Re: Observer pattern limitations


Leslie Sanford wrote:
"David Barrett-Lennard" wrote:

<snip>

By contrast, most spreadsheets use an algorithm called "topological
sort". This involves sorting all the nodes to be calculated based on
the partial ordering defined by the dependency edges.

You may be interested in this:

http://pycells.pyworks.org/Blog

It's a blog about PyCells, which is a port of Cells from a CLOS
extension.

That's about all I know. In reading the blog tonight, a few things rang
a bell with me concerning the discussion we've been having here. So
thought you might find it interesting.

It is Interesting. It is very similar in nature to what I'm doing in
C++. I also use the idea to build and maintain dependencies
automatically.

I'm not sure if they're using an eager or lazy update approach.

A lazy algorithm is more efficient IMO. It also allows for
recalculation to be performed in shared read locking modes allowing for
better parallelism in a multi-threaded system.

Cheers,
David Barrett-Lennard

.

Quantcast