At What Point do We Cut off Support for a Browser?

I know, I know, in theory we should be developing sites that work in every browser and not just targeting specific applications. However, the reality isn’t quite there yet. Support for the different web standards varies massively from vendor to vendor.

Usually we make the decision on how much effort we put into making a site work in a particular browser down to the visitor statistics of that site. If your site only receives a handful of visits from a certain browser, why spend hours – or even days – trying to work around its faults?

So my question is this: just how low should the numbers be before a particular browser gets ‘cut off’? Take, for example, the top 5 browsers in Pixel Meadow’s Mint logs:

  1. Firefox (54%)
  2. Safari (31%)
  3. Internet Explorer (8%)
  4. NetNewsWire, Camino & Opera (2%)
  5. Shrook (1%)

From those numbers it’s clear that I need to support Firefox and Safari (which by extension means support for NetNewsWire and Camino), but what of IE, Opera and Shrook? Do I go out of my way to make sure any future revisions of Pixel Meadow work fully in these browsers, or do I just make sure they’ll degrade gracefully if need be?

Of course, this is assuming an existing site… It stands to reason that a new site with no clear visitor demographics should target as wide as possible until their visitor statistics are known.

Advertisements