14 July 2014
From around 1995 until 2010 there was really only one operating system for client computing: Windows.
Prior to 1995 there were a lot of options, though most not recognizable to users today: 3270 terminals, VT terminals, OS/2, Windows, DOS, CPM, and a bunch of others. Now most of these weren’t “client computing”, they were relatively dumb terminal technologies that provided access to a server (back then called a mainframe or minicomputer). Very much like today’s web browser (sans JavaScript).
Today we’ve returned to a chaotic landscape of client computing: browsers, Windows, iOS, OS X, Android, Linux (for the daring), and of course it isn’t like the pre-1995 technologies went away, they are just mostly emulated in Windows. What is interesting though, is that most of today’s client computing technologies do actually enable smart client software development. This includes the browser which can be used as a smart client technology via JavaScript development, even though the majority of browser “apps” are actually just colorful versions of 1990-era terminal-based computing where the processing is all on a server/mainframe/minicomputer/whatever-you-want-to-call-it.
What is interesting about this return to client-side chaos is that it has reopened the door for third party developer tools as a niche market.
In the early 1990’s there were quite a number of companies selling developer tools for other company’s platforms. Borland with C++, Delphi and TurboPascal, Gupta with SqlWindows, Powerbuilder, and a lot more.
When Windows became the dominant client computing platform most of these dev tools fell by the wayside (not that they went away, they just stopped being mainstream). This was because they couldn’t compete with Microsoft’s dev tools, which were always in sync with the platform in a way that was probably too expensive for third parties to match.
I think it is notable that our return to client computing chaos (or pluralism?) today has already led to numerous third party dev tool vendors that sell dev tools for other company’s platforms. Xamarin, PhoneGap, Telerik’s tools, and a lot more.
What is different to me is that in the early 1990’s I thought it was pretty obvious that Windows would become a dominant platform, and I tended to argue against using third party dev tools because I thought they’d have a rough go of it. As cool as Delphi was, I always recommended VB.
Today I’m not so sure. I don’t see any of today’s platforms becoming dominant in the foreseeable future. It is hard to imagine Windows returning to its monopoly status, but I can’t imagine iOS or Android or OS X displacing Windows as the primary corporate desktop computing environment either.
As a result we business developers need some way to build software independently of any particular platform or OS vendor, because we must assume all our business software will need to run on multiple platforms and OSes.
So today I find myself in the inverse of my early 1990’s stance, in that I’m reasonably convinced that building smart client software (at least for business) means using third party dev tools from vendors that aren’t tied to any one platform.
Of course I’ve spent the last 14 years in the .NET world, so naturally I gravitate toward a combination of Xamarin and Microsoft .NET as a way to use my C# and .NET knowledge across all platforms. I get to develop in Visual Studio on Windows where I’m most comfortable, and my resulting software runs on Android and iOS as well as on Windows Desktop, Phone, and WinRT.
As far into the future as I can see there’s no obvious platform/OS “winner”, so as a developer the question isn’t which platform to target, it is which third party dev tool reaches all platforms with a solid strategy that will stand out and thrive over the next many years.