These ideas are interesting but why is it limited to a drop-down from the taskbar? You could make the entire window pane work like that.<p>The last example, where the user is browsing search results, is just a recapitulation of the entire browser concept inside a taskbar dropdown, with more animation and tear-off results. The animation and tear-off results are an interesting idea, so why not apply them to the whole browser window?<p>But the main idea here is that different verbs work in the taskbar now. We can imagine that the address bar has always had the implicit verb "go to"; now we'll do other things, like "translate".<p>Of course the next step is to make the translation provider configurable, and maybe you even get to choose certain providers, just like Mycroft (the search box on Mozilla browsers). But for configurability, you can't beat HTML. If you think of HTML as "interface configuration downloaded instantly" it becomes the ultimate end of all customization efforts.<p>So maybe what this is proposing is an interface that <i>isn't</i> configurable, that's deliberately constrained. Just like Mycroft forces all the search engines to work in the same way, at least in how the service is <i>requested</i>, but then afterwards it's all HTML pages. I think Aza would prefer to keep control over how the service makes a request and how it presents results, but the easier & simpler way might be to just make Mycroft-like services, where we are just requesting HTML from somewhere, but in a way that's more convenient for a specific task. That also gives service providers an economic model, if they can include ads or other upselling in the results. We can't assume it's always going to be Google's largesse.