Issues

• Two basic approaches: using the vobscene for the info, or storing the info in a different (single cellview instance) object.

  Pros and cons:

  Vobscene:

  • simplicity

  System Message: WARNING/2 (../ffdoc/alph/pegboard/spancellviews--tjl/peg.rst, line 24)

  Bullet list ends without a blank line; unexpected unindent.

  • inflexibility

  Cellview instance:

  • the data can be stored efficiently, and needs to be created anyway

  System Message: WARNING/2 (../ffdoc/alph/pegboard/spancellviews--tjl/peg.rst, line 29)

  Bullet list ends without a blank line; unexpected unindent.

  • complexity

    RESOLVED: we shall use the VobScene, in a slightly special way, as described below.

Basics

CellView A places the contents of a cell into a vobscene under some CS in the hierarchy. ConnectingView B wants to know where a certain subspan of it is. Instances of this problem:

• anchors of buoys
• irregularframing of the linked-to area in a buoy.
• xanalogical comparison of two documents with beams

Version 0.6 allowed this to happen easily by storing the spans as the keys in the VobScene. This worked fine as long as VobScenes were not hierarchical, and no image spans were used.

Now the situation is different: when linking to an image span we want to know exactly where a given subspan is.

The solution

The solution is to define a convention that Image- and PageSpan keys in a VobScene will be used on coordinate systems exactly when the coordinate system is a mapping whose unit square is exactly where the Image- or PageSpan (PageSpan's first page!) is linearly placed.