CSCI 450 - Artifical Intelligence: Types of CS Pubs

1.  Pre-Print, Post-Print, "eprint"

Many articles and some conference papers start as a "pre-prints."  

Pre-prints are publications that have typically not undergone rigerous peer review, but rather have been submitted to an open access repository (such as ArXiv; or an institutional or a shared discipline themed repository) prior to publication in a journal or conference proceedings.  Not all pre-prints make it into journals or conference proceedings.  Once a pre-print has been submitted review with a specific publication (think journal) it may be termed a "post print" where the publisher makes the item available as a "preview" before the paper is placed into the publication and receives standard citation information (issue, page numbers, etc.) if the paper is accepted for publication.

The entire idea behind "pre-prints" in computer science is to make ideas and results available well before they would otherwise appear in a journal or perhaps even a conference. 

2.  Conference Paper

The Computer Science discipline has a very active conference/presentation mechanism for making findings and research ideas known to the computer science community.  Because of the relatively fast time to publication/dissemination in a conference proceedings, this is where many papers get their first wider dissemination if the field of computer science.  Conference proceedings are typically available by the start of a given conference.  Submissions to a conference typically start about 6-9 months prior to the conference itself.

3.  Article

Many (but not all) computer science articles appear in publications of the ACM or the IEEE Computer Society.  The ACM and IEEE Computer Society are professional organizations specializing in the field of computer science and facilitate its progress through professional publications and the organization of annual or themed conferences or special interest groups.

Articles appearing any ACM and IEEE publications (and other various CS related research or academic journals) will undergo a "peer review" process.  Normally the peer review process is also "blind" in that an author submits an paper for publication, the editor of the publication reads it and if it's thought suitable for possible publication the paper is sent along to 2-4 other reviewers that review the paper for style, findings, methodology, general approach, etc. and offer commentary and feedback along with a recommendation to publish, publish with revision, or decline to publish.  This process can take 6-18 months depending on the publication, the number of times the paper is revised and the publication schedule itself (monthly, quarterly, semi-annual, annual).

4.  Technical Report

5.  Book Chapter

6.  Book

The body of computer science literature does include books; just not very many relative to other means of diseminating ideas and findings.  

These typically are summative and overarching in nature and represent a more philosophical approach to the discipline of computer science, or some particular aspect of the discipline rather than the technical means by which something is actually accomplished or the latest findings of a particular research effort.  Books are typically "touchstones" by which an idea is cited as the basis for a subsequent or particular paper or idea.  

Donald Knuth's The Art of Computer Programming is an excellent example; as are various classics as determined by the ACM.