My Perspective on How (not) to do Research

Ok this is a tough page to write.... take it as a blog with some personal thoughts on some aspects of academic and industrial research that i think are wrong, and aspects that i think are right.
I do change mind, so you may see these opinions change over time. In fact i think if we do not change mind over time, it's hard to say we have been learning something.....


As the rest of the web site, it is rather empty, i will fill it up as opinions come to mind. A lot of this is specific to the IT field, i do not know enough of other fields to make comments.
So here is a random collection of thoughts of things i like, or that i don't like, or on how i think things should or should not be done.

Funding research

  • I am puzzled by "innovative projects", high risk projects, etc, that have a detailed long term plan defined a priori. Innovation can hardly be planned and constrained. I think we have enough evidence of this. 
  • This is why i do not believe in detailed Gantt charts, Descr of Work that cover more than 6 or 12 months, detailed grand visions..... or calls for research project proposals that are very detailed on the target outcome
  • To this end, it helps to remember that very few inventions have been used for the purpose they were originally intended.... 
  • This is also why i believe that funding agencies should give money to creative people. To a large extent this includes young people, people who are flexible, and can embrace change.  
  • Funding should be given based on assessing people and the high level ideas, not based on a plan. Once given, people should be allowed to do what they think is right. following a plan written 2 years earlier does not make sense. Then one can periodically evaluate the progress.
  • In general projects with a small set of partners and where the coordinator can have decision power seems to work best. 

Research labs and technology transfer

  • After joining a University, i am convinced that students are the key to research. Not that students have the monopoly: innovative ideas can come from people that are creative and flexible, open to change. In general young people have this characteristic, also as they come with a fresh perspective
  • Incidentally, this is also why I believe that changing areas is good. You bring a fresh perspective as opposed to be constrained in your mental model. 
  • This means that research labs, to function well, they need to have change and openess. What i mean specifically is, less structured / stable personnel and a lot of students flowing in (for a few years) and then out. There are many ways in which this can be done. The large majority of research institutions does not get this, and i think the results speak for themselves (results in terms of actual impact, not in terms of ability to get funds from the government, which is not necessarily related).
  • As for technology transfer, i think it works if you transfer - or anyway connect - people, and have people work together for long periods. And you transfer knowledge more than technology. The other way to "transfer" is to do your startup.

Reviewing research

ok, this is reeeally tough.... bordering the mission impossible. Especially if you want "metrics".
Elsewhere on this site you may have read my thoughts on peer review and others have bashed citation counts.... so what should we do?
i have no clue. One option is to see if research resulted directly or indirectly in a startup or a product that people use. This however takes time to assess and  there may be reason why good research does not result in that, maybe. Maybe. 
it is also not easy to assess. The liquid journal concept is one of the way we hope to help on this. There, you can read our constructive ideas on this matter.
  • Instead, i think it is unlikely that reviewers can properly assess novel papers. This is why when chairing conferences, i tend to see if at last one reviewer really liked a paper, as opposed to looking at averages. 
  • Another thing i am scared about is communities of people doing useless things. Yeah, we all have different opinions on this and we may all be wrong. Yet, there are communities - even large ones - that grow and self-sustain themselves, the create conferences and journals, and they cite each other.... and they get great reputation metrics. Impact in used tools and interest generated in other communities can help identify this
  • Some paper writing practices that i tend not to like (with exceptions, but few)
    • papers that assumes a future technological or social state (eg, 10 years from now, there will be this situation and therefore we will have this problem) and then try to solve a delta-problem on it.
    • papers that contain lots of formulas and theorems but where is it not clear who concretely benefits from the research and how

Universities

  • Many people in universities believe they will change the world through research. But the greatest contributions can be done by good teaching and mentoring of students. 
  • Universities should enforce biodiversity
  • Universities should leave people free to innovate. This implies not having hierarchies where older professors thell young people what to innovate on.
  • overall, professors are not as bad as people think, at least for the sector i know better (IT). On average there are many interesting, passionate persons that work out of their wish to innovate. Workloads end up being much higher than in companies (where people think they are busy all the time.... they should try to be a professor for a while.....) 






Comments