It’s just data

Feed Validator Message Improvements

Zeyad Rajabi: At a minimum, our goal will be to always validate as XHTML 1.0 Transitional compliant code. For a basic blog we will validate as XHTML 1.0 Strict compliant code. For those blogs that use features where we cannot output Word supported CSS, our aim is to be XHTML 1.0 transitional compliant.

It certainly looks like there will be a lot more use of the style attribute in feeds in the future.  The new WikiPedia feeds depend heavily on the use of style to convey exactly what changed.

Aggregators that blindly allow style attributes through may be creating a security risk.  Those that blindly strip style attributes won’t allow their users to take full advantage of those WikiPedia feeds.  Those that attempt to parse the style attributes will find that the full grammar is quite extensive.

As the Feed Validator is now increasingly looking inside of content for feed related issues, it is becoming more important that the issue found be more precisely identified.  Where it previously would simply point to the end of a description and say “there’s a relative URI reference in here”, it now will tell you exactly what URI appears to be incomplete.

Here are some examples from popular feeds: MobileCrunch, Ray Ozzie, O’Reilly Radar, Think Secret.

Fixes: atyle -> style && hiehgt - > height

line 94, column 158: content:encoded should not contain atyle attribute [help]

  ... h?i=JbV3m2" border="0"></img></a>


line 221, column 158: content:encoded should not contain hiehgt attribute [help]

  ... h?i=fvzGmA" border="0"></img></a>


Posted by Stephen Duncan Jr at

Fixes: atyle -> style && hiehgt - > height

Additionally, float; right; => float: right;.  And the img tag — which starts at line 1, column 4 of the content:encoded — is missing a closing quote on the class attribute.  Identifying which tag contains the error is a marked improvment.

Some or all these issues may affect how these entries show up in feed readers.  Not to mention the issues with the two RSS extensions, the first of which affects Dave Winer’s own NewsRiver.

Posted by Sam Ruby at

I could easily be wrong, but I think Stephen may have assumed that it was a typo in the FeedValidator, but yeah, the HTML itself is where the typo is.

Posted by Bob Aman at

Yep, the wrong line numbers & my laziness in actually investigating led me to post incorrectly.  My apologies.  I appreciate Sam giving me the benefit of the doubt though.  :)

Posted by Stephen Duncan Jr at

Add your comment