Hacker News Reader

Comment by Mathnerd314 | Hacker News Reader

> exponential trend

They're market shares & hence bounded by 0 and 1, so exponential seems pretty unrealistic. The logistic curve is a better starting point: https://en.wikipedia.org/wiki/Logistic_function#In_economics...

View on HN

This is what I came here to post, and you can see that, in the mobile chart, Chrome and Safari already seem to be levelling off. Maybe a better model would be like an ADSR (attack decay sustain release) envelope in sound synthesis where the logistic curve represents the attack / initial adoption phase, possibly followed by a mirror logistic curve representing mature saturation then eventual replacement by competitors. That said, some products do come back from the dead, like Mozilla itself; the Apple Mac is another example.

Saying "Chrome won" now feels like saying "IE won" in 2002 or so. Look at the chart here:

https://en.wikipedia.org/wiki/Internet_Explorer_5

2004 was a dark time...

I couldn't find a single chart from invention of the web browser to present, but these two cover most of it: https://en.wikipedia.org/wiki/File:Browser_Wars_(en).svg http://i.imgur.com/kVocSvo.png

What strikes me is the diversity. Chrome has mostly been stealing market share from IE; Safari is growing too and is by no means dead; FF has been declining, but not that much. There are 4-ish strong browsers in the market now when before there was only 1 (IE) and a half (Mozilla). Certainly a different picture from the chart in the article.

+ the fact that the sum of shares shouldn't be over 1

To me on the Desktop only graph it looks like Chrome is meant to be a logarithmic growth curve not an exponential one. It seems he wanted the data to fit the projections rather than the projections fit the data.

Reply on news.ycombinator.com