[satvikpendem]

I should say better, not best, because yes there is no best way, but there are of course better ways, eg spaghetti code vs well architected code.

> When you are less restricted you can let the business requirements drive the design of the application. With a large framework the framework determines the application’s requirements.

This only ostensibly makes sense on first glance. Business requirements are not technical, they are about what the business needs done. In that case, a framework has no bearing on what the requirements are; if I need a button and a form filled out, it can be done in vanilla JS or React, there is no difference to the end user.

And I'm not sure why anyone would compare a programming language to a library, much less one not remotely connected to each other, because the level of abstraction in architecting programs in both are very different and thus the code and architecture will not be comparable.

[austin-cheney]

The code could be spaghetti even with a framework and is not necessarily spaghetti without one.

The big challenge when talking about use of frameworks is that developers always try to make it about some extremely subjective criteria that’s only in their own self-interest like easiness (for me) or best (whatever that means). Real criteria like costs or user concerns don’t up until way down the line. Its a massive difference in perspective that many developers refuse to consider and everyone else looks at as a individual contributor capability/maturity limitation.

[satvikpendem]

> The code could be spaghetti even with a framework and is not necessarily spaghetti without one.

Sure, but some are easier to make spaghetti out of than others.

> Real criteria like costs or user concerns don’t up until way down the line.

This is orthogonal to frameworks versus not, just as you said one can make spaghetti in any sort of way, it doesn't mean that frameworks are worse at ameliorating "real costs."