Question:

Consider the use of the "extent" variable on GoF p. 208 and pp. 213-215. The discussion on 
pages 213-215 is a bit confusing and refers to caching the image extent and saving and 
loading it. Sort this out and explain it. What is the difference in the Save and Load for the 
Image and the ImageProxy? When is each used? Come up (if you can) with a reasonable 
scenario where a client will request the extent of an image that has not yet been loaded and 
it will get the cached extent without loading the image itself. 

Answer:

	You sort of threw out the question in class "is this a good idea
and does it work?"  yes and yes.

	The "cached image extent and file name" refer to the variables 
Image* _image
char* _filename 
that are part of the ImageProxy class. These have been initialized and are 
in cache along with the proxy object. Save writes them to disk. Load reads 
them from disk. 

	There is no difference in the save and load for the ImageProxy and 
the Image. The ImageProxy class handles all calls to save and load never 
passing them on to Image.


void ImageProxy::Save (ostream& to) {
    to << _extent << _fileName;
}


void ImageProxy::Load (istream& from) {
    from >> _extent >> _fileName;
}



	You gave away the answer to the scenario question in class. I have
on my web site torches along the left margin. When the mouse is moved over 
one of these torches an image appears in what had been a blank space. The
image is not loaded unless the mouse event triggers it to be, but the size 
of the image is needed for the page to place items above and below where
it will go.

	
	-------------------------
	|			|
	|visible		|
	|			|
	|visible		|
	|-------		|
	|	|size		|
	|	|only		|
	|	|		|
	|-------		|
	|			|
	|visible		|
	|			|
	|			|
	|			|
	|			|
	|			|
	|			|
	|			|
	|			|
	------------------------