DogFood, take 2

I succeeded in modifying my html5 template in a way that *almost* 
appeases the html5.validator.nu conformance checker.  There is only one 
issue left: the proper placement of svg elements in xhtml5.

It might be worth reviewing the changes I made (had to make?) as these 
may surface use cases that the specification may wish to cater to (or not).

First, I removed the <meta charset> as previously discussed.  As I have 
the charset in my HTTP headers, this is no biggie.

Second, I added what appears to me to be an entirely unnecessary <div> 
element around the search form <input> elements, and then added "body > 
header form div { display: inline }" to my CSS to compensate.  Frankly, 
this harshes my mellow.

Third, I removed the size attribute from the input field and added this 
information back to the CSS style sheet to compensate.  Like with the 
first change, this made me feel a bit like I'm living in a nanny state, 
but other than that, no problem.

Fourth, in my article headers, I surrounded the time element with a div 
tag.  Sigh.

Fifth, some of my articles summaries contain only strictly inline 
elements and text.  Suffice it to say that this is common in weblogging. 
   I found that I had to add a <div> element in such cases.  My guess is 
that few content management systems would bother, and will simply always 
emit a <div> element in this case.

Sixth, even within the div, I have some content that is 'injected': a 
floating image and a visual indication that the content is an excerpt (I 
use a simple ..., others may indicate "more", or refer to content "after 
the fold").  This content needs to be tailored based on whether the 
content it abuts is block or strictly inline.  In the case of demanding 
that my floating image be wrapped in a span vs a div based on the 
content model of the remainder is somewhat counterproductive.  First, 
the image is intending to float in any case.  Second, and more 
importantly, I doubt that many will do this but the content will work 
interoperably anyway, contributing to the continuation of the culture of 
"works in my browser" trumping "conforms to the specification".

Seventh, I did find the issue in my content where a multi-paragraph list 
item did not properly enclose all the paragraphs, i.e., some paragraphs 
spilled out to into the void between the list items.  As it was well 
formed and displayed "correctly" in Firefox, Safari, and Opera, I would 
have been unaware of this if it weren't for the conformance checker.

Eighth (and finally), as my article footers consist solely of an anchor 
tag, I found that I had to add a immediately vestigial <div> element. 
Again, this blunts the benefits of the semantic markup as one needs to 
defend why <footer><div>...</div></footer> is somehow more semantic than 
the now familiar <div class="footer">...</div> approach.

I do think all this needs to be rethought if there is a desire for 
people to adopt html5 and produce conformant content.

- Sam Ruby

P.S.  I realize that there isn't eight separate issues, but sometimes it 
is worth looking at the implications of a single issue from multiple 
points of view to reduce the possibility to settle on a change that ony 
turns out to be a partial solution to the problem.

Received on Thursday, 6 December 2007 20:10:59 UTC