Escalante's Future

Escalante started out as a way to bring in the capabilities and functionality within a managed environment like JBoss Application Server 7 or Wildfly to Scala developers. Among the existing frameworks, Lift has tried to take the most advantage of this combination, so it was an easy place to start at.

Based on our experience of pitching Escalante to Scala developers, it has become clear that managed environments are considered too heavyweight, particularly when it comes to ease of development, and hence most Scala developers building concurrent applications are trying to stay away from them choosing instead to build Akka or Play applications. In other words, Scala developers are focused on building solid, statically typed, applications with the least amount of resistance.

Around the same time as we were working on Escalante 0.3.0 a couple of interesting things happened:

  • First, the Reactive Manifesto came out (July 2013), pretty much laying the foundations for the type of applications that Scala developers have been trying to build.

  • Secondly, Vert.x founder and lead developer, Tim Fox joined Red Hat (January 2013). Vert.x is an polyglot, event-driven application development framework for the Java Virtual Machine which promotes development of reactive applications.

Right until mid-2013, Vert.x's Scala language extension was lagging behind other languages such as Java or Groovy. Tim was very keen to get this effort moving and I started contributing to Vert.x's Scala module, eventually becoming the lead of it.

The amount of feedback received on Vert.x Scala language extension and the community involvement in the project has surpassed anything Escalante had, demonstrating even more where the focus is for Scala developers. They want to easily build multi-node reactive applications that are scalable and fault tolerant. Managed containers, in spite of providing a lot of base middleware and come with good monitoring tools, are rooted in the single node vision and their ease of development is lagging behind.

So, with all this in mind, we have come to the conclusion to not develop Escalante any further. For those Lift users that have tried Escalante, there are other containers out there which should fit your use case.

Thanks to everyone who has been involved in Escalante!