On Wed, 20 Apr 2005, Brad Neuberg wrote:
>
> As someone who works with web application development, here's some of
> the things that would make my life easier. Also, including certain
> methods as part of the standard that I usually have to roll on my own
> would make things more standardized:
Thanks for your input!
> * Have a document.getByPath() method that takes an XPath expression to
> traverse and find any nodes in the document; this would be _extremely_
> powerful and would erase a huge amount of boilerplate code needed for
> walking over the DOM using the standard DOM traversal methods. This
> would have to be a fast implementation though or else it couldn't be
> depended on.
http://www.w3.org/TR/DOM-Level-3-XPath/xpath.html#XPathEvaluator
I also hope to have a getElementsBySelector() method somewhere (either in
DOM3 Style or in Web Apps, whichever I happen to edit first).
> * Have methods for traversing the DOM based on the 'class' attribute. If you
> have XPath type traversal this is somewhat less needed, but I find myself
> rolling methods in most projects to work with my document by class name.
> Example methods I have rolled for myself along these lines:
> * xGetElementsByClassName(rootElement, className, tagName) - Gets all
> the elements rooted at rootElement with the given className, optionally
> restricted by tagName
http://whatwg.org/specs/web-apps/current-work/#selecting
(Can't restrict by tag name, but that seems a bit arbitrary -- why not by
attribute, or whatever? At that point you're back to getElementsBySelector
or some such.)
> * xGetSingleElementByClassName(rootElement, className, tagName) - Same
> as the above, but just returns one element
Just use getElementsByClass(...)[0]. (Since the returned NodeList is live,
it can be lazily evaulated and thus would be no less efficient.)
> * xGetParentElementByClassName(rootElement, className, tagName) -
> Navigates upwards until we hit a parent element with the given class name and
> optional tag name.
Interesting idea. Noted.
> * Formalization of innerHTML as being a standard part of creating web
> applications
http://whatwg.org/specs/web-apps/current-work/#serialization
> * Right now most people directly access an elements className property,
> without realizing that they might be clobbering multi-classed elements (i.e.
> something with class="class1 class2"). I usually have to create wrapper
> methods to ensure that this doesn't happen, such as xAddClass(target,
> className), xRemoveClass(target, className), and xHasClass(target, className),
> but it would be much nicer if the className property itself had better support
> for multi-classed elements. Some example possibilities of what this might
> look like:
> * someElement.className.add("someNewClass")
> * someElement.className.remove("SomeOldClass")
> * someElement.className.hasClass("someClass")
Agreed. Noted.
> * The lack of min-width and max-width in IE's CSS makes things difficult.
Not much we can do about this from a spec point of view.
Thanks,
--
Ian Hickson U+1047E )\._.,--....,'``. fL
http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,.
Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'