3 ms·
A rewrite is almost never the right choice.
by voz_ 3y ago
A rewrite is almost never the right choice.
- pphysch 3y agoI much prefer a SWE culture that embraces iterative development, where "rewrites" and migrations are not feared and can be done safely and efficiently because care was taken during the design process to identify ontologies and system boundaries. Yes, the MVP takes a bit longer as a result. If you don't iterate, and you push what is essentially v0.1.0 to production, you reinforce a culture where deleting code is sin and the concept of iteration ("rewriting") is foreign. http://www.laputan.org/mud/ http://www.laputan.org/mud/
- makapuf 3y agoI think of it as: a whole rewrite is never the answer so don't start one, a gradual one always is, so plan for it.
- kevincox 3y agoIf you have well defined components it is often reasonable to rewrite them. Rewriting an entire project is likely a mistake. I think the mistake is generally the size. If you lock up half of the team and have them rewrite for 2 years then you are likely going to have time. If you slice off components that can be rewritten by a few team members in a couple of weeks you can likely get a good quality result.
- voz_ 3y agoNo lol, not on anything worth working on. Maybe on something small or with no real users.