[uf-discuss] changing abbr-design-pattern to title-design-pattern?

[James Craig]

Tantek Çelik wrote:

> To be frank - the blog post on hAccessibility WaSP was seriously  
> flawed.
> 1. It used a strawman example to argue against.

What about our example was a straw man? Just yesterday it was  
mentioned that Yahoo uses dates without dashes and wikevent was given  
as an example of using the "slightly better dates with dashes." Let's  
use wikevent's "in the wild" example (that includes timezones) and  
talk about what happens with the date portion of this ISO string:  
"2007-05-07T20:00:00+01:00."

I don't have Jaws in front of me, but the time is either going to be  
read as "twenty o'clock zero zero plus one o'clock" or as "twenty  
zero zero zero zero plus one o'clock." Both are nearly useless to  
human ears.

> 2. It recommended known unworkable solutions (using object? are you  
> kidding
> me? that's already been tried and failed - did you not do your  
> homework? see
> my original abbr post, and include-pattern-feedback).  In addition  
> I told
> James Craig *in person* about this at SXSW, so I was a bit  
> surprised it
> still made it to the blog post.

As Andy pointed out, the point of the article was not the proposed  
solutions, but I want to point out that your reason for being hung up  
on the object example is because it was "tried and failed" due to UA  
implementation bugs. The argument you're making here completely  
contradicts the argument you make later in this same email here  
(quoted, but out of order):

> OTOH, not allowing bugs and stubbornness of implementers to retard/ 
> slow/stop
> progress and nor taking a step backward and using span instead.

[...]

> However, I'm against contorting microformats because of bugs or  
> suboptimal
> behaviors in <1% marketshare browsers.

I don't really consider screen readers as "browsers." People who use  
<1% market share browsers have a choice to change or upgrade. The  
people who use screen readers really have no other way to access  
online content. Yes, they could turn off the title attribute  
verbosity, but this would then cause ambiguity of understanding  
other, valid uses of abbr.

I doubt you would agree with the following statement:

"I'm against contorting building code regulations to require  
wheelchairs ramps and elevators in public buildings because of the  
<1% of citizens with mobility impairments."

> So I'm for adding "-" and ":" to get a better and even *usable*  
> result in
> screen readers,

I agree with you that the date portion (yyyy-mm-dd) with dashes,  
though sub-optimal, is better. I told you this in our discussion at  
SXSW. I also immediately mentioned that's only the case with dates,  
not datetimes. The complete ISO timecode is gibberish with or without  
punctuation; I completely deny your claim that it's "usable."

> I think there needs to be a balance.

I agree. I know we all have the specifications' best interests in  
mind, and I'm glad it's finally in full discussion.

James