The reason YAGNI and DTSTTCPW are gaining in popularity and work in practice is because we don't know anything about what future enhancements will be required. Future enhancements don't care about your architecture, code, or technology; they are orthagonal to all of that. They can cut right across your application's carefully maintained layers. It's quite likely that "premature generalization" will not exactly meet the eventual need anyway, which will require a pretty heft rewrite or worse (and more common), two different subsystems for the same thing living forever in your application.
I think premature generalization is much like premature optimization, but it's not nearly called out as often.
This is Rob Meyer's weblog, a weblog focused on software development and system administration based on 10 years of experience. Want to explore further? You can find out more me or see the rest of my website.
Wondering if I've written on something in particular? Try searching:
You might want to take a look at some of the more requested postings (as judged by incoming traffic):
Want more? Subscribe to this site 
or contact me at rob at big dis dot com.
See my writings on: