The reason for it is related to our development environment, where there is a single set of qvs files used to do all of our transform, and I'm expecting that the number of people working on that transform will likely grow in the future.
In order to provide some isolation for each developer I was planning to have a specific qvs included based on who was running the load which would define their isolated development environment (this works well when running the load from the dataloadeditor.
Instead I will probably keep a single source file to define the development environment and each developer will keep their own version of that source file.
I ruled out defining the dev environment within the app, because then everytime a developer works on an app they have to change the app (which if fine for 1 or 2 apps, but with a larger number of apps it gets unwieldy to manage that and be sure you've done the right thing).
I prefer to have separate files for each developer, because then the chance that someone unintentionally checks the file into our revision control system and unintentionally impact other developers is ruled out. But then if we use the task scheduler I cannot do that.
I could possibly do it based on who owns the app - if I'm working on developing an app then it will be in my work-stream, and therefore I must own it - however there is also the possibility that from the QMC I can run an app owned by someone else (I haven't thought through the implications of that one yet).