The time taken to learn the framework was a relatively minor point of the article. I have serious doubts about the payoff—in my experience, by the time you know a framework well, you are fighting against mismatched abstractions, edge cases, performance issues, etc. You spend more time fighting the framework than developing your application. Then if you change to another framework, you have to learn everything over again. But if you use a bunch of libraries instead, you can swap one of them out without touching the others. For example, I can keep the same router but use a different ORM.

The same problems with bad abstractions happen with many applications in the long run, but it seems that using a framework will get you there more quickly. I'd say that the advantage of frameworks is developing many smaller prototypes more quickly.

Sounds like you're using the wrong framework.

May I ask what framework you have in mind with those comments?

I have to say I am very careful about how I budget and spend my learning time. I do not have much time so I make a big decision when I invest in learning a framework and I plan for it to pay off over a number of years. It would be deeply inefficient to be working at a shallow level of understanding across more than two or at most three frameworks.

I've chosen one web server to understand deeply and one front end browser based framework to understand deeply and I'm reluctant to take on anything more than that.

Sounds like you're solving different problems than I am. May I ask what kind of problems you're solving, that the same framework works for all of them? Never mind, my question is really just bait, don't answer it.

My point is that there's a certain arrogance to watch out for, when we claim that our solutions will work for other people's problems, but we don't really understand what their problems are. I've shared my negative experiences with frameworks, but I can't speak universally. I can only suggest that people not use frameworks, based on the experiences I had. I'm not about to—for example—state outright that someone is doing things wrong just because they don't make the same choices I would make. This is the kind of attitude that I watch out for when mentoring junior developers who think code is "messy" and needs to be "fixed" just because it doesn't match the way they write code.

I'm very reluctant to learn any more frameworks because tomorrow's problems seem not to match the frameworks we chose yesterday. Again, this is just my personal experience, and you use different frameworks and solve different problems.

I agree, selecting libraries is a much more fluid approach. Once you get comfortable with your libraries, you basically have your own framework without the commitment.

But with triple the maintenance cost and many times more issues, interfaces and glue.

> Of course frameworks take alot of effort to learn.

Trick is to pick the framework you'd be forced to write anyway for your project if you didn't have one. Then learning comes naturally.