One of the basic expectations about academic writing is that it will only refer to published, or publically available, sources. This is because of the reader-orientated perspective that academic writing is expected to adopt - a source that is not publically available in a library is by definition likely to be hard for your audience to access, and so not one whose use by you they can evaluate. Something your tutor said in a workshop, or wrote on a handout, is not a publically available source, because the accuracy of your reference couldn't be checked by someone who hadn't attended the module, such as the External Examiner.
The web allows access to a great range of very different kinds of material, in a way that makes the distinction between published and unpublished sources very much harder to draw than it used to be twenty years ago. It is becoming increasingly acceptable to include references to online sources, but making use of web-based materials shifts the burden of assessing the quality of these sources onto you as a writer, since most websites have no system of peer review. The same applies to encyclopaedias, such as Encarta. You need to assess for yourself the appropriateness of making use of web-published sources, given the nature of the topic on which you are writing - for recent SF writers, such as William Gibson or Philip K Dick, websites might represent good sources, but at the moment (2005) there isn't a great deal of useful secondary literature available online for anything published before 1960.
Similar points apply to the use of magazines and newspapers as sources. Pieces of journalism, unlike articles in academic journals, are not subject to peer review, so they may represent good sources for contemporary topics, where nothing more authoritative has been published, but will not usually represent authoritative sources for historical work. You will sometimes find literary critics making use newspaper and magazine reviews in their discussions of older books, but this is because they are discussing the response of the reading audience at the time of the book's publication, which these reviews represent, rather than holding these reviews up as authoritative judgements about the book.