Lists Home |
Date Index |
Elliotte Harold <email@example.com> writes:
> Hunsberger, Peter wrote:
> > I don't agree. Where are the graph serialization standards? Where
> > are the best practice algorithms for graph traversal?
> Where are the
> > standardized languages for graph transformation? Where are
> the "graph
> > databases"?
> I think we're talking past each other on two different levels here. I
> was talking about the generic principles of handling graphs
> (which are
> well understood, as are those for handling trees) and you
> were talking
> about common, agreed upon formats for data.
I had a feeling where you were coming from, but I was trying to get a
little pragmatic. Theory's one thing, commodity competition between
vendors is a whole 'nother deal....
> I don't think there are any standardized formats for graphs,
> similar to
> what XML is for trees. It might be an interesting project to
> invent one.
> It might even be useful. I suspect it would look a lot like
> RDF, minus
> all the URI baggage.
> The graph databases do exist, though I'm hardly enough of an
> expert to
> point to one.
At a vendor level?
> The best practice algorithms are well-known too
> within the
> graph theory community, of which I am not a member so again I can't
> point to them for you, but they are out there. Ask a practicing
> mathematician for details, and then be prepared to listen for a while.
Sure, and I'm familiar (or used to be) with a bit of that world, but I
guess I should have been more clear; well known implementations of best
practices. I want the equivalent of Saxon for graph traversal...
> I don't know if there are any standardized languages for graph
> traversal. That might also make an interesting project.
Well, as I point out elsewhere, if you think about it, in a way xPath
already does graph traversal. The question of what is "root" is
problematic, but if you had a way of just saying "treat this node as
root" wouldn't you be well on your way? Alternatively, if you can
bootstrap yourself into the graph somehow then you can treat everything
as relative traversal with xPath (though I'm sure that would soon become