Earlier quoted context omitted.
A fundamental principle of the Web is that site-specific clients or apps shouldn't be necessary. That they are only underscores Google's massive blunder here. Yes, there've been CLI wrappers around web queries before. Until the past few years, these simply addressed search format, URI arguments, and namespace. They launched in the user's choice of browser, text or graphical. Surfraw is the classic, I've written a few…
> A fundamental principle of the Web is that site-specific clients or apps shouldn't be necessary. I think this ship, if it hasn't sailed already, is at least starting its engines and getting ready to leave the harbour. User agents and servers are increasingly trending towards an adversarial relationship, where the user doesn't want to do much of what the server is asking of it. This has been true from the first pop-…
That doesn't diminish the fact that the original intent was to have a common, freely-available mechanism for accessing, viewing, and presenting content.
(I'll probably be asked for citations. TBL has probably written on this, and Tim O'Reilly had an essay on his early response to the WWW as opposed to alternative, proprietary, systems, when O'Reilly & Associates were plotting their early course.)