I'm busy writing a paper for the OBZ resources page, explaining how to go about building an app with licensing & protection in mind.

One of the areas I touched on was that a non-free app that requires deployment has zero value until it can be deployed, which involves not only licensing and protection, but installers (or whatever framework you're using, e.g. click-once for .NET or Capistrano for Rails). My thoughts so far on the matter:

If you're building software that must be deployed, you can have all the features you want, but without an installer they are completely inaccessible to the customer and thus valueless. If the project were to be wound up tomorrow, all of the effort put into those features would be wasted.

On projects with a deployment requirement, we encourage customers to prioritise basic installer functionality very highly. Often the product of the first iteration is an application skeleton and an installer for same. This also allows for automated deployment testing at a very early stage, which helps identify platform issues.

Would be really interested to hear feedback on this ...