Brooks' Law states that adding people to an already late project will make the project later. I think there's a similar law that I pose here for the first time (to my knowledge, please correct me in the comments if I'm wrong):
Adding developers beyond one to the bug-fixing effort on an app will increase the number of bugs.
I find this to be very appealing and very close to Brooks' Law, especially when working on a Death March project. I believe this applies even if the developers that are added are incredibly good developers. This doesn't require a badly-written app. This happens regardless of the quality of the app.
I do not have a real proof for the law, which, as a mathematician, annoys me to no end. It is one of those things that to me and my experience, it seems really intuitive. Basically, no developer you add will know all of the requirements and previous/current bugs that need to be fixed/implemented, so adding them increases the possibility of them adding a new bug. With each new person, that probability increases exponentially because they may now step on everyone else in the group's toes/bugs in some way. At the same time, when there are multiple stakeholders (which is almost always true), it's guaranteed that nothing is ever going to get fixed properly and completely.
What do you think? Agree/disagree? Let me know in the comments.
Subscribe to:
Post Comments (Atom)
I agree to a point. Whoever's running the project should never put a new dev on a bug that requires extensive knowledge of the business domain. But There are often little bugs that don't require any knowledge. For example, I was thrown in during the last week of Adecco to knock out a bunch of SharePoint + CSS + JQuery bugs. I killed them all, and to this day I have no idea of even what all that site did.
ReplyDeleteHeh, fair enough. However, how do you know you didn't cause a bunch of other bugs (if it was the last week of the project, after all)?
ReplyDelete