I've been searching everywhere I know for research concerning the topics listed below, but have been unable to find anything specific. The best I can come up with is general literature. Specifically I am looking for research on the following subjects:
The effects of feedback on an artifact given in-context of artifact versus out-of-context, and the reasons behind designing UIs that support either. A concrete example is feedback on a report; why would one choose to provide functionality for writing feedback in the report margins (in-context, e.g as when commenting PDF documents) instead of in a separate document (out-of-context)?
Context switches in workflows and why they are bad/good in general; why do we want to avoid them in user workflows (and do we?).
Related to 2: theory supporting integration of related workflows into a single computer supported workflow. A concrete example is Microsoft Word which has a feature to edit images inside the application itself even though it is a word processor. Why is this preferable compared to e.g. editing the image in an external application and inserting it afterward? And is it (always)? Are there special cases where it may be counter-productive?
I was advised to post these questions here as part of my broader focus on feedback on COGSCI SE: http://cogsci.stackexchange.com/questions/5763/research-on-properties-of-effective-feedback