DEV Community

Discussion on: Why you should deploy on Friday afternoon

Collapse
bizzy237 profile image
Yury

I like the idea because it allows to automate almost everything outside of coding and complicated manual tests. I don't think I'd be comfortable with extremely fast feedback loop though unless someone else writes the tests for the feature I'm adding (just because it works the way I expect it to doesn't mean that's what you were asking for)

Also, a little offtopic but I need a "dev" environment because I want an environment as close to production as possible and my work PC is not strong enough to handle a running app (including all necessary stuff like database filled with test data, proxy server, whatever) on top of a dozen of Chrome tabs, a few Firefox and IE tabs, a couple of messengers, a few projects open in IDE and a text editor

Collapse
quii profile image
Chris James Author

I don't think I'd be comfortable with extremely fast feedback loop though unless someone else writes the tests for the feature I'm adding (just because it works the way I expect it to doesn't mean that's what you were asking for)

I would argue you need to get out of this habit. You mustn't fall into the trap of not understanding what you're making and hoping someone else will know. This is why discussing what you're writing before you do it is important. Collaboration is key.

And ultimately lets say you do somehow make slightly the wrong thing. Well in most cases, it wont actually be a big deal. And here's the kicker, you can change software! Perfect is the enemy of good.

Also, a little offtopic but I need a "dev" environment because I want an environment as close to production as possible and my work PC is not strong enough to handle a running app

Perhaps a way around this is decomposing your app into smaller systems. Remember CD (and agile in general) encourages small stories so you shouldn't need to spin up the entire system in order to be confident your change will work.