Naming conventions and data

[alabarre]

Hi,

I just pushed a few separation results in the theorems directory. Before I push the others, I find it best to agree on naming conventions and I'll follow the established ones if need be. Let me just lay out what I had in mind:

• separation results: I see that theorems are named T000000....md; I had started using AnotB.md, where A and B are ISGCI ids to make it clear that the markdown means "if a graph is a member of class A, then it is not a member of class B". This makes less sense if our intended scope is broader than separation results, but it makes it easier for me to track what has been done already. If we follow the T000000....md, no worries, this just means slightly more work to retrieve the information and convert it into another format.
• graph classes: likewise, I use more evocative names for my files, like some_id=the_class_name for the same reasons, but I can switch to P000000....md.

I also have a few questions:

• should graph classes be stored in properties, or in spaces? From the data that's already here, I'd guess the former, but please confirm (and then I'd like to know what's supposed to go in the other directory; apparently, that'd be dedicated to specific graphs instead of graph classes);
• some ISGCI classes are known to be equivalent (e.g., bipartite = "odd-cycle-free"); do we want to represent this in the system, and if so how? a property for "bipartite", a property for "odd-cycle-free", and then a theorem showing the equivalence?
• I would also like to add references to papers; sometimes results are trivial or can be proved with two sentences, but in some instances referring to Theorem X page Y in paper Z is the more reasonable option.