32
xonya
1y

We have a huge codebase, built during the last 10 years, with a lot of problems caused by legacy dependencies. We are trying to modernize this gradually but it is very challenging because we have a lot of features to maintain and test coverage is low.

Today, a guy hired three days ago just proposed to rebuilt everything from scratch stating that he did the same thing in his personal project, so it wouldn't take too much to develop what we need.

Manager gently invited him having a quick call. I would pay to listen that conversation.

Comments
  • 6
    Personal Projects != 10 year old systems that are in active development.

    How many years does the new guy have? Backing up his statement with personal projects does sound like he is a junior
  • 4
    Oh yeah I did that too when I was 3 years old and didn't like the shit from before, so I sat down and shat another pile.^^
  • 4
    I know that type I always tell ok make a branch and do it.
    Sometimes it takes a day, sometimes a week but everyone resigns or tell that he need more time that we don’t have.
    They never apologize and I just do my shit and look at those failures of “smart” knowing everything young people hit by real life.
  • 2
    same thing at my work, the problem is that even refactoring some parts is hardly approved by the project manager cause of course we don't have time for that, and after a month he asks why this part and why that part is laggy and buggy..
  • 1
    @gitpush It seems he has already some years of experience. I think it's an ego problem.
    A huge one :D
  • 0
    @vane Yeah, that sounds the right approach :D
  • 5
    Rewriting such huge LEGACY apps from SCRATCH is a piece of crap idea and only leads to more bugs and headache.

    What you need is a slow one-step-at-a-time rewrite module by module. It is like making multiple unit-level commits on your git repo. If something breaks, because of better traceability there is an easier way to revert to the last known working state
  • 4
    @xonya when he’s ego has crashed and burned into a small pile of ashes, send him my way, I have a few legacy projects that are in dyer need of rewrites 😂

    Although the shock and awe approach generally end up being a clusterfuck of hell larger then the mess we started with.

    The best approach is slow and steady, As changing 1 small feature can result with a flow on affect leading to other features needing adjustments to cater for your original changes.
Add Comment