What is to be done

The future of Complex Science

by Mikhail Simkin

…complexity of complexities; all is complexity.

Figure 1

Figure 1 shows the distribution of complex-science posters in the rank-frequency representation introduced by Zipf. The discussion list obviously achieved a remarkable success, as is evidenced by the similarity of the posting distribution to Zipfian distribution. However, we should not close our eyes on the fact that the correspondence with Zipf’s law is not perfect. There appears to be a kink around rank 15 with posters above and below that rank following Zipfian distributions with different slopes (see Figure 2).

Figure 2

Fortunately, there is nothing here what can not be fixed. For example, we can stick to the slope dictated by the bottom of the distribution and challenge our top-rank posters to match. The proposed distribution is shown in the Figure 2 by the blue line. To make this happen our most prolific posters should post as specified in the table. The rest of the members of the list should limit their participation to reading those postings.

Post one for the Zipfer

Rank	Poster	Actual number of posts	Should be according to Zipf	Additional posts needed to achieve Zipf
1	Stanley N. Salthe	764	6661	5897
2	Sungchul Ji	532	2383	1851
3	Gavin Ritz	378	1306	928
4	John MIKES	319	853	534
5	James N Rose	268	612	344
6	E. Taborsky	230	467	237
7	Don Mikulecky	223	372	149
8	Russell Standish	193	305	112
9	Yaneer Bar-Yam	169	256	87
10	Guy Hoelzer	124	219	95
11	val	124	190	66
12	Paul Prueitt	116	167	51
13	Peter McBurney	116	148	32
14	John McCrone	113	133	20

Although there is not the slightest doubt that our most prolific posters are perfectly capable of the task, alternative approaches are worth exploring. We can stick to the slope dictated by the top-rank posters (magenta line in Figure 2). To make this happen our most prolific posters should abstain from posting and divert their energy into recruiting 6,000 new members.

Also note that we are not bound to extrapolating the top or the bottom of the distribution. We can select an intermediate slope. The most attractive value seems to be 1, because this value of the exponent was envisioned by Zipf himself. The proposed distribution is shown in Figure 2 by the green line. Our most prolific posters should continue to post and in addition 1,300 new members should be recruited.

Now we need to decide which way to go, make a commitment, see it through, and together we will achieve a perfect Zipf.