Re: Things in HTML that I disagree with (Was: evidence of harm)

On Thu, 25 Jun 2009, Rob Sayre wrote:
> On 6/25/09 3:13 PM, Ian Hickson wrote:
> > Aesthetic reasons, different priorities (e.g. whether migration to and 
> > from XML is desireable), opinions based on circumstancial evidence 
> > that that don't have as much evidence as other things (e.g. guesses as 
> > to the psychological effect on authors of including certain features), 
> > etc. I always defer to the argument that is based on the strongest 
> > reasoning or research; sometimes, that just doesn't happen to match 
> > what I would like, because I can't argue my case as convincingly as 
> > other cases.
> 
> These are things you don't like. You put up with them because people 
> have made (in your opinion) cogent arguments for their inclusion. Things 
> you disagree with don't make it in the spec. I suppose it's possible to 
> play word games with the definition of "disagreement", but the spec text 
> provides a concrete delimiter.

I disagree that we should allow "/" talismans or "xmlns" talismans or 
"xml:lang" talismans in text/html content. I don't just dislike the idea, 
I actually disagree with the fundamental concept that there is value in 
making it easy for authors to transition to and from XML. I think that 
allowing this will further confuse authors about the difference between 
HTML and XHTML and will lead to more "tag soup" and will actually harm 
the ability for authors to truly use XML effectively.

However, I cannot back up this argument with any kind of data.

Meanwhile, the "pro-/>" advocates can demonstrate with data that authors 
are using this technique today in text/html and that making this non- 
conforming is harming authors who are trying to use validators, as they 
cannot see the woods for the trees when they have this error everywhere 
in the validation results, crowding out other more important errors.

I don't just dislike having "/>", I actively disagree with the premise 
that we are helping authors by letting this slide.

>From an objective point of view, however, I can see that my opinion here 
simply doesn't have the support from the data that the pro-"/>" opinion 
has. Therefore, I don't continue to argue my case on this topic and the 
spec allows "/>".

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

Received on Thursday, 25 June 2009 23:15:44 UTC