Ask HN: Are Branches (SVN/Git) Evil?
Their reasoning for this statement is that all code changes can be made in small increments and be done in such a way to preserve existing behaviour while allowing new development to occur. This can be achieved in a number of ways, for example if/else statements where the else branch contains the new feature but is not executed (except when the dev changes the expression to return false). By this line of reasoning, it follows that any new or dev commits which lower the quality or reliability of the code will be in some code-path which is deactivated and so they won't actually lower the quality.
In summary, my argument for branching was you can use branches to help to ensure release quality (ie: have branches for master, release and development) by segregating changes in SCM according to whether they are under active development, ready for release or currently in production. Their counter argument was that ALL changes in SCM should be production-quality and be ready to release at any second (and if not, you should change your dev process).
What do you think HN, are branches indicative of a "broken development process"?