COMP2011	mst assignment		Addenda

The following points need clarification:

20 May:
    1.	The program should read from standard input, at least. This means that
	Modula solutions need only use procedures from InOut and RealInOut. 
	We will only test the program this way
		mst <graphfile

    2.	You may use library procedures freely to assist in parsing the input.
	Thus RealInOut.ReadReal can be used when a real number is expected.
	Note that InOut.ReadString can be used for the vertex identifier
	but not the label, since that may contain several words, and ReadString
	stops at the first space.

    3.	To assist in choice of implementation, the program is free to reject
	any graph that has more than 1000 vertices. This allows an array of
	vertices to be used for the graph instead of Kingston's representation,
	and avoids the difficult task of dynamically allocating a heap array
	for the priority queue.

    4.	The program is to determine the minimum spanning forest in the event
	that the graph is not connected. Apply the MST algorithm starting
	at every vertex.

    5.	The graph should be scaled for display. That is, the x and y
	coordinates should be multiplied by a factor (< 1) to ensure they
	fit on the screen. The factor depends on the maximum and minimum 
	x and y values. To obtain the screen dimensions, pass 0 values
	to MingInit and then call MingGeometry.

    6.	Display instructions:
		(a) Display the original graph with dashed edges.
		    Include vertex identifier and/or labels as desired,
		    positioned adjacent to the vertex (don't try to avoid
		    edges or other vertices--that's a hard problem)
		    Display edge weights at the midpoint of the adjacent
		    vertices. Use contrasting colours in case of too much
		    clutter, and a small text font (where supported).
		(b) Calculate the MST (MSF actually) as a subgraph containing
		    the same vertices but a subset of edges.
		(c) Ring the bell (07C) and read a character from the keyboard
		    using Terminal.Read.
		(d) Display the MSF with solid edges.
	You should have a single graph display procedure with a few options
	to distinguish the two uses.