Re: Why the emphasis on graphics?

I think Narc is saying, if you look at the output from a graphical 
designer, the coordinates produced, it does not look like something 
that could be created by hand. For example, if you convert an arc 
into a bezier path, you are going to get some control points you 
would never compute by hand.

On the other hand, a hierarchical text based interfaced (xml) is 
great for seeing the logical struture of a GUI or a graphical 
component.

The interesting thing is, that tree view doesn't have to be xml. It 
is easier to edit using a real tree control, not a text file, as 
long as you can do all the same things on both. This is how the best 
xml editors work. But it is not necessary for hierarchical data to 
be expressed as xml, to build a tree-view style editor. Binary data 
can be edited that way as well.

State machines are typically expressed as diagrams (Harel or now the 
UML equivalent).

You are right that editing xml is more clumsy than editing code -- 
especially if you are using a text editor and not a high-end xml 
editor. But one advantage of xml serialization is that inherent tree-
view visualization. Take a VG.net Picture, export as MyXaml, and 
compare to the code-dom serialized version. The MyXaml version is 
easier to read, simply because of the tree-view. The data is 
identical.

Frank