Totally agree about the amount of coordination overhead. That is a huge amount of time to do anything.
But even so, it’s even slower, by a lot, once you pull the ripcord and need to keep the site working while you update it.
Prior to release you don’t need to have branch and release then QA then deploy. You can just modify schemas and drop existing user data without needing to migrate anything. You can change the look of the interface without angering users who generally hate change.
Just the cycle of releasing new features carefully is a ton of overhead.
I’ve spent entire days just rolling out code to change which domain name is used to refer to some images because doing it quickly would overwhelm the image servers due to the caches being unpopulated. 100% of that would be unnecessary prior to going live.
Totally agree about the amount of coordination overhead. That is a huge amount of time to do anything.
But even so, it’s even slower, by a lot, once you pull the ripcord and need to keep the site working while you update it.
Prior to release you don’t need to have branch and release then QA then deploy. You can just modify schemas and drop existing user data without needing to migrate anything. You can change the look of the interface without angering users who generally hate change.
Just the cycle of releasing new features carefully is a ton of overhead.
I’ve spent entire days just rolling out code to change which domain name is used to refer to some images because doing it quickly would overwhelm the image servers due to the caches being unpopulated. 100% of that would be unnecessary prior to going live.