Ben Gillis is on the right track wrt some of the benefits we believe Oslo could help provide to IT in the long run. Specifically, his focus on the costs involved in getting answers to questions is spot on:
"Many applications are so large and complex even the most knowledgeable working on them can’t answer the above questions beyond generalities. And, general answers aren’t good enough for a lot of development and support scenarios. ... As for labor-skills strategies, almost always it requires not only an expert with the technologies themselves, but also the application, with all its various nuances and such. A guru with the implemented technologies is still greatly restricted in terms of productivity until he/she knows the application. Even the gurus can’t answer questions like those above without labor-intensive research, and it’s manual research that’s error-prone. The consequences can be drastic if something is overlooked such as a call, something in the metadata or mentally trying to work out how code executes as you read through it. And, there’s “walking knowledge” risk; lose a guru or two and the venture will feel the pain. ... The time it takes to get answers is greatly reduced. The skills it takes to get answers is greatly reduced. The headcount needed to support an application can be reduced, sometimes drastically. The accuracy is much greater than manual analysis. I’ve built a couple of these the hard way and experienced all these results. Having this out-of-the-box would be a big productivity booster."
"Many applications are so large and complex even the most knowledgeable working on them can’t answer the above questions beyond generalities. And, general answers aren’t good enough for a lot of development and support scenarios.
...
As for labor-skills strategies, almost always it requires not only an expert with the technologies themselves, but also the application, with all its various nuances and such. A guru with the implemented technologies is still greatly restricted in terms of productivity until he/she knows the application. Even the gurus can’t answer questions like those above without labor-intensive research, and it’s manual research that’s error-prone. The consequences can be drastic if something is overlooked such as a call, something in the metadata or mentally trying to work out how code executes as you read through it. And, there’s “walking knowledge” risk; lose a guru or two and the venture will feel the pain.
The time it takes to get answers is greatly reduced. The skills it takes to get answers is greatly reduced. The headcount needed to support an application can be reduced, sometimes drastically. The accuracy is much greater than manual analysis. I’ve built a couple of these the hard way and experienced all these results. Having this out-of-the-box would be a big productivity booster."
Getting answers to questions is all about your ability to access and analyze data and the ease and speed with which you can access and analyze data is directly impacted by how you capture, relate, and store that data in first place. Today, application (meta)data is strewn across a wild west of distant, isolated towns with fractured infrastructure, poor communication, and little to no law and order. This is true both on the Microsoft platform and across the industry at large. It's time for something better. At a minimum we need to clean up our own house so Microsoft customers can get better answers to their questions about their applications built using Microsoft technology and hopefully there's an approach in here that is amenable to the broader industry so customers can get answers to questions that span implementation platforms.