Recently, I've been thinking about how it might be possible to improve the current model of how commercial software development is done today.
I was thinking along these lines:
A company has a Github repository with several tasks. Independent developers choose tasks, fork the repository and upon completion of the task send a merge request. The company would then review the committed code and upon approval pay the developer.
While this at the first glimpse seems like a pretty simple and feasible model there are obviously quite a few gotchas:
1.) How can companies make sure developers don't use their source code for their own purposes?
2.) How can a company make sure it's not wasting time on a developer? There would have to be some sort of rating system.
3.) How could you break down more complex tasks so developers don't have to spend weeks just to understand the software they're working on.
4.) Is it possible at all to accomplish non-trivial tasks on a complex software system one doesn't have at least some sort of training on?
5.) How would you make sure a company actually pays the developer upon task completion?
Quite a few problems, indeed. Nevertheless, do you think such a concept might have legs?
I do! I have been thinking about this much like you BjoernKW.. my dream is individual bid able pivotal tasks, that are part of a coherent whole agile project. When organized the right way, the fastest person to complete them on a distributed team wins, and there would be no limit of team size.
I actually ran a request on Bountify, and it was cool. There was some pushback for programmers who thought it was too much of a project for low $, while others enjoyed completing it for the challenge (and actually got paid).
Plenty of challenges (in making this really work for companies), but I believe Bountify has a good start here.
I was thinking along these lines:
A company has a Github repository with several tasks. Independent developers choose tasks, fork the repository and upon completion of the task send a merge request. The company would then review the committed code and upon approval pay the developer.
While this at the first glimpse seems like a pretty simple and feasible model there are obviously quite a few gotchas:
1.) How can companies make sure developers don't use their source code for their own purposes? 2.) How can a company make sure it's not wasting time on a developer? There would have to be some sort of rating system. 3.) How could you break down more complex tasks so developers don't have to spend weeks just to understand the software they're working on. 4.) Is it possible at all to accomplish non-trivial tasks on a complex software system one doesn't have at least some sort of training on? 5.) How would you make sure a company actually pays the developer upon task completion?
Quite a few problems, indeed. Nevertheless, do you think such a concept might have legs?