Pretty sure they were discussing technology while playing Dota.

Man fuck frameworks I can write my own shit.

@AdrianD Exactly. Your own shit. So when will you write the next ORM that blows Hibernate out of the water? ; - )

@Makenshi I actually did that :)

ORMs are often overkill for what you're trying to do. All we want is a better, more fluent communication with the database. ORMs do that, but at a very high cost. You have to map stuff, update stuff, use tons of annotations, are limited in the power of SQL, etc.

I usually did my own stuff - just some dynamic mappers, that worked almost as fast as working directly with the database. Now I don't have to do that any more, as more and more people adopt this approach, and there are some that are better than what I would usually cook up. For example Dapper.

In the past, I used to use all kinds of bloatware. Like "do it the right way, use enterprise stuff", so it was Castle, NHibernate, and tons of other huge shit. And then nothing worked and nobody knew why.
Now, I understand that simplicity is key. Dapper, ninject and full speed ahead!

@apisarenco I don't see where you did what you claim? You wrote some framework yourself, which in scope and richness of features is nowhere near something like Hibernate.

Of course you shouldn't use some framework just because it's "standard". The right tool for the job.

Netherless the argument "Man fuck frameworks" is still utterly stupid. It's the overestimation of one's own abilities that your are able to surpass the work of a whole team of actual software engineers with your homebrew solution.

The first thing that comes to mind when I read something like "In the past, I used to use all kinds of bloatware... [...] And then nothing worked and nobody knew why." is that you don't seem to know how those frameworks operate, if you have trouble telling why something doesn't work. So you cooked up your own solution where you wrote everything alone. Of course you would know how everything works there.

1/2

@apisarenco One of the qualities a good software developer imo should have is that he reuses good solutions. Why invent the wheel over and over again? Especially when your solution has nothing new to add.

In the long run you will also just need longer for your implementations as you rewrite parts of applications that where perfectly fine solved elsewhere. That will make you less competitive on the market.

2/2

@Makenshi it was faster than NHibernate, which means that it did what I would use NHibernate for, but did it faster. Because I wouldn't have used its entire feature set anyway.

You should not use a huge ass framework just because it has a ton of features. Use only what you need, and if you don't need even half of what it offers - consider getting something smaller, leaner, faster.

@Makenshi "is that you don't seem to know how those frameworks operate" - OR, it was far too bloated, and we struck a case where it just didn't work. Don't worry, there have been plenty, and not even 100 point bonuses on Stackoverflow helped solve the issues.

You totally misunderstood my comment. It was not a criticism of frameworks, fuck no. It was a criticism of loading up tons of huge ass frameworks just for a tiny microservice.

@Makenshi "One of the qualities a good software developer imo should have is that he reuses good solutions" - then why did the guys at Stackoverflow have to write their own stuff?

Judging by your logic, good software developers never innovate, and just use existing software. Which is totally absurd.

The amount of times when I made something that worked better than what the framework provided is too damn high. And I know you'll probably say that I just didn't know the framework and I'll stop you right there: In those cases I initially read the entire documentation of the said library/framework. Has it come to your mind that maybe there are instances where people come up with better solutions than what's provided in the mainstream?
It's called innovation.

If we didn't do it, we would be still writing COBOL, at best.