I think the most difficult part is sorting and collecting the texts. Even if you choose a wrong software - if you make a good choice of the material, you can switch to another software later.
Generally, I would recommend using MediaWiki, the software used by Wikipedia. You don't get other software tested by so many users. I have repeated experience with people suggesting using other kinds of wiki, because they have this or that additional feature, only to find out that the main feature - editing pages - is full of bugs. I prefer if the software does one thing and does it well. -- I don't know what those "social networking features" are specifically, but I would guess they don't really add much value to the project. The wiki has talk pages, and you can always install a forum or chat as a separate software.
As a first iteration I would probably do a main page with a list of topics, a page per topic (such as "programming"), and a separate page for each significant material where the main ideas could be summarized.
The difficult part will be collecting the material, describing the material, and protecting the site against vandals or mindkilled people (as self-improvement is often dangerously close to self-delusion).
I took part in a recent discussion in the current Open Thread about how instrumental rationality is under-emphasized on this website. I've heard other people say similar things, and I am inclined to agree. Someone suggested that there should be a "Instrumental Rationality Books" thread, similar to the "best textbooks on every subject" thread. I thought this sounded like a good idea.
The title is "resources" because in addition to books, you can post self-help websites, online videos, whatever.
The decorum for this thread will be as follows:
I think depending on how this thread goes, in a few days I might make a meta post on this subject in an attempt to inspire discussion on how the LessWrong community can work together to attempt to reach some sort of a consensus on what the best instrumental rationality methods and resources might be. lukeprog has already done great work in his The Science of Winning at Life sequence, but his reviews are uber-conservative and only mention resources with lots of scientific and academic backing. I think this leaves out a lot of really good stuff, and I think that we should be able to draw distinctions between stuff that isn't necessarily drawing on science but is reasonable, rational, and helps a lot of people, and The Secret.
But I thought we should get the ball rolling a little before we have that conversation. In the meantime, if you have a meta comment, you can just go ahead and post it as a reply to the top-level post.