It's been a whilst since we've had one of these on HN. What a bit of fun.
My 2c. There's a choice about the work you do. does your work satisfy a business requirement, or does it satisfy your ego.
I need a site for work for a 3 year project, CRUD, users, auth, sessions. Django's perfect for my team of 1.5. It wouldn't be responsible to roll my own, when I can have the bare bones of that site up and ready to go within a day.
This is like complaining that X company is locked in because of a chipset, and suggesting they branch out into lithography to ease that burden.
Is your core business getting things done, or is it flexing by rolling your own framework.
I did some Django but in the days, and I completely agree that as long as you have CRUD, Django shines. The issue starts when you cannot really fit your uses cases into simple CRUD and you start to fight against the framework to achieve this.
The author also made a point of negligible gains - even if you spend a week to setup auth, sessions etc. if still negligible for longer running projects
I've worked on to many long dead projects with scant doco and the best ones are always where people colored inside the lines. Remember google web toolkit? That flamed out hard 15? years ago.
I've recently had to make a mod to a decaying app written in GWT. It was easy because no one got creative. I think theres a long term risk with rolling you're own that you box yourself into a lot of corners without knowing it.
You can also get really creative with frameworks, specially if you need to quickly overcome some strange limitations imposed by framework. There is also a long term risk with using frameworks - somebody here mentioned method renames in Laravel, that makes upgrading difficult.
Im not talking about writing your own http server, or building the whole database access library. You can use some ready made libraries. As long as you keep proper boundaries between IO and business logic, you can fairly easy change the IO libraries.
Of course there is quite some overhead there, and for some projects it makes not sense.
I need a site for work for a 3 year project, CRUD, users, auth, sessions. Django's perfect for my team of 1.5. It wouldn't be responsible to roll my own, when I can have the bare bones of that site up and ready to go within a day.
This is like complaining that X company is locked in because of a chipset, and suggesting they branch out into lithography to ease that burden.
Is your core business getting things done, or is it flexing by rolling your own framework.