Apache => Nginx
Sendmail => Postfix
Wordpress => _______
That is room for an enterprising open source developer.
The problem is that everyone needs a different 20% customized. This is what Joel Spolsky basically argues in Bloatware and the 80/20 Myth: http://joelonsoftware.com/articles/fog0000000020.html . He's talking about desktop software, but you can see the same essential thing in blogging software; I'm using Wordpress for Grant Writing Confidential at http://blog.seliger.com chiefly because it has everything I want and is easy to use.
The extra stuff I use includes themes and a couple of plugins, like one for generating an XML sitemap and another for cacheing. I know that blogging platform X probably has the particular set of features I need -- until I find something that it can't do.
I don't trust Sendmail but there's an emotional component to that; I would have laughed at you for using it in 1999, but I think 2 decades of analysis has to produce some result.
There's something more going on with Wordpress than simple complexity.
Apache doesn't have security issues at all like Wordpress, you are confusing FLEXIBILITY with INSECURE software.
Even if Wordpress didn't have security issues there'd be room for an nginx of blogging.
Need, no. Want? Yes. Literally "infinite"? No. But variable beyond belief? Yes.
Elsewhere in the thread, you complain about hardcoded URL structures! That's pretty far out there on the "I demand customizability!" front, if you consider the horrible things that will do to your platform's internals. Random users running code in their templates, another feature you request, implies that we need to sandbox and resource limit the templates, if we also want to maintain the ability to have "cheap" hosting (which breaks if our users start writing infinite loops into their templates). (Sorry, did I say "if"? I meant "when".) Take your needs, and two other customers with the same level of detailed desires, but different desires, take the union of them, and you've got a freaking complicated product spec'ed out. One that, incidentally, is going to be pretty complicated to actually customize. Oh, and the odds of it suffering the classic "endless interacting-feature bugs because nobody ever said no to a feature" is basically 100%.
I mentioned more of our features below: http://news.ycombinator.com/item?id=1328512
That's why nobody's "solved" this problem in the last ten years; you can't have programming language flexibility without programming language complexity.