Though this one isn't far off: https://bugs.php.net/bug.php?id=18556.
You can point out all the flaws, but that's like pointing out the flaws in McDonald's. Sure it's cheap food, but it's consistently cheap and there's a much larger market for cheap food than gourmet food.
Perhaps "other language" developers should start thinking of themselves as gourmet chefs. And like gourmet chefs, not many people can afford their products in the real world.
People have sentimental feelings about McDonald's, prefer the style of fries, etc. - it isn't that McDonald's products are incomparable or irreplaceable and that the only alternative is expensive, difficult "gourmet" food.
How exactly? I always see vague assertions that PHP is somehow better for "stuff I do in the real world", but nothing concrete. Guess what, the rest of the world is real too. I am not writing haskell code for my imaginary friend, I am writing it for real users to use in the real world, and a real business relies on it to provide real people with real incomes to feed their real families. How exactly would PHP be "more efficient all around"?
Like those, who failed for years to actually consult the results of the test suite before shipping a new release? http://gcov.php.net/viewer.php?version=PHP_5_4
Like those who change stuff incompatibly to the worse, while introducing security holes? http://php.net/manual/en/function.is-a.php https://bugs.php.net/bug.php?id=55475
Like those were security fixes create even larger holes? http://thexploit.com/secdev/critical-php-remote-vulnerabilit...
Like those who can't program a sane line of code in the language PHP is written in? "OMG, this calculation overflows, let's replace the int by an float!"
You can't make that shit up. I have nothing against stupid people, but against stupid people who fight against every kind of process who could prevent their own stupidity like
- "learning from mistakes",
- "learning the language they actually program in",
- "handling bug reports in a way it doesn't alienate reporters" or
- "having a decent test coverage".
Instead of thinking about yet-another-documentation site, what about fixing the devastating quality of their API documentation (especially the comments)?