Anatomy of a Web Browser which is User-Configurable

More about Anatomy of a Web Browser which is User-Configurable

As discussed in the previous picture, let's explore (lol) how we'd implement a User-Configurable Web Browser. The main purpose of such a browser is to give the user control of the layout and style of what they view, and thus improve user contentment of the WWW as a whole, and frequently-visited websites in particular. CSS (cascaded style sheets) was a sort of attempt at this, but few browsers expose a suitable interface. Basically when the browser reads a HTML page, it parses tags (such as <h1> and <table>), displays text or graphics, and runs scripts (source code procedures) embedded in the page. In a FSM parser such as ::SHE+ILA:: 's front end, the input stream is first lexed (split into tokens or "atoms") and then this stream of tokens "trains" a state-machine by setting the probabilities (Markov Weightings) of various tokens preceeding or following it. Thus a ::SHE+ILA:: browser can dynamically learn the structure of a webpage (or, in fact, any document) and adjust it to fit the user's preferences, such as the screen position of various items. how clickable links are displayed, whether certain JS or VBS scripts are allowed to run, and even add their own SheilaScript procedures and macros. In fact, general AI principles could be applied to the rendering of the web page. Caveat : the copyright situation of such processing would need to be clarified with a competent computer copyright lawyer (which I am not). E&OE.