Some things would be better, but some notably worse. I am definitely limited by Discourse in some important ways for my purposes, but I think one of the primary ones is simply navigation, which seems semi-easy to address if the will and thus dev resources are there. I have an idea I’ll be proposing (with some funding) soon that I hope will help.
I also think it’s valuable to think about the idea of extending the Moderator concept in particular knowledge generation contexts into being something more like a “Curator”. Solo knowledge generation combines generator and curator roles/activities. But collective knowledge generation probably requires - or at least would significantly benefit from - trusted people whose role or part of their focus is to quote, promote, organize, polish, and otherwise surface content, ideas, etc. that others generate, and make it more accessible to those who could benefit from it. In theory the wiki functionality in Discourse could be part of the solution here, but it would need some adjusting.
There are also lots of interesting things to think about in terms of collaborative knowledge generation, “digital gardening”, etc. Is the goal to end up with singular documents that represent a collective perspective and understanding? Or to represent multiple perspectives simultaneously? Or some combination of the two. I can see many possible approaches to handling these and other questions.
Ultimately the challenge is that CDCK probably is not that interested in these uses for Discourse (which I can understand, their utility and thus potential for revenue is much more nascent and unclear at this point). And meanwhile few - if any - other platforms that do focus on knowledge generation, e.g. Wikipedia/MediaWiki, really address the conversational, discussion aspects well enough, and especially the interaction between the two. In a perfect world high quality discussion over long periods of time could incrementally add to a distilled output of knowledge, ideas, perspective(s), easily and fluidly, maintaining attribution while also allowing for well-formatted, consistent “document”/artifact output that would actually be enjoyable to read. Wikipedia is again a good example of the current model that works, but it’s highly imperfect and newer tools and methods are needed to get beyond simply documenting knowledge to actually generating insight, exploring new ideas, etc.
Discourse can be used for these things now, in some ways more easily than others. But it can be done, it has most of what is needed…