3 ms·

I kinda of worry that if this site succeeded, it would slowly do to papers what news aggregators and the internet already did to news. Grind and incentivize the

I kinda of worry that if this site succeeded, it would slowly do to papers what news aggregators and the internet already did to news. Grind and incentivize them down into clickbait with no meat, with as few things as possible that people won't immediately understand. For example, looking over my own papers for examples, a one page note [1] got more scites than a paper I contributed to that introduced several techniques cumulatively dropping the projected cost of simulating a chemistry thing by more than a factor of a million [2]. That's a cherry picked example, the pattern isn't that consistent, but having seen what happened to news sites it still worries me. ...not that the incentives for how to structure and publish papers are that great to begin with. [1]: https://scirate.com/arxiv/2106.11513 https://scirate.com/arxiv/2106.11513 [2]: https://scirate.com/arxiv/1805.03662 https://scirate.com/arxiv/1805.03662

Syzygies 5y ago
On the contrary, I've sat on Ivy League tenure panels where humanities members intoned solemnly about H-indices, oblivious to the differences in value of the Lira and the Pound. I also consider much of baseball statistics to be junk science, but at least they have "wins over replacement" that could help with the endless "does author order matter in this subfield?" debates. I welcome a platform that could allow more sophisticated mind-mapping. I've argued to no avail for years that MathSciNet should be open to machine learning, try to make itself a premier playground for machine learning, but it's an AMS cash cow and leadership prefers a 19th century model. Out to lunch recently with a former AMS president, he suggested I should do this for the ArXiv. That's not the point; one wants to encourage significant existing projects to land on our data. Math is pretty much bifurcated between tightly-bound field tribes, only reading papers within their field, and roving players who recognize ideas in any form and chase them across fields. Of course, tribal mathematicians would dispute this. Better mind-mapping would reveal the connections between what's out there, rather than imposing a tribal mindset on its organization. It would be good for mathematics to change this culture. It takes more courage than one might imagine to be a roving player. At hiring time, most departments let the tribes take turns. Just like growing up on the streets, pick a gang for survival, if you want to get hired.

Syzygies 5y ago

On the contrary, I've sat on Ivy League tenure panels where humanities members intoned solemnly about H-indices, oblivious to the differences in value of the Lira and the Pound. I also consider much of baseball statistics to be junk science, but at least they have "wins over replacement" that could help with the endless "does author order matter in this subfield?" debates. I welcome a platform that could allow more sophisticated mind-mapping. I've argued to no avail for years that MathSciNet should be open to machine learning, try to make itself a premier playground for machine learning, but it's an AMS cash cow and leadership prefers a 19th century model. Out to lunch recently with a former AMS president, he suggested I should do this for the ArXiv. That's not the point; one wants to encourage significant existing projects to land on our data. Math is pretty much bifurcated between tightly-bound field tribes, only reading papers within their field, and roving players who recognize ideas in any form and chase them across fields. Of course, tribal mathematicians would dispute this. Better mind-mapping would reveal the connections between what's out there, rather than imposing a tribal mindset on its organization. It would be good for mathematics to change this culture. It takes more courage than one might imagine to be a roving player. At hiring time, most departments let the tribes take turns. Just like growing up on the streets, pick a gang for survival, if you want to get hired.