a. General context of code generators
This is written in the context of neferti.
We examine the problem in a broad sense to set the stage for smart contract generators.
The presence of higher level languages (C, Pascal, C++, Rust, Go) can be said to be a success of generators, as compilers/parsers generate machine code or lower level constructs, freeing us all up from writing in assembler or even lower level code (which would have been impossible). For interpreted languages the Virtual machines do this generation dynamically.
I have been involved in the world of generators for close to 35+ years. What I call generators are not just compilers/parsers. In the early days it was felt that code could be generated to solve any problem; especially to reduce dependency on programmers who were in short supply. These generators creates code in higher level languages from a visual representation or a template.
Where I have had success in this is by utilizing specific generators like LINC (which builds code, the database schemas, the network definitions and Job Control ) or for generating html using XSLT. I have also used tools to build parsers and compilers using yacc and lex for embedded systems. Most of the successes were domain specific and addressed specific functionality.
General purpose code generators have been problematic, similar to the challenges faced by General AI. Hence we should probably use very targeted generators in the context of NFTs. At first generating a smart contract parametrized by certain inputs for . The inputs can come in the form of JSON files which in turn can be generated by entry into a UI. It will be nice to be able to deploy the smart contract into a specific Blockchain. Another part could also generate some kind of a wallet plug-in for issuing and transferring the NFT.
b. Concrete projects that we are aware of (Oracle generators for ERC-721 variants unifying TTF and yeoman generator)