T-Shirts for Fans of Obvious Mathematical Statements
“5 > 2”. So now you know.
“5 > 2”. So now you know.
The conformance checker technology preview has been stuck in a time warp for a while, because I have been writing about it instead of working on it.
I am happy to report that the schema layer has now undergone a major update. (The Java parts have the same bugs as before. I have not forgotten about those. They just are not what this update is about.)
The WHATTF have today decided to reveal our plans for HTML6. The objective is to rewrite HTML5 based off the new ISO standard OOXML specification as the serialization. The architectural model will adhere to the principle of separation of semantics from presentation using RDF for semantics and XSL-FO for presentation. An XML Schema will be provided for semantic validation.
Advantages:
We firmly believe that new HTML version should maintain backwards compatibility and specification writers must not reinvent the wheel. We've come to the conclusion that the OOXML specification adheres to these goals, making life easier for everyone: users, developers, implementors and spec writers
More details will follow soon.
html5lib 0.9 is now available for your parsing pleasure.
html5lib is an implementation of the WHATWG HTML parsing algorithm in Python and released under a MIT-license It enables malformed HTML to be parsed into standard minidom and ElementTree structures,in a way that is highly compatible with the behavior of major desktop web browsers. As well as parsing to trees html5lib contains a DOM to SAX converter; it is hoped that by supporting these standard APIs, toolchains based on draconian XML parsers can be repurposed to process HTML content with minimal effort.
In addition to the HTML parsing capability, html5lib 0.9 contains an experimental liberal XML parser based on the WHATWG algorithm without the HTML-specific error handling. This is suitable for parsing XML from sources that cannot guarantee wellformedness; e.g. web feeds.
The 0.9 release is expected to be the last major release before 1.0 and no new features will be added before 1.0 is released. Instead we will work on any remaining correctness issues, other bugs, and on improving the messages reported when parse errors are encountered. Bug reports are very much appreciated. Users or people looking to get involved are encouraged to join the mailing list or visit the #WHATWG channel on freenode.net
this post refers to the "Write" interface from WordPress utilized to post comments to WHATWG blogs.
there is well-intentioned, but mis-implemented markup in the edit form; namely, improper implemetation of the FIELDSET
.
for a proper FIELDSET
, one needs to do 4 things:
FIELDSET
(which this form does)LEGEND
for the FIELDSET
(which this form does NOT); the natural candidates for LEGEND
are the level 3 headers (H3
) classed dbx-handle so instead of repeatedly hearing "click to open this box", i would also get the pseudo-box (which i would call sub-forms) LEGEND
as an indicator of what i am about to open or close. i would also make the alt text device independent - instead of "click here to open this box", i would propose "show sub-form" and "hide sub-form"FORM
controls to their textual labels by use of the LABEL
element and the for/id
mechanism that ties the form control (which takes the "id") to a LABEL
(which takes the "for") or multiple labels; the LABEL
should contain the actual, textual label, and NOT the FORM
control, as in this form; this form has the attribute set set correctly to bind the LABEL
to the FORM
control, but since the LABEL
element is opened PRIOR to the INPUT
element, no labeling is available to the user - in my case (i use a screen-reader) the sub-forms that appear when one opens a FIELDSET
to reveal a FORM
appear unlabeled to my screenreader, because of invalid markup.FIELDSET
(which this form does)