Is it better to write "Related Work" at the start or in the end?

Question

As per title, which one is a better way to go:

to write the Related Work part right after Introduction and before System Architecture

OR

to write the System Architecture right after Introduction and dump Related Work in the end just before the Conclusion?

In my field, I have read both ways. I personally find Related Work a bit unnecessary and I usually skip them while reading a paper. It is reasonable in the sense that a researcher would have already read nearly all the related work within his or her field. So he or she may not bother to read those sections of the similar papers.

In this sense, I prefer to adopt the latter that dumps this part at the end. However, I see most of the papers that I have read adopted the former.

What are the pros and cons of these two approaches? How do I choose wisely?

I am in EECS field, but any generic advice is also very much welcomed

Answer 1

The main advantage of putting related work near the beginning is that it allows you to very clearly define what is new in your approach. I like to do that when I'm attacking a very well-defined problem and my contribution is closely related to what has gone before; e.g., if I've modified algorithm X so that it runs twice as fast, then I must review prior work on algorithm X. I try to work the review into the flow of an expository structure -- i.e., introduce prior work as you move naturally through the introductory and motivational concepts.

The disadvantage of putting related work near the beginning is that the reader must wade through it before getting to the important part of the paper -- your contribution. It also may interrupt the flow of exposition. So I put the related work nearer the end when possible; for instance, if I'm solving a problem nobody has considered before, or exploring some newly discovered phenomenon.