- From: Marie Hanotte <m.hanotte@acti.fr>
- Date: Mon, 23 Nov 2015 09:31:43 +0100
- To: "Michael [tm] Smith" <mike@w3.org>
- Cc: www-validator@w3.org
- Message-ID: <5652CEEF.8030101@acti.fr>
*Hello Michael,* Thank you very much for your answer ! I now understand better how it works. You're also right about the "search" role that should be on the form (not on the input) : I misread the specifications about accessibility. Thanks again, and have a nice day ! *Marie Hanotte * <http://www.acti.fr> *marie hanotte* intégratrice multimedia // /interactive designer/ m.hanotte@acti.fr <mailto:m.hanotte@acti.fr> tél. : +33 4 37 37 25 10 acti.fr <http://www.acti.fr/> acti.fr/blog <http://www.acti.fr/blog> Experte Accessiweb <http://www.acti.fr/blog/une-introduction-a-laccessibilite-du-web/> Renforcer la stratégie d'influence d'une marque <http://www.acti.fr/success-story/rcf/> Le 23/11/2015 06:50, Michael[tm] Smith a écrit : > Marie Hanotte <m.hanotte@acti.fr>, 2015-11-20 11:47 +0100: >> Archived-At: <http://www.w3.org/mid/564EFA3F.6020701@acti.fr> >> >> When I use the validator on a HTML5 page I'm working on, I get this error : >> *Bad value “search” for attribute “type” on element “input”.* >> >> I'm a bit surprised, as the reference lists *input type=search* as a valid type : >> http://www.w3.org/TR/html-markup/input.search.html#input.search >> >> Is there something I am missing here ? I don't know if there is something on >> my side that is not fitting, or if it's a bug in the validator. >> >> I'm sorry I can't provide a link to my page, as it's not yet publicly > ... >> <form action="/content/search" method="get"> >> <label for="Search">Votre recherche</label> >> <input type="search" name="SearchText" id="Search" role="search" /> > Remove role="search" and that error message should disappear. > > The error message being emitted for this case is suboptimal but > unfortunately given the current design of the checker backend, it’s > difficult to improve it. > > Basically what’s is saying is, “An input element with role=search can’t > have type=search.” But what we rightly should have it saying instead is “An > input element with role=type can’t have role=search.” > > Anyway, as far as I understand ARIA, there is never any need to mark up a > type=search input element with a role attribute. Accessibility tools > already natively/automatically recognize for what it is—a type=search input > element has “implicit semantics” equivalent to either role=textbox or > role=combox, depending on whether it has a “list” attribute or not. > > So there’s no need to put role=search on a type=search input element. In > fact, as far as ARIA 1.0 goes, type=search input elements should never have > any role attribute at all. I think role=search is intended for marking up > an entire search form—e.g., at the level of the <form> element. > > It looks like ARIA 1.1 adds role=searchbox to mark up the equivalent of a > type=search input element. But I’m not sure that’s supported yet in > accessibility tools—and I think even once it is, you wouldn’t need to used > it explicitly, because I expect the intention is that browsers would > automatically implicitly map type=search input elements to that role. > > —Mike >
Received on Monday, 23 November 2015 08:32:15 UTC