[Gzz-commits] gzz/doc/pegboard/vobuml--tjl peg.rst

[Tuomas J. Lukka]

CVSROOT:        /cvsroot/gzz
Module name:    gzz
Changes by:     Tuomas J. Lukka <address@hidden>        02/12/31 07:59:45

Modified files:
        doc/pegboard/vobuml--tjl: peg.rst 

Log message:
        More

CVSWeb URLs:
http://savannah.gnu.org/cgi-bin/viewcvs/gzz/gzz/doc/pegboard/vobuml--tjl/peg.rst.diff?tr1=1.1&tr2=1.2&r1=text&r2=text

Patches:
Index: gzz/doc/pegboard/vobuml--tjl/peg.rst
diff -u gzz/doc/pegboard/vobuml--tjl/peg.rst:1.1 
gzz/doc/pegboard/vobuml--tjl/peg.rst:1.2
--- gzz/doc/pegboard/vobuml--tjl/peg.rst:1.1    Tue Dec 31 05:08:48 2002
+++ gzz/doc/pegboard/vobuml--tjl/peg.rst        Tue Dec 31 07:59:45 2002
@@ -3,8 +3,8 @@
 =============================================================
 
 :Author:   Tuomas J. Lukka
-:Last-Modified: $Date: 2002/12/31 10:08:48 $
-:Revision: $Revision: 1.1 $
+:Last-Modified: $Date: 2002/12/31 12:59:45 $
+:Revision: $Revision: 1.2 $
 :Status:   Incomplete
 
 With PP, the structure of the hierarchical coordinate
@@ -15,8 +15,34 @@
 Issues
 ======
 
-Vob-UML
-=======
+- Should Coordinate systems use the generalization relationship
+  to their type? I.e. "coordinate system of viewport frame 
+  inherits from orthoboxCS"?
+
+    RESOLVED: No. In the current context, the coordinate system
+    type is a *metaclass*, not a *parent class*. While a template
+    might be considered, the template notation is clumsy.
+    Therefore, we shall use the *stereotype* notation
+    for coordinate systems: defining stereotypes such as
+    "orthoboxCS" for all coordinate system types, and
+    "*CS" for coordinate systems of indeterminate type.
+
+- Should we define images for the Vob and Coordsys stereotypes?
+
+    RESOLVED: Not yet.
+
+- How should the interpolation key be represented?
+
+- How should the sequence of Vobs be shown? Specifically, if 
+  there are alternatives (e.g. in Stencil, the border may be 
+  left out).  UML statechart -type things?
+
+- Should this spec discuss umltool as well?
+
+    RESOLVED: Besides giving clear names to the relationships, no.
+
+The objects to be described
+===========================
 
 There are two kinds of "objects" (in the following, the quotes
 will be dropped) in the Vob system:
@@ -48,6 +74,32 @@
 - Type
 
 - Parameters
+
+Vob-UML
+=======
+
+Both coordinate systems and Vobs are represented as stereotyped UML
+classes. For coordinate systems, stereotypes ending in ``CS`` are used,
+and for Vobs, ``Vob`` or ``Vob0`` or ``Vob1`` or ``Vob2`` is used.
+
+Relationships between coordinate systems
+----------------------------------------
+
+Both types of relationships between coordinate systems
+in the same VobScene are represented as arrows; the arrow for
+transform parenthood (A ``transform`` B == A is child of B)
+is thin and possibly contains (at the arrow end)
+the index of the parent as a role. The interpolation parenthood
+(A ``submatch`` B == A is interp child of B)
+is represented by a thick arrow. 
+
+Relationships of Vobs
+---------------------
+
+The containment of a Vob in coordinate system(s) 
+(A ``vobin`` B == Vob A is in CS B) is represented 
+by a thin, dashed line. The coordinate system end is optionally
+adorned by the index.
 
 
 Examples