deinst

I am of two minds about this. It makes a lot of sense that edges in an undirected graph behave symmetrically. On the other hand, the wonderful thing about this...

It makes sense, yes, and there is nothing preventing you from making an undirected edge type such that `edge(x,y) == edge(y,x)` and making an undirected graph with it. I object...

Just for reference this does not hold in networkx. This is partially a distraction, as networkx is very much node (vertex) based. ``` >>> import networkx as nx >>> G...

@lindahua That explains it. I see that you have cloned the BGL quite faithfully, I probably should have been more aware of that. I'm still not sure that it makes...

Very good. I may look at this this weekend. If I do change things, I'll keep the old interfaces and complain appropriately. If we do change though, it would be...

This will have to be on hold for a while. Keeping the old and new parameter orders drives the julia dispatcher completely nuts. As soon as that is fixed, I'll...

Thanks. I'm still learning the julia module system. (and the language in general)

Do I really need to import Base.setindex! ? As I understand things, setindex! is an operator, and so should be implicitly imported. I'll fix the other export screwups.

I too think it is a good idea. It is a bit more complicated, as the inspectors are also carriers of type information, so the last example would have to...

@lindahua We also need to implement `vertex_property_requirement` but your point is extremely valid. I probably tend to be too prejudiced towards requiring too much typechecking of parameters.