History for High Level Design

??changed:
-
Common Area for High Level design documents

For those of you with XMI capable Design Editors I'll put the XML here 

http://www.fourthought.com/projects/DigitalCreations/hld.xmi

Here is a link to the class diagram from the above XMI

http://www.fourthought.com/projects/DigitalCreations/hld.gif


General Design Discussions:

Last nights Use case analsys meeting drove a high level design.  The !XMLDocument implementation will be seperated into 2 focus areas.

1.  Importing an XML document into Zope as a single object.  This new single object will overide the Element interface to allow traversal of childnodes, etc as if it where a tree of objects.  Objects returned from get_childNodes will be transient (not persist in the Zope Database).  The reasoning for this is to allow an !XMLDocument to be used in all places a Zope tree (or DOM tree) can be used, but not pollute the Database with zillions of objects.

2.  Importing an XML document into a Zope tree.  This will take an XML Document parse it, map it based on a user defined mapping from Elements/Attributes to Objects/Properties, and create Zope Objects based on the mapping.  ALL INFORMATION IN  the original XML document that is not mapped to a Zope Object or Property will not bew represented in the Zope Hierarchy.  This option is inteded for import use cases only.

General Note:
	In all places where importing of an XML document is referenced, we will include, edit from scrathc, load file, and load URL.

General Note:
	When a Zope tree that contains !XMLDocument Objects is serialized (A.K.A. pretty print)  The !XMLDocument Object will serialize into a seperate file, and the original file will contain a xlink replace atrtibute, ie the original file will reference a full XML file.  This is because a valid XML file cannot contain another valid XML file.



Mike