This documentation is archived and is not being maintained.

A Portrait of the Document/View Architecture

Documents and views are paired in a typical MFC application. Data is stored in the document, but the view has privileged access to the data. The separation of document from view separates the storage and maintenance of data from its display.

The view accesses its document's data either with the GetDocument function, which returns a pointer to the document, or by making the view class a C++ friend of the document class. The view then uses its access to the data to obtain the data when it is ready to draw or otherwise manipulate it.

For example, from the view's OnDraw member function, the view uses GetDocument to obtain a document pointer. Then it uses that pointer to access a CString data member in the document. The view passes the string to the TextOut function. To see the code for this example, see Drawing in a View.

The view might also interpret a mouse click within itself as either selection or editing of data. Similarly it might interpret keystrokes as data entry or editing. Suppose the user types a string in a view that manages text. The view obtains a pointer to the document and uses the pointer to pass the new data to the document, which stores it in some data structure.

In an application with multiple views of the same document — such as a splitter window in a text editor — the view first passes the new data to the document. Then it calls the document's UpdateAllViews member function, which tells all views of the document to update themselves, reflecting the new data. This synchronizes the views.

What do you want to know more about?