Currently there is a proposal on the table to move MSBuild out of the .NET redistributable and ship it instead as part of a lightweight developer tools redist (subset of the .NET SDK). This would mean that your dependency on MSBuild has the requirement of installing an additional redistributable before MSBuild can run.

Before we make any decisions of this magnitude, we need to better understand how this decision would impact you and your decision to adopt MSBuild. We would like to assess a few things:

  • What key requirements would a tools redist have to meet in order to be viable for your needs, e.g. how small, distribution channels, available to repackage/refactor for your needs, etc.?
  • Assuming a redist met the requirements, there would be an additional redist step in your distribution story. Does that threaten or completely break any of your key scenarios or value propositions? Which ones?
  • What kinds of innovations do you expect in future versions of MSBuild? Are you willing to accept a dependency on a future refresh of the redist in order to utilize those innovations?

I would love to hear your feedback! If e-mail is not a convenient medium to give us that feedback, let me know and I'll gladly schedule a quick 1:1 phone meeting.

Alex Kipman (