Up: Nanoc

This note is a brain dump first and foremost. There is no work happening on Nanoc 5 at the moment — at least no coding, but definitely some thinking.

That said, here is what I would love from Nanoc 5:

• True parallel compilation.

• View components. See also: It is hard in Nanoc to use reusable view components. This might mean making the distinction between partials and layouts more clear.

• Usable as a library.

• Easy to install without needing Ruby. This might mean a total rewrite, and I don’t know how existing filters would still be usable. See also: Nanoc’s use of Ruby is not an appealing choice these days.

• Minimal memory usage. Especially for big sites, loading everything into memory is slow and causes performance issues. See also: Nanoc’s in-memory approach limits scalability.

• Focus. Remove checking and deploying, or at least have those as separate products.

• Support for extracting metadata from items that don’t use a traditional frontmatter. See Embedded plaintext metadata formats.

• Intermediate formats (e.g. Markdown ASTs). This would allow more efficiently converting from one input format to multiple output formats.

• Multi-part documents. Though this might overlap quite a bit with intermediate formats as described above.

• Dynamic content generation: allow a document to have multiple outputs from embedded content like Mermaid diagrams or TeX formulas.

• Pure. No preprocessor. This will require some work on figuring out how to replace the preprocessor, because it’s rather essential at this point.