4 ms·
A Few Valid Reasons to Reject a Bug Fix
- DrScump 11y agowhy isn't the #1 reason to avoid introducing regressions, i.e. changing a behavior that (even if it runs contrary to spec or documentation) existing users/applications rely on, knowingly or unknowingly? In such cases, fixes shouldn't be rejected per se but instead be deferred until the full impact on existing code bases can be estimated. Such a fix should be deferred to the next major release AND the user base notified that the change in behavior is coming (and/ ideally, give affected users a way to check their apps for reliance on that soon-to-end behavior and/or provide a backward-compatibility switch to retain the old behavior long enough to get affected apps converted). Few things alienate your existing user base more than having long-working applications suddenly broken by "fixes" whose implications were not warned of in (at least) release notes.