| Homepage | https://www.spickermann.com/ |
| https://www.linkedin.com/in/spickermann/ | |
| StackOverflow | https://stackoverflow.com/users/2483313/spickermann |
| Homepage | https://www.spickermann.com/ |
| https://www.linkedin.com/in/spickermann/ | |
| StackOverflow | https://stackoverflow.com/users/2483313/spickermann |
I read a couple of message in which people were surprised that Twitter is still online with only 25% of their employees left. Some even argue that those employees must have been redundant because the site is still up.
I once was the last developer of a small startup and was let go when it ran out of money.
My experience: It is possible that web apps run for years without any maintenance. Some feature might stop working, but it can take years before an application finally dies.
I worked on translating legacy authorization rules into Pundit policies this week.
Pundit is such a great gem. And when you add verbose specs to test them, it is a piece of art.
Even complex permissions rules are easy to read and understand – not only for developers, but also for product managers.
In every startup I worked in the past, the same question came up: Where to host the application?
My answer is always the same: Build a monolith and host it on a PaaS (Heroku, Render, etc.) until you have enough people to maintain your own servers in the cloud.
Do not overthink this. Once you outgrow PaaS, the switch is likely not that hard. We just migrated our whole infrastructure in just one week (planning + execution). With <15 min downtime.