title attribute and abbreviated class names (Was: [uf-discuss]Currency Quickpoll: Preliminary results)

[Scott Reynen]

On Oct 14, 2006, at 3:42 AM, Mike Schinkel wrote:

>>> I think your use of the title attribute in these examples  
>>> contains two
> bad practices....
>
> Hmm.  I see your point, and being new to this I'm learning from your
> examples.
>
> OTOH, I also see that the proposals I first viewed as being very  
> complex and
> I'd fear many people simply won't implement them until there is a  
> direct
> benefit, and there will likely be few direct benefits until lots of  
> people
> start implementing them; a classic chick and egg problem.  Is there  
> not a
> way to significantly reduce complexity, at least in the 80  
> percentile case
> and still maintain proper semantics?  I know I'm new and might be  
> schooled
> to understand the downside of my current view, but currentky if I  
> had to
> between the two, I'd vote for semantics that don't fit perfectly over
> significantly greater required complexity per each marked up amount.

We should minimize complexity, but not at the expense of clear useful  
semantics.  Without clear useful semantics, there's no point in  
microformats.  Your examples seem to leave a lot of ambiguity about  
what things mean, and this reduces the benefit of use, which will  
hurt adoption.  Small businesses don't want to get a bunch of  
payments submitted in the wrong currency because some parser guessed  
wrong.  A microformat should leave no room for guessing.

>>> It's a minor problem, but it's also a minor solution - typing  
>>> four extra
> letters.
>
> Point of note, my concern wasn't typing extra letters, it was the  
> need to
> transmit extra bites over the wire.

This is also a good goal, but also a lower priority than clarity.   
Microformats are going to require some amount of additional markup.   
There's no getting around that.  Reducing this markup by making the  
class names more ambiguous isn't worth it.

> Imaging a very large volume site that
> has hundreds of prices to mark up per page, and they server  
> millions of
> pages an hour. It might add up to be a concern.

Do you know someone who actually has this problem?

> For example, why does
> google use "q" instead of "query" on it's search box?  I'm assuming to
> reduce unnecessary characters.

Take a look at the HTML source of any Google page.  It's full of  
comments that don't provide any functionality at all, and inline CSS  
and JavaScript, which could be cached separate from the HTML to  
significantly reduce bandwidth.  I see no evidence unnecessary  
characters is a real concern for Google.  And I don't see any other  
high-volume sites doing this kind of micro-optimizing for bandwidth.

Peace,
Scott