This may result in issuing a "bad patent" ... i.e., a patent that (rather than promoting the progress of science and the useful arts - as patents are meant to do) may needlessly impede real innovation or technologic progress.   Once issued, a patent is presumed to be valid.  Though this presumption is rebuttable, it can be expensive and time-consuming to challenge a patent's validity in court.  The CPR project is designed to help ensure that bad patents do not get issued in the first place.

There are two important criteria: (1) the content of the reference must relate to one or more claims of the patent application; (2) the date of the reference must be earlier than the priority date of the patent application (ideally, more than one year earlier.)

Under the relevant US patent statutes, a publication that discloses a claimed invention will be an absolute bar to patentability if it is dated more than one year earlier than the priority date of the patent application under discussion.  

Not necessarily.  If the publication is dated before the priority date (though less than one year prior), it may show that the applicant was not the first to invent.  The USPTO may use such a publication to challenge the applicant's claim to be the original inventor of the claimed subject matter.  However, a prior art reference dated subsequent to the priority date of the patent application will not be considered by the USPTO.

In general, the effective date of a publication is the date that it became sufficiently accessible to the public to qualify as prior art. For the most typical forms of printed publications (books; journal articles; etc.) this would be the publication date of the book, or journal issue containing the relevant text. For less typical forms of prior art publications (dissertations, for example) it might be the date that the material was included in an index generally available to the public. Issues such as publication dates for web pages, or commented source code, are probably not yet susceptible to easy definition given the rapid changes in technology, but if one of these forms of prior art is identified as a strong candidate for submission in the community patent review process then the reviewer proposing this material should specify the a date when the material is known to be generally accessible, and provide any relevant commentary about how the date was determined along with the material.

Since claims serve such an important legal function (specifying the metes and bounds of the invention), their language and form has evolved to be somewhat arcane - taking advantage of evolving bodies of relevant case law. The task of a reviewer here will be simpler if some general principles can be kept in mind: For one thing, those drafting patent applications will often include sets of claims phrased to include multiple statutory classes - method claims, and apparatus claims, for example. The reviewer should attempt to focus on substantive differences between claims or sets of claims, and the task at hand may be simplified if the reviewer recognizes that parallelism between claims will allow similar claims to be thought about together. It will frequently be the case that a single piece of prior art will be equally applicable to claims phrased in several different ways.

No.  Even if the reference shows only some elements of a claim it may be an important reference.  For example, two such references might be able to be combined to show all elements of some claim.  Such references may show that a claimed invention is obvious.

Obviousness in the patent sense is not the same as obviousness in the ordinary sense of the word. For example, many simple inventions, once described, might be thought obvious in retrospect. But hindsight is not permitted in making a legal determination of patentability. Therefore, for the purposes of the CPR project, it is best to focus on exactly what a prior art reference shows - and to compare it to the claims of the patent application under consideration. If the reference shows some portions (elements) of a claim, it is worth considering for submission. If it does not show (or very strongly suggest) some claim elements, it should probably not be submitted. Obviousness in the patent sense is a determination that while no single prior art reference discloses all elements of a claim, the differences between the claim and those combined references would not be significant to a person of ordinary skill in the art.

No, the Community Patent Review Process is specifically focused on identifying and commenting on prior art. Comments directed to other matters (such as the adequacy of the disclosure) will not be considered by the US PTO.