> On Jun 16, 2015, at 8:35 PM, Logan Streondj <[email protected]> wrote:
> 
> I've always had trouble understanding graphs, lines and boxes.
> What are these topologies? can they be expressed in text?


Think of it as relationships between shapes embedded in space. 

Humans have an intuitive conceptual understanding of relationships in Euclidean 
3-space because that is the environment we live in. An interesting and 
well-known problem in ordinary databases is that people generally have no idea 
how to represent spatial relationships such that a computer can understand it 
the way a human does. Consequently, spatial relationships in data can only be 
interpreted correctly by other humans. We assume that shared context in the 
encoding of the representation. But again, a lot of this is because the vast 
majority of software can’t represent topological relationships, only graph-like 
relationships. Humans have the ability to represent and reason about shapes in 
3-space built-in. It is an impedance mismatch. 

There are some interesting theoretical requirements that make this much harder 
to solve than it sounds and which are missed by virtually every computer 
scientist that dabbles in it. You cannot tractably reason about N-dimensional 
topological relationships in N-dimensional spatial representations. In other 
words, if you want to reason about the physical world (a 3-space), you actually 
have to transform that 3-space into a significantly higher dimensionality 
complex topology, the manipulation of which efficiently produces the expected 
3-space results. 

Your conventional human computer scientist is not accustomed to doing what 
feels like intuitive, ordinary reasoning about relationships in the physical 
world by mathematically folding, stretching, twisting, etc an exotic 
7-dimensional surface or similar. Nonetheless, it is the only way to do it. 
That aside, it also has the advantage of naturally being an extremely 
parallelizable method of computation, even for traditional data models.


In principle, you would expect a parallel computational system that is 
optimized for reasoning about an N-dimensional topological data models to also 
be effective for topological data models of lower dimensionality and graph-like 
data models. Not so much for anything else. It might be coincidence, but human 
reasoning capabilities look like what you would expect from a computer 
optimized for 3-space topological data models.

-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com

Reply via email to