Hacker News Reader

Comment by TruthAndDare | Hacker News Reader

> Derek de Solla Price is an interesting character. 50+ years ago while stacking and organizing scientific journals by year, he noticed the stacks grew in height in a logarithmic curve. From there, so much.

I'm afraid I don't understand, especially not the last sentence. Could you explain?

View on HN

In 1951!

"The number of scientific papers published each year may be taken as a rough indication of the activity displayed in any general or specialised fleld of research... The growth factor of the exponential portions is such as to double the amount of literature every ten or eleven years in both the general and specialised cases."

http://garfield.library.upenn.edu/price/pricequantitativemea...

This is 15 years before Moore's Law and he's predicting the growth of science and citation relationships. His books basically lay out PageRank and network effects.

Thanks. That has much more sense than "grew in height in a logarithmic curve."

More detail but less beauty. I'll take a second whack at it. Ignore the details, consider the premise. The university library was under renovation. He wound up in charge of safekeeping a pile of journals. So he stacked them neatly by year and decade:

  - The 1900s Physics Journals: 1" tall
  - The 1910s stack: 2" tall
  - The 1920s stack: 4" tall
  - The 1930s stack: 8" tall ...

I'm reminded of Feynman's tale of the Bell Labs scientists who had a view of suspension bridge construction. Marked the progress week by week on the glass, watching the cable slowly fall into parabolic shape.

This contradicts the earlier description of "growing in a logarithmic curve" -- it's an exponential curve. A logarithmic curve would look like this:

    - The 1900s stack: 1" tall
    - The 1910s stack: 2" tall
    - The 1930s stack: 3" tall
    - The 1970s stack: 4" tall
    - The 2050s stack: 5" tall ...

"Ignore the details, consider the premise." :) Have an upvote and go read the guy's Wiki article.

https://en.wikipedia.org/wiki/Derek_J._de_Solla_Price

The "logarihmic" premise was false from the start, that's why the people (I too) couldn't understand what you initially wrote.

If you stack books horizontally on shelves by year it would form the textbook exponential curve. If you stack them vertically on the floor it makes the textbook log curve. It's a visual metaphor. Apparently not a very good one.

http://science.larouchepac.com/gauss/ceres/InterimII/Arithme...

http://4.bp.blogspot.com/-47zv3839-6U/UP37FwkfOhI/AAAAAAAACU...

EDIT: Wow, this is still ongoing? You can mount a hockey stick horizontally or vertically on a wall, guys. Think about it.

So there are 10 magazines from 1910s, 20 from 1920s and 40 from 1930s. How would you arrange them so that I can recognize the log curve?

> If you stack books horizontally on shelves by year it would form the textbook exponential curve. If you stack them vertically on the floor it makes the textbook log curve.

It seems to me you have that exactly backwards:

If you stack books horizontally on shelves, which each shelf corresponding to a same-size range of years (e.g., a shelf per decade), with earlier periods on lower shelves, it would form a quantized approximation of the textbook logarithmic curve.

If you stack books vertically on the floor in similar periods with earlier periods on the left, it would form a quantized approximation of the textbook exponential curve.

Found it, I think:

https://en.wikipedia.org/wiki/Derek_J._de_Solla_Price#Scient...

Price's Law generally predicts the growth of social networks, which is why it's important today, but he discovered it in the context of number of scientific citations.

Reply on news.ycombinator.com