[wfjackson]

Not just that but it's abusing statistics.

StatCounter attempts to measure browsing volume, not people using a particular browser.

Net Applications attempts to measure people using a browser, so IE9's reported share is three times higher, around 9%.

As an analogy, take the toothpaste market with only two players. Lets say 30% of people use Colgate and 70% use Crest. But for some reason the first group uses more toothpaste, so 60% of toothpaste sold in the market is Colgate and 40% is Crest.

Now, which one has a higher "marketshare"? Crest or Colgate?

It's funny how everyone gets so confused about this.

Anyway, Net Applications is the more apt comparison here because she's attempting to compare people having a disability, not the amount of browsing done by such people.

[bunderbunder]

> Now, which one has a higher "marketshare"? Crest or Colgate?

Trick question.

I can't answer that question until I've decided whether I want to be a Crest fan or a Colgate fan. Sort of like how I can't decide whether to measure smartphone market size by units shipped or dollars spent without first deciding whether I want to promote Android or iOS.

[ntietz]

I think that in this case, it's fair to compare the number of people who have a disability instead of the amount of browsing they do, because it speaks to the potential market.

Suppose that there was no technology to help blind people use computers. Then blind people would have 0% of browsing, but they're not 0% of the population -- should we create technology to help them even though they do no browsing? (Yes.)

[lilsunnybee]

Regardless of if IE9 use is 3% or 9%, the point still stands. It's still an at least higher or comparable number of internet users that are ignored and marginalized by the workflows and decision-making of most internet app projects.