Good observations about "Oslo" from David Ing
http://www.from9till2.com/PermaLink.aspx?guid=029e433d-db41-41e9-bcbd-aafe677e1606
There are lot of good observation here, but one I want to highlight…
While ‘less code’ is an admirable mission statement, the real cost of projects is the two bookend elephants of the requirements phase and the maintenance/operation phases of software, at least when judged by the money/per mistake ratio. Agile methods have tackled the requirements side by stressing the importance of working code as soon as possible, and a model can certainly fit with that ‘fail fast’ mentality probably better than just a straight API and some tests. A visual DSL could really help here and get you working with real data. Just like areas of Domain Driven Design allow you to talk in the language of your customer, a DSL/Model should be able to express the problem a lot closer to the domain a lot more interactively than just straight UML-as-Sketch (which is how I think a lot of people model today). For operation/maintenance then, again, if the system is expressed in a higher abstraction then monitoring/execution host systems have a better chance of reuse.(*)
When I wrote my Why “Oslo”? post, I really wanted to drive the simplest story about value of “Oslo” to mainstream developers. This is actually a key part of our strategy; “Oslo” must be an important and effective tool for mainstream developers. Certainly we talk about the leverage we get in a system like “Oslo” for the requirements & operations phase of a project (in fact, that is where a lot of the project was focused initially), but there is a huge pitfall in making that the centerpiece of our strategy.
Put simply (and this is a personal judgement), “modeling” efforts have failed to gain mainstream acceptance for a key reason — they are not relevant for developers — they do not make their lives easier. Today, we see modeling before the coding and modeling after the coding, but little use of modeling — save for the model-driven aspects that have organically emerged in the platform and large-scale applications — in the coding phase.
We will absolutely leverage “Oslo” for the entire lifecycle, but we have a made a very important decision to directly target developers as the key success criteria for “Oslo” v1.