- From: Laurens Holst <lholst@students.cs.uu.nl>
- Date: Wed, 16 Jan 2008 20:06:44 +0100
- To: "Hallvord R. M. Steen" <hallvord@opera.com>, "public-html@w3.org" <public-html@w3.org>
- Message-ID: <478E55C4.2090907@students.cs.uu.nl>
Hallvord R. M. Steen schreef: > Hi, > I've seen these new attributes cause problems for real-life websites. > I think we shold discuss re-naming them to make conflicts with > existing content less likely: Just for the sake of clarity: are you talking about XML attributes or DOM properties here? > INPUT action > - breaks web applications frequently. Both GMail and Yahoo mail (the > new Oddpost-based version) use input/button.action and were seriously > broken by WF2's action attribute. > I propose renaming it "actionurl". And still there are people who claim that namespaces and XHTML are useless :). > SELECT data > FORM data > > These break > a) scripts that set them, read them back and are confused because they > turned into resolved URLs I think changing XML attribute values when they’re set is very bad practice (and impossible to implement properly in a standard DOM). It might in some instances be necessary for ‘compatibility with the web’ (aka IE), but IMO should be avoided otherwise. I still remember that at some point Opera used to resolve all src and href attributes (on <link>, <a>, <script> elements, etc.). That was one of the Opera compatibility issues our product had back then (Backbase), because it was not really possible to retrieve the original values. Only the DOM properties corresponding to the attributes should resolve the URLs. I looked at the specification, and I could not find where it specifies that the data attribute values have to be replaced with their resolved versions. If someone could point me there, I would appreciate that, and I would like this to be changed. Otherwise, it seems that the browser that you tested (which?) is wrong and should correct its behaviour. However, I did see that the DOM ‘data’ property does not explicitly state that it should be a resolved absolute value [1]. I think this should happen, similar to similar DOM properties such as a HTMLFormElement.action [2] (note: does not explicitly state it either, but browser do implement it this way), HTMLAnchorElement.href [3] (does explicitly state it must return an absolute uri), etc.. Can I get an acknowledgement that these comments are noted by the editor? > b) scripts that define global functions called data and try to call > them with a SELECT or FORM element in scope (for example from event > handlers) > > I propose renaming it "dataurl" But all in all, I’d say that the cases you are citing contain very sloppy code, and deserve to break. It seems obvious that adding such obvious custom attributes without prepending a prefix of some kind is a compatibility problem just waiting to happen. I am not in favour of renaming these attributes and properties for the sake of those sites. ~Grauw [1] http://www.whatwg.org/specs/web-forms/current-work/#data1 [2] http://www.w3.org/TR/DOM-Level-2-HTML/html.html#ID-74049184 [3] http://www.w3.org/TR/DOM-Level-2-HTML/html.html#ID-88517319 -- Ushiko-san! Kimi wa doushite, Ushiko-san nan da!! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Laurens Holst, student, university of Utrecht, the Netherlands. Website: www.grauw.nl. Backbase employee; www.backbase.com.
Received on Wednesday, 16 January 2008 19:07:07 UTC