Hacker News Reader

Comment by probdist | Hacker News Reader

The doctoral thesis already is already essentially universally several papers stapled together in experimental science and engineering. The shortest compelling way to explain your results is a paper. Furthermore, the notion that your work over 3-8 years is one cohesive story has been violated by every graduate student I've ever met.

View on HN

My thesis (in math / CS) was essentially several papers stapled together, except that I made the papers into chapters, expanded/rewrote bits and pieces, and added narrative to glue the chapters together.

This took about six months. In retrospect, I think this time was well spent. I developed my understanding of the material quite substantially by reworking it. But doing so was certainly not nearly as fun as writing new papers.

Same here except that for me it feels like wasting a lot of time, unlearning all my other skills and not staying up to date with current trends.

Having to rewrite polished and reviewed papers is aweful and not publishing stuff (except in a thesis) is unacceptable for many institutions, so I think making papers to chapters is the way to go as long as there isn't something completely different (e.g. no prose but making all data, hardware, software, documentation etc open source)

Yep, my math PhD was literally 4 submitted papers placed together with an introduction section to the topic area. This wasn't 100% common in my department (US public state university), but my advisor strongly suggested it as the most efficient way to go.

Depends very much on the field. Everyone I knew in experimental physics was focused on a single experiment or type of apparatus, generally an attempt to measure a single number (or rather, because it was all "physics beyond the standard model", an attempt to set a new limit on a single number.)

What constitutes a staple-thesis varies a bit though. My advisor required 3 top-tier papers in a well-defined topic. Papers in other topics or workshop papers does not count.

Reply on news.ycombinator.com