Re: Bug 7034

On Fri, 26 Mar 2010, Henri Sivonen wrote:
> 
> I do agree with the notion that markup becomes more maintainable (and 
> potentially more compact) if it doesn't use features like <h1 
> align=right>, though. But then, markup also becomes more maintainable if 
> it doesn't use style="...".

This was my position for a while, and the HTML5 spec in fact disallowed 
the use of style="" for a long time. However, it was suggested -- and 
somewhat convincing arguments were presented to support this -- that the 
style="" attribute can actually aid in the maintenance of documents when 
used for the twin purposes of rapid prototyping and stylistic exception 
handling.


> I think it would be useful for validators to offer a mode for flagging 
> <h1 align='right'>, style='...' and other cruft like 
> xmlns="http://www.w3.org/1999/xhtml", language=JavaScript and 
> type=text/css, but doing it all by default makes validators unusable for 
> those migrating existing content.

Validators are able to do this today. Nothing in the HTML5 spec prevents a 
validator from prioritising certain errors over others -- indeed HTML5 for 
a while explicitly required certain errors to be given a lower priority. I 
think it would be perfectly reasonable for a validator to have a mode 
(even the default mode) wherein it lumped all presentational markup (that 
which is defined by the spec's obsolete section as being better handled by 
CSS) into one error in the UI, marking the document as "valid HTML except 
for the use of presentational markup" or some such.

This is exactly the kind of feature validators should be competing on.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'

Received on Friday, 26 March 2010 17:04:02 UTC