As a developer I like to first create a map in ESRI's desktop software first and then re-create it using code (design and build). For example, I may want to add some rasters, TINs, featureclasses, etc. and then exclude the unnecessary features, label the features, render the layers correctly, etc., and then try to reproduce these same conditions in my code. While most of the time this is pretty straight forward, there are many occassions, especially with ArcGlobe or ArcScene, that this is very difficult, if not impossible, without having to contact ESRI and find out how to repeat what's been done at the desktop level. It's like shooting in the dark. On many occassions it's just a matter of knowing what property to set on some interface or what method to execute. While it may be just one thing that's missing the difference between knowing and not knowing can mean success or failure.
What ESRI needs to do is develop a tool for us developers that once you've built your map or scene you then open a tool from the the desktop software and it tells you all of the objects and interfaces that were used to build it. In fact, it would be nice if it completely generated the code for you in the language you need or at a minimum generated the pseudo-code and a list of all the objects, interfaces, properties and methods. This would be tremendously helpful and not only save ESRI users a lot of time, it would be a great teaching aid, reduce the guess work, and generally speaking provide ESRI developers with a better user experience.
What's your take on this?