- From: Daniel Brooks <db48x@yahoo.com>
- Date: Sat, 29 Jan 2005 19:41:57 -0500
I was just reading through the spec and noted a few things that caught my eye. I took some notes, which I'm sending to you with only slight elaboration. 2.4.2: for controls with date and time types that don't have timzone information, would it be better if the zero point where 1970-01-01T00:00:00.0 (the epoch in the local time zone)? 6.2: why specify that image controls are treated as two controls with specific names if they're just going to be skipped? (step 4, paragraphs 3 and 4). The same thing is specified for form submission, because that's where it matters. 8.1: :indeterminate ? what good is it? If it's defined in a different standard and you're clarifying how it interacts with this one, perhaps a note to that effect? The rest are just editorial in nature. 2.15: 'form script' => 'for scripts' 5.6: grammar? 'In the tables below, cells that specify that the case should be handled as another represent cases where the given attribute combinations are otherwise meaningless. In such cases, the encoded data set is created from the fallback encoding type, not the specified encoding type.' This just feels awkward. Had to read it several times to figure out what it means. 5.6.1: combine put and post columns as in previous table 5.6.4: 'Thus the replace attribute is effectively ignored when enctype is a mailto URI.' should be '?when action is a mailto: URI.' 6.1: mentions type attribute on select element, but then states that all attributes on select element are ignored. Probably just needs to be reworded. 7.1: '?much like happens for the default action of change and input events.' should be '?much like what happens?' Also, do you know of anyone with plans to implement this in Mozilla? db48x
Received on Saturday, 29 January 2005 16:41:57 UTC