[uf-discuss] Appeal for Issues: Empty spans in value-excerption-pattern

[Ben Ward]

Hi Rob,

On 7 Nov 2008, at 11:40, Rob Crowther wrote:

> <span class="dtstart"><span class="value" title="2008-11-04"></span>
> Barack Obama was elected the first African American president of the
> United States of American, and he was really pleased about it, on 4th
> November 2008</span>
>
> Basically I'm not clear how forcing the empty element to be the first
> thing guarantees that it will then be close to the plain text date.
> Forcing it to be anywhere would seem to guarantee there are situations
> where it can't be close to where the plain text date is if the
> original example is valid, because it's easy enough to have that
> sentence the other way around.

Requiring the machine data to be the first child element is not to  
force it to be close the human form (which is in an indeterminable  
position in the structure), but instead to force it to be close to the  
*property declaration*.

The idea is that when you read the mark-up, you read the machine-data  
value right next to the property name; a visually clearer association  
that will make it easily to see and maintain the value when working  
the code. Visual developer aids (such as CSS rules to display the  
@title attribute of elements with class=value during development) will  
cause the value to be displayed at the beginning of property blocks,  
again promoting a stronger visual association.

It means that the machine-data value doesn't get lost within a block  
of content, it's there right at the start, and that attempts to  
alleviate the violation of DRY.

I most cases, I expect patterns will be simple, and end up containing  
just the machine value and displayed form with no additional content.  
By forcing the machine-data to the front though, it remains consistent  
no matter now much content is contained in the display form.

B