Different glyphs for different classes

[ghost]

Ext Chart now computes the 'orders' associated to different EXTTerm objects that it displays. These orders should be illustrated by using different glyphs to draw the terms (rather than the uniform choice of a solid black dot we have presently).

Ideally, this would be user-configurable, but it would be sufficient for now just to distinguish between the infinite-order and finite-order cases. A common glyph for an infinite order class is an unfilled square.

[ghost]

I believe we should implement #75 before this. If a term cell has rank ≤ 3 and at least one term has ≥ 2 homology representatives with both finite and infinite orders, we need some underlying (view) model mechanism to choose which glyph should be used in which position, making sure that (differential) lines connect to fixed (and correct) endpoints.

[ghost]

Implemented by commits

c65514e
2a32357
5798359
92c6309

I need Eric’s eyes on this.

In the demos, A(1) May shows filled dots, Serre with fibration S^1 → CP^2 → S^5 shows unfilled squares, and KU HFPSS shows a mix of filled dots and unfilled squares depending on the page.

[ghost]

Lest I forget: EXTChartViewModel and EXTImageTermLayer would benefit from Eric reviewing my choices of words for classes, properties, methods and comments in the changes introduced by issue #75 and this issue.

[ghost]

After this has been verified correct, we need to port hReps & glyphs over to EXTShapeTermLayer, which in turn may be substantially changed because of issue #92.

[ghost]

The visual effect of this is definitely correct (and I corrected a model bug surrounding it this afternoon in dbd6b79 ). Looking through the source just now, some of it doesn't mean anything one way or another to me, but everything that I understand looks sanely organized and implemented. At some point in the future we'll want to give the user some manner of control over how things the glyphs get assigned, which is mostly what I have my eye on, and things look well-written toward that eventuality.

I'm happy to call this closed if the EXTShapeTermLayer thing exists in a separate issue, or to proceed with that in any case.