|Author:||Tuomas J. Lukka|
For reorganizing xupdf in view of the coming article deadline, we need to look at showing spans/enfilades in a cell in a flexible way.
Two basic approaches: using the vobscene for the info, or storing the info in a different (single cellview instance) object.
Pros and cons:
RESOLVED: we shall use the VobScene, in a slightly special way, as described below.
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 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.