I asked the wikipedia research list when one might publish results. Many of the ideas for wiki in scienc were spawned from Samuel Klein's response. email He says, I don't know... how about:
You have a good project idea someone should do. You publish it.
You know some people doing interesting work in the area who need x,y,z to tackle such a project, and add that.
You start a project. You publish a pointer and project name.
Some collaborators join. You publish names.
You get a target to take data from, have a meeting, and publish.
You finalize procedures and start implementing. and publish. ★
You get first data. and publish. ★
You get context for the data. And publish. ★
You find time to look at the data, organize the context, add a summary, and publish. ★
You compile a full schedule of data, and run analysis, publishing your error logs and lab notebook pages on the fly. ★
You give a paper bag talk with slides (and publish)
You draft an abstract for peer review (and publish)
You finish an abstract and submit it for review (a. p.)
You get feedback from the journal you submitted to (a. p.) and revise (a. p.)
You get included in a major quarterly Journal, with polish (a. p.)
You get public commentary, cites, criticism; and make better talk slides (a. p.)
You add suggestions for your students or others to extend the work in future papers (a. p.)
Various fields adopt various subsets of the above; most have only a handful towards the end. The stars mark items for which federated wiki most directly applies.