This is probably a place where we will disagree.
Based on what I’ve seen so far, Blockly isn’t a bridge to “real code”. It is going to be the first and only language most of the non-technical users will use. That will be even more the case as the Blockly library builds up with more choices on in the marketplace. I would not be surprised if it becomes the most used language over all. Forcing these users to have to understand and avoid type errors, the number one problem these users face, because it’s good for them, is, IMO, user hostile to these types of users (and a big pain in my side as forum support since, as I said, type errors are the most common problem these users face).
Even before 3.2, with all the limitations and missing features, the uptake in Blockly use was breathtaking. It’s already very popular.
Like I argued in the GraalVM issue, we need to try to avoid forcing work and expectations on those users least able to carry the burden. If the leap between Blockly and text code for rules is just a little bit more, I think I’m OK with that.
tl;dr, treating Blockly as a stepping stone towards “real” code is a mistake; Blockly needs to be a first class rules language because that’s where a ton of users will stop.