## Idea Exploration * Not enough to just write an idea - Easy to come up with ideas, but not good ideas * Need to force your mind to: - Explore the ideas deeply - Ask the five whys in problem solving, problem framing, context, and more - Hard to get your mind to go there sometimes, to do it (at least for me) ## Fictional Abstracts * An abstract contains a research paper in miniature: - Context - New argument or concept - Description of work done - Summary of results and conclusions * In most research you have a good sense of the abstraction _before_ doing the work - The _form_ of the "answer" is intertwined with the "question" - Research results are almost entirely (socially) constructed in CS, not discovered - We often even know what results we expect or would like to see - Even in ordinary science, a hypothesis is key * We'll explore this more with Rittel and Webber ## Each good research agenda should yield at least 3 core papers * Not an exact number, just a rule of thumb. * If it has less than 3 core papers of material: - Not broad enough - Less likely to contain a fundamental shift in thinking * Really speaking should result in many more than 3 papers: - Each of the 3 core papers should explore orthogonal aspects of the agenda - Each will require followup, in the form of more papers - For example, in applied areas, each core paper would likely benefit from one followup on theoretical foundations and one followup on engineering/implementation/applications * However, beware of mechanistic changes that force re-thinking of everything - Easy to devise a superficial idea that requires many papers -- examples? - Such examples are often not worth exploring - How do we tell them apart from fundamentally-important, new advances? ## Write or die (exercise) * We'll be using https://writeordie.com/ - Change settings to 400 words, 15 minutes, 5 sec grace period, kamikaze mode - Remember to copy the text from the Write or Die buffer to your doc after each * Explore the three research agendas, one at a time, through fictional abstracts - For each research topic, write 3 fictional abstracts - That's _9 total_ fictional abstracts you're going to write in 45 minutes - Each abstract should be distinct / non-overlapping - The abstracts should span the research topic - Good research agendas should have more than 3 papers worth of research to be done - 45 minutes in total, 15 for each of the three ideas. * Then we'll pair up and provide peer feedback