Show that we need PageRank

[vmarkovtsev]

Related to #2
Simple removal of all edges from Go to * makes:
| 22.76% | java |
| 15.73% | go |
| 14.31% | c |
| 13.46% | c++ |
| 6.83% | c# |
| 6.12% | python |
| 5.70% | scala |
| 3.98% | rust |
| 2.93% | php |
| 2.02% | objective c |
| 1.54% | ruby |
| 1.07% | swift |
| 1.00% | node |
| 0.65% | kotlin |
| 0.40% | r |
| 0.30% | matlab |
| 0.28% | haskell |
| 0.24% | pascal |
| 0.21% | cobol |
| 0.13% | erlang |
| 0.11% | fortran |
| 0.10% | visual basic |
| 0.07% | lua |
| 0.05% | clojure |
| 0.02% | perl |
| 0.01% | lisp |

[vmarkovtsev]

Surprise-surprise!

[erikbern]

i'm not following. how does this prove that we need pagerank?

[vmarkovtsev]

Results from PageRank:

| 19.22% |                 java |
| 13.49% |                   go |
| 13.16% |                    c |
| 11.33% |                  c++ |
|  7.00% |               python |
|  6.33% |                   c# |
|  4.95% |                scala |
|  3.52% |                 rust |
|  2.96% |                  php |
|  2.76% |          objective c |
|  2.02% |                 ruby |
|  1.76% |                swift |
|  1.49% |                 node |
|  1.04% |               kotlin |
|  0.99% |                    r |
|  0.93% |              haskell |
|  0.90% |               matlab |
|  0.80% |               erlang |
|  0.77% |               pascal |
|  0.74% |                cobol |
|  0.69% |              fortran |
|  0.66% |              clojure |
|  0.66% |         visual basic |
|  0.65% |                  lua |
|  0.61% |                 perl |
|  0.59% |                 lisp |

I can conclude that PageRank does not help in this case. I am disappointed - this seems to be very unstable. How can Java be the dominant if the only thing I did was restricting the transition from Go to *

[ellisonbg]

I think the big issue is how the diagonal element are handled. Making minor changes to the off-diagonal element doesn't fix that.