A good topic is one from which you can get a "take home message". Also, The bulk of the paper/talk needs to be explaining the take home message, so you need to be able to explain how it works to a non-science and science audience.
Consider the topic "bar code scanners". The science in this topic is how a scanner can read the code the bars represent, regardless of the orientation on the bar code. A discussion of how the computer, having read the bar code, looks up the UPC in a database is a discussion of databases and not bar codes.
One example of a particularly poor topic is quantum computing. A typical quantum computing paper goes something like "quantum computers will avoid the one-atom limit (false), uses qubits for data storage and computation, can do true multitasking (misleading) and has applications in cryptography." The science in the paper and talk is all in explaining how quantum-specific algorithms work, which is very difficult and results in a science-free paper and an uninteresting talk.
Topics need to have some depth. Consider the topic "how fireworks produce different colors". If the content is limited to a brief discussion of emission lines (for the non-science audience) and a list of which chemicals produce which colors, the content will not be sufficient to interest your science colleagues during your talk.
Astronomy topics are difficult to do, because the science tends to be
vague. Also, avoid topics which are chiefly historical in that your
paper reads like a story rather than a science paper.
To cite this page:
Principles of topic selection <http://www.physics.ohio-state.edu/~wilkins/writing/Assign/topics/topic-sel-princs.html>
[Saturday, 16-Dec-2017 17:24:14 EST]