[avmich]

I've read all, and got both interesting points - useful to me angles of view - and certain statements and conclusions from then which I don't agree with.

1) Author says object and data structure are different things basing on stated difference, but doesn't explain why this couldn't be a different point of view on the same thing.

2) When author says "adding a class is easy when you need a new entity, while adding a function is easy when you deal with data structures" he doesn't mention that only "packaging" of code - into switch cases or separate methods - is different. To add Triangle, one still has to add both fields which contain Triangle data and all the methods which work with that data - area, perimeter, center. Similarly, adding data structure "Triangle" also needs all those additions - and only them - but in a different form.

And many derivations from this perceived difference are, um, questionable.

Yet I like author's explanation of impedance mismatch. True, data storage has to satisfy many needs (roughly, we don't want to copy data - normalization), while logic may focus on specific goal at hand.