->removeDecorator(‘label’)
->removeDecorator(‘HtmlTag’);

… stacked and it works, for hidden and hash elements.

I don’t care if it does not validate.

Could you confirm this, or did you prepare your elements in a special way (changing the decorators etc).

But, of course, one can take the best of both worlds and do it like:

public function render(Zend_View_Interface $view = null) {
    foreach($this->getElements() as $element) {
        if($element instanceof Zend_Form_Element_Hidden) {
            foreach($element->getDecorators() as $decorator) {
                $decorator->setOption('class', 'hidden');
            }
        }
    }
 
    return parent::render($view);
}

This will yield valid XHTML code and reside in the parent form. It’s not a matter of right or wrong, but rather of personal opinions on design patterns.

Moreover, that implementation skimp lines of code and new classes creation with the same results and no different business logic nor logic error.

Btw, this is my interpretation.

And, as always, there are more than one ways to get things done and people adhere to different philosophies. I’ve used the view helper approach because this is a presentation issue (how the form will be displayed) and presentation logic should stay at the “view level” of the MVC paradigm.

I created a mother class (that extends Zend_From) for all my forms, and rewritten render() function:

public function render(Zend_View_Interface $view = null)
{
	foreach ($this->getElements() as $element) {
		if ($element instanceof Zend_Form_Element_Hidden) {
			$element->setDecorators(array('ViewHelper'));
		}
	}
 
	return parent::render($view);
}

So no different code is written in forms or views.
You can use even Tudor solution with my implementation.

P.S.: Tudor, make possible and easily visibile for all how to comment code plz.