Visual Studio LightSwitch

~ the goodbye ~

Distance on this one is professional rather than geographic. LightSwitch was an office tool, aimed at the analyst with a department to run and a database full of orders, who had no intention of ever becoming a programmer. My whole career has sat on the engineering side of that divide, and tools pitched at that user are ones I know from their documentation and nowhere else.

Microsoft released Visual Studio LightSwitch 2011 to MSDN subscribers on 26 July 2011, with general availability two days later at $299. You pointed it at a data source, described your screens, and it generated a functioning line-of-business application you could deploy to the desktop or out to Azure. The code underneath was real .NET, and you could drop into it when the templates ran out, which was the part that distinguished LightSwitch from the drag-and-drop app builders that came before it.

On 14 October 2016 Microsoft published the notice. The reasoning was that the ground had shifted since LightSwitch was conceived, naming mobile and cloud as the things that had changed, and that the company was consolidating its efforts behind PowerApps. The wording was careful and final in the way these notices are: Microsoft no longer recommended LightSwitch for new apps, and recommended that nobody start new development with it.

~ the chronicle ~

~ the Silverlight problem ~

LightSwitch 1.0 generated Silverlight clients, and that decision aged badly inside eighteen months. Silverlight needed a browser plugin at the exact moment the industry was deleting browser plugins, and it had no path onto an iPad, which by 2012 was the device the department head actually wanted the app on.

Microsoft did respond. The 2012 release added an HTML5 client, and by Visual Studio 2013 that client was the default path. By then the tool had already acquired a reputation as the thing that produced Silverlight apps, two years after the first version shipped.

~ the move to a subscription ~

PowerApps took over LightSwitch's argument from a different part of the company. The pitch is the same one: business users building their own applications over their own data without going through the development queue. What moved was where the thing lived. LightSwitch was a boxed product you installed into an IDE, and PowerApps is a subscription attached to Office, which is where its intended customer already had a login and a budget line.

That is the pattern worth taking away from this grave. The idea was never the thing that failed, and Microsoft has proved as much by shipping it twice more. What failed was a delivery model that asked a non-programmer to buy and install a developer tool first.

~ nine years, to the day ~

Visual Studio 2015 was the last release to include the LightSwitch tooling, so the tool's real expiry was tied to that IDE's support window rather than to any announcement of its own. Support for Visual Studio 2015 ended on 14 October 2025. The notice that ended development was published on 14 October 2016. Nine years apart, on the same calendar date, which is a coincidence of Microsoft's second-Tuesday lifecycle calendar rather than a plan.

~ the apps nobody ever saw ~

The applications LightSwitch produced are the least documented part of its history, because they were internal by definition. An expense tracker here, an inventory system for a single warehouse there. Software with no user outside the company that commissioned it, written by someone whose job title did not contain the word developer.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »