PLCgen: (too?) many lines of model overview in generated code
@koenveldik and @lmoormann indicated to me that their generated code for Lock Linne contains a model overview that spans 12K lines! They like the part with the version, and the explanation of how it links to CIF concepts. But, the overview of the components and their variables/events doesn't give them overview at all. It could maybe serve as reference information. However, most of the information is also in the event functions, which is where they look for the information. There it is much more scoped. They have to scroll quite a bit to get to the relevant part of the file. Also for people that see the MAIN
file for the first time (and this is the file you open as the first file), there is so much model overview, that they may just give up reading the file altoghether.
One question we had, given this first example of how it works out in practice, is what the purpose of the model overview is, if in real models it is so long? It does not serve as an overview. Is it need for traceability? Are there other purposes?
We wondered if we would be an option to remove the components with events/variables part of the overview altogether? What would we lose? Could parts that we lose be put in a better place? We do realize that the @doc
annotations of automata that synchronize over multiple events get duplicated in different event functions, but we are OK with that.
If we can't remove it, an alternative could be put it in a separate code file. But, will users find it there?
Addresses #679