Re:Startup Engineers and Our Mistakes with MongoDB
by meesterdude on Wednesday July 19, 2017 @01:30PM
(#14806075)
A really great writeup. A revealing exploration of the perils of mongoDB, and of the greater issue of selecting technology for a project.
There is never any mention of usecase, in companies. Everyone is still trying to be google or facebook, and uses "their tools" without a use case for why. It's just because that's what everyone else does.
All technology has a place somewhere for some use - but we need to know why we are choosing them.
For example, I choose Ruby on Rails because of it's gains in productivity, reliability, security and pleasure of use. I can quickly sling maintainable code that can scale up to millions of users. And most importantly, I don't have to make a lot of decisions. Best practices are defaulted into the framework. Some people shy away from frameworks, favoring "simpler" tools - but that only means they have to re-solve all the problems the framework has already solved.
Likewise, with MongoDB, all those useful things SQL databases do for us need to be resolved. But, that's a database decision, and many people making the decision are just developers in search of a data store. Transactions and ACID are often not even on their radar.
Overall, i think this is a symptom of a lower barrier for entry. Where before you needed a CS degree to be a programmer, now you just need to go to a bootcamp. People are not properly trained, aware (or interested) in the underpinnings like databases, networking or security - and to their own peril and the peril of the companies they work for.
There is never any mention of usecase, in companies. Everyone is still trying to be google or facebook, and uses "their tools" without a use case for why. It's just because that's what everyone else does.
All technology has a place somewhere for some use - but we need to know why we are choosing them.
For example, I choose Ruby on Rails because of it's gains in productivity, reliability, security and pleasure of use. I can quickly sling maintainable code that can scale up to millions of users. And most importantly, I don't have to make a lot of decisions. Best practices are defaulted into the framework. Some people shy away from frameworks, favoring "simpler" tools - but that only means they have to re-solve all the problems the framework has already solved.
Likewise, with MongoDB, all those useful things SQL databases do for us need to be resolved. But, that's a database decision, and many people making the decision are just developers in search of a data store. Transactions and ACID are often not even on their radar.
Overall, i think this is a symptom of a lower barrier for entry. Where before you needed a CS degree to be a programmer, now you just need to go to a bootcamp. People are not properly trained, aware (or interested) in the underpinnings like databases, networking or security - and to their own peril and the peril of the companies they work for.