First of all, thanks for the feedback on the topic, something to think about.
Why is that a problem? I love nodes, I know how to use them… I treat nodes/nodetypes as a sort of “virtual DOM”, a data abstraction that allows us to attach semantical meaning to the data when there is such meaning and to tighter control the rendering, querying and aggregating the data as opposed to a blob of HTML that you see in most other CMSs. It’s a base building block that you can use according to the task, so if you have just a blog, I feel like it’s perfectly fine just to have Text + Headline + Blockquote + Image + Grid nodetypes, they describe as much semantics as there are in this piece of content, and if this content is not deeper structured and requires just aesthetics, then so be it.
Anyways, I’m no good at having such theoretical discussions in written form, so if we’d find ourselves in strong disagreement here, I’d love to move that discussion to a Hangout.
Not to get too theoretical, I’ll try to write some user stories:
A) As editor I want to be able to quickly insert other nodes in between paragraphs
(Visually it could look like that you hover a paragraph, and next to it you see a ‘+’ button. Then it does what is described in solution 3.)
B) As integrator I want to have more data represented as nodes, e.g. heaving headlines as a separate nodetype allows to build up ToC, customize rendering depending on context, insert anchor links etc.
C) To achieve B), as editor it must be really easy to select a paragraph and convert it into a separate headline, blockquote etc. Then it would feel like a combination of solutions 1 and 3.
So yes, I think we should start with solution 3 and see how far we’d get away with it.