Its you who are being biased. Really complex web apps need as much JS perf, as they can get out of the browser!
And I mean apps! Not just comment systems, etc.
This lead us to the conclusion that we need complex benchmarks, not only measuring many single features, but also complex algorithms which are on the “egde” of what is possible, since web is not set in stone. Its expanding, and more and more things can be handled in web!
For example editing photography, can be too much for current JS!
And do not forget that web is for everyone poor, or rich. Stationary or mobile. Inter or ARM or whatever ….
V8 is biased toward Chrome, Kraken was biased toward FF. MS do not have JS test of its own (pity!!).
So if you put them together you can measure how much progress is been made!
Last but not least Sun Spieder now have too small resolution. Even large advancement is translated into small improvement of score.