AUIML/idea/2

AUIML idea #2: Component-Driven

Any application will need to support a continuous trialog between the user, the user's screen or other output medium, and the executable code, so does it really make sense to think of it as a document? I'm thinking instead that we get this sequence of events:

• User requests application (via URL in an AUIML frame/browser, aka the client)
• Application sends:
  • any preliminary setup information, such as name, version, security stuff, and a spec for which types of events it would like to receive from the client (and related details).
  • a set of initial components (display elements) that it wants the client to display.
    • Each component includes recommended positioning data (relative to other components: next to, inside...).
• Client sends events to server as requested/allowed while initially drawing the requested components.
• Client and server continue to update each other as needed after the screen is drawn.
• User interacts with screen components, which may also send events to the server.
• Server can send update information to any component at any time, changing its value or state; can also request additional components (including windows), or remove any previously drawn (e.g. closing a window).

In this model, we'll need

• a set of standard component types
  • a standard name for each type
• A lot of other things... (I'm out of time for now.)