That's literally how IBM block terminals operated, and it was AFAIK most visible with CICS which was designed around similar model to Web 1.0 apps - blocks of code ("transactions") in CICS would send a (possibly multipage) form to terminal, then when the terminal send back the response (usually just the filled in fields of the form) another transaction would get fired and perform processing on it, and so on and so on.

Of course the screens had rudimentary markup to support this including client-side validation of sorts.