The issue with the documentation is the longer term discussion between Torben's approach (openacs-handbook), writing all from scratch, tailored for different users and the old way (http://openacs.org/test-doc/
The former has the intention of a categorized approach where pages are structured and presented differently depending on the target audience.
The latter has the intention of being a book.
My opinion was and still is that we should have the BOOK as the primary reference and add categories in a smart way so that we do have multiple entry points into it. But others might disagree. And due to this impasse I don't know where to fix documentation or upgrade it and most likely others don't either. So not much get's done.
If you go to /xowiki you will also realize that we have package documentation there twice (http://openacs.org/xowiki/ACS_Admin vs. http://openacs.org/xowiki/acs-admin). In two separate categories. User Installation. There exists a couple of pages, which one will be deemed current?
Another example: If you go to http://openacs.org/xowiki/en/openacs-system-install (which luckily is prominently linked from the main page), on the left hand side you see a ton of pages and options, all flat in one category. Here the category approach clearly fails in my understanding and having sections and subsections would be much better.
Usually I am a person who would get his hands dirty if he feels pationate about it as I do about user documentation. But there exists a documentation team which has different approaches, and this shows. Furthermore, not everything old is bad (read: /test-doc). It is just old and needs updating. But the old documentation was well structured.
Therefore, we might have to open the box of pandora again and discuss what the documentation should do for us, then talk about structure and last but not least talk about implementation. Not to mention the fact that e.g. I do all my user documentation in screencasts and not in writing anymore.
Currently the best way to learn OpenACS is to use google search. Why? Because they also index the forums smartly and we have a lot of hidden gems in there. Sadly it also increases the noise. Maybe .LRN 2.4 could be a documentation release (as it was with ZEN) while OpenACS 5.4 will be the "clean up bug-tracker" release? (the differentiation is purely to suggest who is in charge. .LRN consortium for documentation, OCT for bug-tracker cleanup)