While working on a recent project (signed direction) I made a breakthrough: a faster, cleaner way of exploding paths.

I explode paths (break them into individual curves and line segments) all the time using my explode node. For years I've searched for ways to make this node simpler and more efficient without success. But last week I finally saw a way!

And, initially, it looked like it might be MUCH faster, as much as SIX times faster. This could be revolutionary because many of my other nodes have an explode node inside them, including nodes like Clip, Occlude, and Sub_Path that bog down when put under strain.

It turns out, though, that my new explode node is only significantly faster when used on long, complicated paths. And it turns out that such paths are rare. The other nodes that use explode already organize things to avoid exploding long paths all at once, so when I tested them I found no significant speed gains at all. Days of testing for a disappointing result.

Since there is no significant speed gain, I don't feel it's worth the risk and bother of updating all those other nodes. But the new explode node does work perfectly well, does run faster in rare cases (and no slower the rest of the time), and is cleaner inside. So I am going to update it.

Also, in all my testing, I did find one other node that was slightly improved: even_sample. The main reason for this is that, in one subnetwork, is was exploding things twice when it only had to do so once. When I replaced the explode nodes and fixed that one problem, even_sample did run up to twice as fast when in "Curves Only" mode. So I am updating that one too.

Attached please find a demo with the newly updated versions of both of these nodes (see screenshot). You can start using them now, or wait until I include them in the next release of my library.

Meanwhile, if anyone out there has a project in which one of my library nodes seems to be taking too long, please post a note on the forum and let me know. I will add the new version of explode inside other nodes if I can find compelling use cases where doing so would make a noticeable difference.