is pretty specific to the particulars of the Selenium Java local-end implementation. 108621 3 ato 2014-07-01 16:19:35 +0000 After careful reconsideration I think Marc is right here. We should actually not talk about the local end at all because the spec is now remote end only. The augmentation is a WebIDL thing only and doesn't imply how a particular language needs to implement an API for the spec.