Context: Emmanuelle posted a [message to the mailing list on the evaluation of template vs generated code https://lists.w3.org/Archives/Public/public-wai-ert/2015Feb/0010.html] I tried to post a response but it got somehow lost in the cyberwebz, I am posting it here as an issue.
For convenience, an extract from Emmanuelle's message follows, and my response/issue below:
I think in that paragraph, in addition, it is important to emphasize on what
should review the tool and alert to the possibility that, if not done, is
being generated a false positive or a false negative.
I explain it with a specific case that I found:
AChecker review tool is integrated into the LMS ATutor, which enables
content authors review the accessibility of content they create using the
WYSIWYG editor, usually the TinyMCE.
The tool checks the original markup of content created with TinyMCE, instead
of the generated markup. For example, aChecker not find any fault with a
content created with specific formating code used in ATutor [1] as:
I think Emmanuelle has raised a relevant point. I just started thinking as well of related issues after we coincidentally revised the "document fragment" feature in the latest conf call. I would suggest adding something to the list of content formats, as this formatting code Emmanuelle presented is not plain HTML anymore, it is TinyMCE markup.
I don't think this is covered by our current definition of "markup formats", which are oriented to HTML/XML-ish heavy markup languages. Think instead of Markdown different flavors, MediaWiki's wikitext, BBCode, or any other of those lightweight markup languages for user-generated contents that are pervasively employed to generate HTML. Tools could (and some attempts have been made) evaluate the accessibility of contents in those formats.
On the other side, what about template languages such as those of AngularJS, Jade, React, HTMLBars, etc.? Many developers of web applications currently create their web contents using those languages. Tools could evaluate the accessibility of those templates in their native form (I am not taking about PHP, JSP, VBScript or Ruby, which are maybe too complex in practice, but only about templating languages; besides, I am not say that any test must be conclusive —neither are they when evaluating HTML contents). Are they a different kind of content, different from markup resources? Otherwise, should they be mentioned together with the markup resources?
Context: Emmanuelle posted a [message to the mailing list on the evaluation of template vs generated code https://lists.w3.org/Archives/Public/public-wai-ert/2015Feb/0010.html] I tried to post a response but it got somehow lost in the cyberwebz, I am posting it here as an issue.
For convenience, an extract from Emmanuelle's message follows, and my response/issue below:
I think Emmanuelle has raised a relevant point. I just started thinking as well of related issues after we coincidentally revised the "document fragment" feature in the latest conf call. I would suggest adding something to the list of content formats, as this formatting code Emmanuelle presented is not plain HTML anymore, it is TinyMCE markup.
I don't think this is covered by our current definition of "markup formats", which are oriented to HTML/XML-ish heavy markup languages. Think instead of Markdown different flavors, MediaWiki's wikitext, BBCode, or any other of those lightweight markup languages for user-generated contents that are pervasively employed to generate HTML. Tools could (and some attempts have been made) evaluate the accessibility of contents in those formats.
On the other side, what about template languages such as those of AngularJS, Jade, React, HTMLBars, etc.? Many developers of web applications currently create their web contents using those languages. Tools could evaluate the accessibility of those templates in their native form (I am not taking about PHP, JSP, VBScript or Ruby, which are maybe too complex in practice, but only about templating languages; besides, I am not say that any test must be conclusive —neither are they when evaluating HTML contents). Are they a different kind of content, different from markup resources? Otherwise, should they be mentioned together with the markup resources?