Re: Semicolon after entities

Philip & Le Khanh wrote:

> for all browsers to handle something that /alleges/ to be HTML5
> consistently if the document is defective (poorly formed).  Indeed,
> if browsers /do/ vary wildly in their treatment of ill-formed
> HTML5 documents, there will be far greater pressure on /hoi polloi/

Unfortunately, that is not how the market really works.  What really 
happens is that one vendor becomes, at least temporarily, dominant. 
Users then start complaining to the other vendors that their browsers 
are broken because they don't handle input correctly when the author 
thinks it is valid, even though it is really invalid.  Vendors are then 
forced to reverse engineer the dominant product, and imitate its 
behaviour.  If you actually reject a construct, users will consider it a 
false positive.

The result has been that, if you look at www-style or www-svg, it is 
clear that all the vendors want there to be no input that is either 
rejected or produces undefined results.

"HTML5" would have value to user agent developers if all it did was to 
document real life HTML.  However, this knowledge is, at best, not 
helpful to authors and, sometimes, dangerous knowledge.  Moreover, at 
the same time, it is extending HTML to compete with ASP.NET and XFORMS 
and with Flash and SVG.

Received on Saturday, 28 April 2007 13:56:41 UTC