I as soon as had a dialogue with my boss, and we by no means got here to a conclusion as a result of he’s a fail-safe man, and I am a fail-fast man. So, who’s right? Or is that this merely a private selection?
Why I Choose Fail-fast?
On this article, I will share with you 3 the explanation why I want fail-fast offensive programming type.
1. Validate My Assumptions
Fail-fast validates my assumptions. In coding, we frequently make assumptions and our assumptions principally incorrect at first as a result of we merely do not totally perceive the system. Fail-fast / offensive programming type can rapidly validate my assumptions. Every time my code fails, it makes me a greater programmer.
2. Much less Conditional Logics
Fail-safe code normally comes with extra conditional logics. If this fails, do one other factor and if that fails once more, do one thing else. When correct clear structure is NOT being executed accurately, the fail-safe code might turn out to be very messy. Fail-fast merely removes this pointless conditional checks.
3. Pinpoint Root Trigger Quicker
As a result of fail-fast code fails the code instantly, the reported error or exception usually fairly near the precise root trigger. This reduces the debugging time considerably. When fail-safe code fails, you’ll have a tough time discovering the foundation trigger as a result of it’s not imagined to fail within the first place.
Once I Use Fail-safe?
Properly, I nonetheless use a fail-safe method solely when coping with code (e.g. exterior libraries) that isn’t inside my management. You need your system to be sturdy to deal with any errors or exceptions from this exterior libraries. One other instance is coping with exterior dependencies. For instance, a community name or file system name which is able to throw exceptions.
Ultimate Ideas
For my part, fail-safe defensive programming type is just too preferrred and never sensible as a result of you’ll be able to’t 100% assure bug free and canopy all situations. When there’s a bug, it’s tougher to seek out the foundation trigger. Have you ever ever work on the code base when one thing fails, and you haven’t any clue in any respect? It takes you a lot hours or days to determine the issue.
So my coding method is I at all times do fail-fast first to validate all of the assumptions. If the software program fails, and I am unable to repair the difficulty (because of not inside my management), I’ll convert the code to be fail-safe. Ultimately, the fail-fast method also can result in a sturdy system.
Which method do you favor?
