Geometry generation vs tessellation

The discussion on geometry generation vs tessellation on GitHub – Efficiently Working With Large Data Sets In Dynamo – makes me wonder if taking a group of nodes and turning them into a Design Script block would do the same/similar to creating a Zero-Touch Node (which is outside my current skill set) in terms of not generating the tessellating the geometry at every step. I have a lot of nodes that just move and reorganize thousands of lines that don’t need to be tessellated until much later in the process.

The conversation here – http://dynamobim.org/forums/topic/dynamo-nodes-vs-code-block-who-is-faster/ – suggests that code blocks are faster. This suggests to me that, if your graph is finished, it might be a good idea to convert as much of it as you can to DesignScript with Code2Node… I mean, so long as you don’t need to edit it!

Yeah generally speaking if you can compress all of the nodes to a single design script codeblock it runs faster. And it can be edited, just have to learn to read/write design script similar to how others read/write Python or C#.

I haven’t yet seen any Python vs DS vs Zero Touch speed tests but it would be great if we had one.

Also I would like to point out that Design Script is changing often so that expertise in how to write/edit needs to be kept up with as the language develops. Had to edit the crap out of some nodes so wrote for 0.8 the other day and it wasn’t much fun. Python may be better as a result of this, though API changes may make that as difficult. Not sure as I don’t know enough Python to say for certain.

For sure, but I’ve seen the same thing just in the various nodes I’ve used over the years. Many of those where before the 1.0 release, however, so that is likely to be expected but I’d expect that to continue in some manner over the years. Changes to the API alone will wreck havoc on many workflows!