The article tries very hard to pretend there are problems that don't really exist, and does that by going through great lengths to ignore any possible solution.
Take for example the blogger's thesis that OO doesn't help swap out data structure choice at a later date. The blogger goes off on a tangent about procedural programming, and somehow fools himself into believing that modularity was the defining trait, and that only major benefit OO brings is that "some modules should not depend on other modules." The blogger then proceeds to make the bold and clueless claim that "when we’re just considering code within a module, or a couple of related modules, OO does essentially nothing to help make data structure changes less painful." Bold claim indeed.
In the meantime, here am I looking at the standard library of a random programming language and I see collection interfaces implemented by generic linked lists, queues, stacks, arrays, vectors, maps, etc. In turn I see concrete implementations of these containers further implement more specialized interfaces for their class of container, both in terms of concrete data structures and data structure traits such as whether it's ordered or not, offers concurrency assurances, if the container is const or read-only, etc.
All of this is implemented with polymorphism, inheritance, encapsulation, abstraction, etc.
But somehow the blogger refers to setters and getters.
It's a poorly written blog post. Not worth a click.
mcc1ane•2mo ago
(2019)
f_klem•2mo ago
I just clicked... indeed, the article is so misinformed that even some simple half-trues (whatever that might be) mentioned in the article are completely out of place...
But this kind of articles show me a different problem: the relation between seniority and expertise. I'm growing tired of an industry where more years of 'experience' is conflated with more expertise. Maybe one of the culprits is not OO itself (which I personally do not regard as a bad paradigm), but the myriad of really bad and lousy texts and examples teaching and explaining OO concepts: I think there is out there a complete generation of developers that just don't design good software.
locknitpicker•2mo ago
Take for example the blogger's thesis that OO doesn't help swap out data structure choice at a later date. The blogger goes off on a tangent about procedural programming, and somehow fools himself into believing that modularity was the defining trait, and that only major benefit OO brings is that "some modules should not depend on other modules." The blogger then proceeds to make the bold and clueless claim that "when we’re just considering code within a module, or a couple of related modules, OO does essentially nothing to help make data structure changes less painful." Bold claim indeed.
In the meantime, here am I looking at the standard library of a random programming language and I see collection interfaces implemented by generic linked lists, queues, stacks, arrays, vectors, maps, etc. In turn I see concrete implementations of these containers further implement more specialized interfaces for their class of container, both in terms of concrete data structures and data structure traits such as whether it's ordered or not, offers concurrency assurances, if the container is const or read-only, etc.
All of this is implemented with polymorphism, inheritance, encapsulation, abstraction, etc.
But somehow the blogger refers to setters and getters.
It's a poorly written blog post. Not worth a click.