It is different for each field and each individual researcher, but generally you should only publish original ideas. But that can also be the application of an idea to another area or an incremental improvement to what you've been previously doing.
While you can't publish the same work twice (and you'll get in trouble if you get caught) you can break up a big project into chunks that can get published in different journals and conferences.
About a year ago I completed a large research project and have published different results in two conferences. These weren't the same paper though. One, for example, was much more focused on applications and methodology of the research, while the other was based on the technical advancements our group made in the instrumentation and in finer implementation details.