XHTML Validation and Accessibility

Question

We've recently completed developing some Application Express systems for a customer, and they were disappointed the pages did not pass W3C XHTML validation to meet their accessibility targets.
I've had the chance to do a bit of a post-mortem and nailed some issues that were down to our templates. I've also noted some previous threads here (notably Re: Validation!!! and Re: HTML DB and Accessibility (AA-Compliance) and look forward to resolution of markup validation issues in future releases. In the meantime, here are the most frequent problems we encountered. I'd appreciate any explanations, and any suggestions on how to improve matters in the short term.
INVALID ENTITIES
The majority of the validation errors are like:
Cannot generate system identifier for general entity "p_fname"
General entity "p_fname" not defined and no default entity
Reference not terminated by REFC delimiter
Reference to external entity in attribute value
Reference to entity "p_fname" for which no system identifier could be generated
Entity was defined here
Where the App Ex engine is generating internal URLs and not escaping ampersands, that are then interpreted by the validator as forming part of an entity. This isn't disastrous, but an early fix in a future release would be welcome as it would reduce the number of errors and make it easier to spot any arising from invalid template code.
Am I right in thinking that storing CSS, images and scripts externally to App Ex would eliminate this?
STANDARD HTML BUTTONS
All "HTML Buttons" result in:
Value of attribute "type" cannot be "BUTTON"; must be one of "text", "password", "checkbox", "radio", "submit", "reset", "file", "hidden", "image", "button"                    
There is no attribute "onClick"                         
End tag for "input" omitted, but OMITTAG NO was specified
Where the engine is generating an attribute value in uppercase and attribute names in camel-hump case instead of all lowercase. We can work round this by creating an "HTML Button" template with valid attributes and values.
ITEM LABEL
Reference to non-existent ID "P1_DATASET_CATEGORIES"
Where the label is the overall label for a checkbox item with multiple checkboxes, all of which have their own valid label. Should the engine include the checkbox item as a hidden item? Can we work round this by doing that manually, e.g. in Post Element Text?
HEADERS
Reference to non-existent ID "GV_VALUE"
Where a report has report has Headings Type "None" and hence no <th>s for headers attributes to refer to. Is there a workaround for this?
REPORT BREAK FORMATTING
End tag for "br" omitted, but OMITTAG NO was specified
<br> tags inserted by break processing within the engine are not terminated " />". This is not controllable by any template (that I am aware of) - anything that can be done?
TEXT FIELDS
This is rather puzzling, as it seems to be random: items on some pages are affected, those on others are not. Anyone have an explanation?
There is no attribute "TYPE"
There is no attribute "NAME"
There is no attribute "SIZE"
There is no attribute "MAXLENGTH"
There is no attribute "VALUE"
There is no attribute "id"
Element "INPUT" undefined
End tag for "INPUT" omitted, but OMITTAG NO was specified
Reference to non-existent ID "P5_ENTRIES_DISPLAYED"
Where the engine is generating an element and attribute names in uppercase instead of all lowercase. From earlier threads, I'm presuming nothing can be done at present.
HYPERLINK
There is no attribute "HREF"
Element "A" undefined
As above: element and attribute names in uppercase instead of all lowercase.
TABLE HEADING/DATA ALIGNMENT
Value of attribute "align" cannot be "LEFT"; must be one of "left", "center", "right", "justify", "char"
Attribute value generated in uppercase. Any workaround?
REPORT SORT ARROW
End tag for "img" omitted, but OMITTAG NO was specified
Sort arrow <img> element not terminated " />". Any workaround? (Or can this be suppressed completely/replaced?)

Answer 1

Hello,
Most of your issues are things we are aware of and hope to get fixed in 3.0 as they are produced from an engine specifically using htf package to render elements which we will be changing. Unfortunatly with most of those issues I see there I don't think you can do much.
I doubt the html the engines produce will pass xhtml strict in 3.0 but we will pass xhtml transitional (other than maybe urls in links) with a custom DTD that takes into account things we output in our namespace.
Carl