|Think about Loose Coupling|
Just some comments…
As a Class-DBI Con you have
Single author => single point of failure
but then, in the DBIx-Class Cons you have
Multiple contributions => less coherence expected.
What do you think is better now, more or less contributors?
Also, DBIx-Class seems to be both, the "Second worst performance" and the "Second best performance" which both are Pros and Cons in your list. Is it really both a pro and a con when it is in the middle?
No support for customizable column inflation/deflation. (is this true?)
No, it's not. There are already some column inflation extensions available on CPAN.
No distinction for "primary/essential/other/all" column kinds except for primary key and "others". (Maybe a performance issue for large columns.) (is this true?)
DBIx-Class lets you specify what fields you want to have in the resultset.
Also, what do you mean by No support for static class metadata generation.? If you mean DBIC can't create the classes for you, you haven't taken a look at DBIx-Class-Schema-Loader.
Also, your DBIx-Class Con list contains both of these:
Does NOt have seamless support for handy database domains (production, test, dev, staging, ...) and types (main, archive, reporting, ...).
Ordinary morality is for ordinary people. -- Aleister Crowley