Thanks to Liberty I had the opportunity to attend a morning session by Scott Ambler around Disciplined Agile Delivery.
The selling point around DAD seems to be that it offers a more complete view of both project and organisational delivery than what Scrum does. The argument being that Scrum gives you a framework within which you can execute project delivery through sprints but doesn't itself address what happens before (where did the product backlog come from) or after (how are we actually releasing the product to production). So Scrum teams start adopting other practises to roll their own, more complete, process. Of course this sort of approach should really be done by "Process Experts" and as most people are not "Process Experts" the resulting process is typically sub-optimised. DAD suggests that rather than starting at the baseline and optimising to the middle, why not just start at the middle?
This is somewhat analogous to the old libraries vs frameworks debate in the technical world. Frameworks make decisions for you and as such are great when venturing into a new technology space but as you become an "expert" in that area the helpfulness of frameworks diminishes and they can become rather restrictive. At this point "experts" begin to craft their own solutions typically utilising a small number of libraries. I like to consider my self well versed in development of solutions and as such I typically avoid big frameworks and favour smaller, more lean bespoke development.
In this sense DAD is a framework, it gives you everything, and in comparison Scrum could be considered a library, giving you a small set of tools and as such I'm inclined to tend toward rolling my own process utilising various methodologies and processes. This has worked for my projects.
But here's the rub - While I may be a "technical expert" (tending toward the use of libraries over frameworks) I am by no means a "Process Expert". My current experiences have been with smaller scale projects and a bespoke process will fit well at that scale without much expert knowledge. We do, however, need to move beyond that; toward proper organisational change and without "Process Experts" the approach of mashing some principles and methodologies together will result in a sub-optimised process tending toward failure.
The process aspect of large scale software delivery is my "new technology space" and as such it may be prudent to investigate a large focused framework until such times as we can adapt and move beyond its limitations. Will it be DAD? Will it be the Scaled Agile Framework? Perhaps it will be like the carefully governed but organically grown Spotify style model. I don't know yet, but I do know its time to start moving toward using agile in the large and I doubt a slapdash approach will be enough until I developed a broader understanding into larger organisational challenges.