One of the core goals of Civilization is to support better indexing than current forums do. Maciej (archived) talks about how fandom uses complex tagging schemes combined with the wisdom of the crowd to determine the appropriate keywords for an item. Our problem is that in all likelihood Civilization, at least as it's meant to be used by FortForecast, will only host between 300 and 3000 people on an instance if it's successful. That means we have a lot less data to rely on wisdom of crowds for, and thus have to put more effort into active indexing. You have to have features to support indexing, and make sure people somehow go through the trouble of indexing work to make it accessible.
I remember someone remarking that it's incredible to think there are furry porn sites whose nitpicky zealous indexing is better crafted than the systems we use for the most important knowledge in our society. One of the major things that is meant to set FortForecast apart is a healthy respect for the task of building an academic library rather than a holy book or canon.
Contents
Thread Purposes
As part of the design for Civilization Hypothesis noticed that how to summarize a thread often depends on what the thread is, structurally. Not all forum threads have the same purpose, so he went out and cataloged a bunch of them to see what kinds of things people create forum threads for. From this he derived a set of thread purposes which a user can tag their thread with to get indexing benefits at thread creation.
Proposal: Structured Post Markup
Inside a thread depending on its type it can make sense to alert machines to the presence of certain data and give some indication of how to collect and format it. For example in a coworking thread it could make sense to markup the purpose of a time session and how long a person expects to work before they post again. A program could then comb through the thread and create an artifact automatically from the data markup in posts.
Tagging
Mandatory Tagging
One of the best ways to get high quality indexing is to use a controlled vocabulary (archived) as part of a taxonomy. This requires some work from contributors and upkeep to make sure the indexing is accurate and consistently applied.
Crowd Tagging
Another way to get high quality indexing is to let individual users tag the work of others and then use the collective wisdom of many to determine a sane set of keywords for the document. While this method is viable it requires many users to work well. Civilization intends to support this pattern because it is both useful to individual users to be able to tag works for later retrieval and because it helps allow the generation of new tags. Furthermore because in the context of FortForecast users are expected to make high quality contributions even "ad-hoc" tags should be quite useful to people searching the database.
