When it comes time to select a new project or to do some new research, it's important to find out as much as possible about the subject area. Often times lots of research has already been done and written up as a journal or conference paper ("technical paper").

Why we read papers:

To see and understand what other research has been done, which is important, more even more important is to understand the advantages and disadvantages of their proposed technique, and whether its worthwhile to implement or do more research extending their solution.

What To Expect:

Papers are typically written targeting others in academia or industry, but the reader's knowledge of the material is not necessarily assumed. Papers are usually written with a short summary of previous research, in addition to what the paper proposes to contribute.

Citations are made to other papers to identify the source of material, but also to give a place where the reader can find more information about a given subject.

When I'm research a paper, I'll read it and then try to find all the papers that are referenced and see what those contain.

What Not To Expect:

While detailed pseudocode is usually present, complete source code to their entire program is almost never included in the papers. Full source code usually isn't included in the papers not because the authors are trying to protect their work, but rather that it is just their implementation working with their constraints (time, hardware, framework, graphics system, etc), whereas the reader's environment and implementation will be completely different. However, source code and executable programs are sometimes available on the website of the author.

As the reader, a description of how things work is all that is needed, because exact source code also makes it harder to read and understand what is going on. The most important thing is to understand how and why the technique works, not how every line of code is written.

Differences Between Papers and the Web

Ideally a paper will present new ideas or extend other ideas. There is a generally more formal layout/construction of the paper (abstract, previous work, new work, conclusion, bibliography/resources) and the technique is more formally described or other details referenced elsewhere. Really, with some formatting, any research or experimentation that you're working on can become a paper. A large factor is how you view and document your work.

More Information

For examples of excellent technical and research papers:

• debevec.org: Excellent work in the area of computer vision, reconstruction, high dynamic range images, and much more.
• Siggraph 2002: The best graphics papers are published at the annual Siggraph conference.
• Google: When you want everything you can get on any topic. Try keywords "siggraph", "cvpr", "eurographics".