- From: Jukka K. Korpela <jukka.k.korpela@kolumbus.fi>
- Date: Mon, 08 Sep 2014 13:31:30 +0300
- To: public-html-comments@w3.org
2014-09-08 13:08, Jens O. Meiert wrote: > The first line seems to sum it up quite well: “I propose an extension > to the fieldset element: a required attribute. The required attribute > is similar to the required attribute of the input element.” More exactly, the attribute would have an integer value and its value would specify the minimum number of fields that need to be filled in. The most important thing with extension proposals is usually the existence and relevance of use cases. I think for this, use cases are easy to imagine and describe. A very simple and common case is that the user is expected to provide at least one way of contacting him. Another scenario is a form where the user can select (say) up to three optional items from a list. The next question is implementability. This does not sound excessively complicated, even though it introduces yet another aspect that browsers would need to consider in form validation. I don't think the name "required" would be suitable, since an attribute with that name is already in use as a so-called boolean attribute, and an attribute with a different set of values would be confusing. On the functional side, it would be logical to have also an attribute for specifying the maximum number of fields that must be filled in. I wonder if "min" and "max" might be suitable names for such attributes. Using <fieldset> in this context is problematic, since it has default rendering that suggests close association between fields. Introducing a new element for grouping sounds awkward. Introducing the attribute for the <div> element would be somewhat odd too. But before going into such issues, a discussion about the usefulness of a feature at an abstract level is needed: should there be a possibility in HTML to specify that some minimum number of fields among some designated set of fields must have values, to pass form data validation in a browser? And should there be a corresponding way of specifying the maximum too? -- Yucca, http://www.cs.tut.fi/~jkorpela/
Received on Monday, 8 September 2014 10:31:57 UTC