2. I see, but I still think a way of finding url's which don't have any metadata set at all would be nice.
Well, if you really mean "By URL", that's what sh404SEF already allow. If you actually mean, "I have to look at each URL, understand that it's a Joomla content URL, or a menu item or a catgeory blog, and check if the corresponding Joomla content item has some meta data or not, it'll take minutes before the corresponding SQL queries can be run.
Of course, that can be organized as a one off thing, like a "Check my pages" button that you run once in a while or otherwise. I have things in store for that in the future!
3. Perhaps it's an idea to make a plugin for OSMap, so that it gets the URL's from sh404sef?
That's exactly why there's no sitemap system in sh404SEF right now. You usually don't want to simply dump all URLs into a sitemap. A sitemap is supposed search engines what your "good" URLs are, and so you want to be sure you only feed them with no duplicates, pages iwth good metas, etc
So you have to be able to process them. Dedicated extensions such as JSitemap pro have (very complex) systems to extract pages from extensions, with dozens of plugins just for that. It can also get very cpu intensiev to generate a sitemap, espcially on large sites (and a sitemap is of nearly no use it you don't have a large site).
I don't want to get into that at all, but I do hope to find a simpler way.
Rgds