Research suggests elite scientists impede progress in their field until they die [pdf] | Hacker News Reader

Research suggests elite scientists impede progress in their field until they die [pdf]

econ.upf.edu

43 pointsteslacar7 comments

There's much more data in the Programming Languages field, which contradicts this study: Let's look at the popular dynamic languages, where progress is either hindered by the original author still working on it or stepped back, or where progress succeeded, when the original author is there or not there.

* ruby: matz still there, stagnating progress (sound but slow design, desperate implementation problems)

* python: guido still there, steady but slow progress (sound but slow design, proper management at least)

* php: rasmus still there, but thanksfully not really. with less rasmus, more progress. esp. after hhvm and php7. (unsound design, unsound community, but external pressure improved it radically)

* perl5: larry went away, decline (unsound design, unsound community)

* parrot: authors went away, death. (even with sound design, which was later destroyed by community)

* perl6: founders still there, progress. (sound design, huge implementation problems)

There are many more examples, but this is the easiest summary. esp. with the lisps, which all have a sound design and a single author.

So we have both, positive and negative influences:

* when the original design was sound, but the author went away, it will decline.

* when the original design was a hack, and the author went way, it might progress.

* when the original design was sound, and the author is still there, perfect.

* when the original design was a hack, and the author is still there, decline or death.

Read the paper! No contradiction whatsoever. The authors refer to 452 eminent life scientists (not programmers) 'who passed away while at the peak of their scientific abilities'.

Programming language development is extremely different from scientific experimentation. I honestly don't know why you'd even compare them; the authors of this paper certainly aren't making a general statement that the existence of an important person makes systems worse.

This concept has a long history: https://en.m.wikipedia.org/wiki/The_Structure_of_Scientific_...

I almost feel as if there's a corollary in the tech industry, specifically with long time employees. When an employee has worked somewhere for a long time, certain systems are considered untouchable (as if they were made by the hand of God), until the brilliant person who came up with them leaves. Then there's a flurry of development to replace unmaintainable pieces, or get to that next level of performance, etc.

Plus, new people will avoid that area, considering it plowed and looking for whatever new green pasture will net them that recognition and promotion.

So if universities want to push a field, how can they protect new ideas and accelerate there recognition? Force into every publication by the "field-leader" one contradicting page of the other, newer thought school?

I think it's less about universities and more about grant agencies and publications.

The last sentence of the abstract points to the broader problems in academics, of which this is a part:

"Intellectual, social, and re- source barriers all impede entry, with outsiders only entering subfields that offer a less hostile landscape for the support and acceptance of “foreign” ideas."

The issue is access. It's not enough to do research, or even write the paper, you have to publish it or get the grant, and that means the approval of peers. That often increases the quality of the publication or grant, because it has to run the gauntlet of criticism, but it also means that if peers don't understand something or aren't open to it, it dies before it sees the light of day. It's sort of like "if a tree falls in a forest, does it make a sound"? If research is conducted, but isn't allowed to appear in a peer-reviewed journal, did it happen?

What's really necessary in academics is something closer to what happens in comp sci, where people post to blogs regularly, and peer review is left for the hard work of proofs and whatnot, to force detailed scrutiny. Grants also need to be restructured dramatically, so that indirect cost requests are eliminated or minimized, and that grants are given by lottery or something for grants above a certain score, and that grant review panels are more randomized, to reduce nepotism within grant-panel agencies.

Reply on news.ycombinator.com