Almost any interactive website breaks down into logical divisions. For example, there is sometimes a "People" section of the site where you can create an "account" for yourself, log in to your account, edit your account, etc; on e-commerce sites there's always an "Order" section where you create order, check an order's status, cancel an order, update an order, etc. In the webbox architecture each of these functional areas is made into a PHP object with a thin wrapper around it to allow each of the object's functions to be called and receive arguments via URL, form action or via PHP script. The object plus its thin wrapper layer and other supporting files, I call a "webbox". The functions that can be called on the webbox, I call "boxactions". I use the term "webbox" for the wrapped object because this entity isn't really a pure PHP object anymore (it can't be subclassed) and because of the one feature that distinguishes its "boxactions" from regular PHP class functions -- that they can be called directly from URLs or form actions. The implementation of this feature is the whole purpose of the webbox architecture. Without it, "webboxes" are nothing more than regular PHP objects, and I should not use any new terminology.

The boxactions of each webbox should be able to do anything PHP can do -- write HTML out to the browser, perform queries on relational databases, send e-mails, look in LDAP directories, launch programs on the host system, create GIF images, encrypt data, create PDF files, parse XML documents, open network sockets, etc.