I also implemented a small caching framework in Racket so it will automatically download whatever screen-scrape source material is missing from its cache directory. Downloading each source document only once and then caching them locally ought to be more polite to the people hosting the package repositories.

This is very cool - I hate my hacky code around the Make-get, Racket-http-get and local-file-read and wanted to go back to caching, once I got slib support running and learned whether this needs features additional to the Raven processing.

I'd like to thank Frank for the excellent work in adding a RavenSc.com scraper which I did not figure out how to do - and also for writing a Makefile.

You're welcome ;) I'm currently on the slib part - with no clever idea yet whether/how to consolidate the package redundancies between the various repos/package-managers. Poco a poco, as the Italians say ;)