Brian Ingerson wrote:
> As much as I loathe comparing scripting language structures to
> mathematical theory, I will make my best attempt to internalize this
> document. The first thing I needed to to do was find out what a
> "vertice" is. Apparently it isn't anything, as "vertices" is the
> inflected form of "vertex". FWIW, after searching google it appears that
> you are not the only math person to have made up this word.`
Me bad. It should be vertex, of course. Too much Hebrew-isms when studying
graph theory too many years ago...
> Just one last gripe before i dive back in: Do you really think that it's
> in YAML's best interest to define YAML models and YPATH in pure
> mathematical graph theory? I've pushed against this from the beginning,
> because I know that 99.999% of YAML users will not visualize it this
I think its important to be _able_ to do that, as it allows applying formal
theory to YAML. But I think that the most important point is the
When we discuss YPATH between ourselves, and when we explain it to people,
we should be able to refer to _some_ standard way to visualize it. Last time
we tried to talk about YPATH, I had one visualization in mind while you and
Clark had other(s) - that caused a lot of misunderstanding.
Since there's only one way to visualize graphs, I had hoped that casting
YAML as a graph would provide this standard visualization. I've spent some
time trying to come up with some other way for a standard visualization of a
YAML document (beyond the YAML text itself, of course).
The big problem is the introduction of a "pair" as a first-class entity in
the visualization. Since the same (identical) node can be used as a key in
multiple maps, I think there's no way around it, no matter what mechanism
you choose. Once one understands the need for the existance of "pair" in the
model, then using a graph provides a convenient way of visualizing and
discussing the issues while ensuring we all follow each other's terms.