How to not Support Windows Platforms

This analysis presented by Ed Bott on 2013-01-01 reminds me of something that has been nagging me for the past couple of years.

It is amazing to me that web-site and application developers manage to be in complete denial about the distribution of operating-system and browser usage by those they proclaim to want as users and customers.  Here’s what I see as the bottom line.

• Internet Explorer has 58.36% of browser activity, exceeding the total of all other browsers combined.
• The combination of Windows 7 and Windows XP represent the most successful desktop operating-system deployments ever, with 78% of current usage. 
• Windows 8+8.1 usage, at 9.3% exceeds that of all non-Windows operating systems combined, and it is early days for Windows 8.x.
• The most-heavily used non-Windows desktop operating system is Mac OS X.

How is it that developers of web sites and of desktop applications are able to ignore this reality?  It seems to be ideological and the echo-effect of believers.  If you don’t support Internet Explorer, of course the visitor statistics will demonstrate a different population of access, distorting (and confirming desired) reality.  If your wares appeal first to a technologically-oriented population, of course you might see statistically-significant differences in your usage and supporter population.

The question is, who are you attempting to reach and how do you want to serve them?

I am reminded of this question every time someone at Coursera or Codecademy hand-waves the justification for Internet Explorer being discouraged for their sites. 

Is that self-selection consistent with an undertaking’s expressed goals?  As far as I can tell, that rationale simply excuses developers from testing against Internet Explorer.  It’s not fact-based.  It is convenient folklore.

Although I am a mere nano-scale open-source developer, I can tell you how I intend to target and develop for desktop platforms and browsers.  Deliver on Windows first.  Target the latest version first.  Confine platform dependencies to a portably-adaptable set that brings older versions of Windows and other platforms into practical reach.  Ensure that open-source/free tools can do the job, allowing others to contribute platform adaptations. 

It’s simple, really.

The accommodation of mobile operating systems and consoles is a different avenue that is also appealing to nanoISVs such as myself.  In that case, choosing an ecosystem might be an useful factor in determining how to broaden and possibly scale-out a hit.  It depends on the application.  For gaming, social software, and personal clouds, it will be interesting to see how this approach may also work as an avenue of entry for software offerings.