Benjamin posts an interesting response to Joel Spolsky’s interview on Microsoft-watch. Basically Joel says he’s not interested in Indigo since it is just a “big communications architecture” that makes it easier to build connected systems. Here are my thoughts in response to some of Benjamin’s questions:

There is already a business imperative that applications communicate. The method they communicate by, however, varies greatly, and the majority of them don’t do it “the SOA way!” The easiest way for applications to be connected is to share the database. Of course, this breaks all the rules in bypassing the security, business rules, etc. that are contained in all the application layers above the data. It works but it’s fragile. And this is where Indigo will come in. It will be easier for me to do the right thing (a la the Pragmatic Programmers).

I like the goals of Indigo. I believe that frameworks like Indigo should handle the underlying complexity the majority of the time. Every additional layer of abstraction, although dangerous if it is leaky, improves productivity. That way I can worry about autonomous services and not network protocols (for example). At a previous employer the introduction of ASP.NET and the huge gains in productivity over classic ASP allowed for fewer developers to do much more with a better end product. When the users notice this, then you know that the benefits are real. I can imagine the same thing for Indigo.

Steve Maine also has a good post, calling Indigo the Web Services Kernel.