Re: Support Existing Content

On 2 May 2007, at 09:50, James Graham wrote:

> Gareth Hay wrote:
>
>> If the page does not conform, is not well formed, an error is  
>> displayed.
>> During the creation process of the page, the author will see this  
>> error - long before it is released into the wild - and if they so  
>> desire (as in they can't be bothered to fix it) they can change  
>> what they claim to write to "tag-soup" and the browser is free to  
>> do it's best - albeit different browsers will do different things,  
>> but on the author's head be it.
>
> This ignores the fact that many sites are dynamically generated  
> from a combination of author supplied content and external content  
> such as user supplied content and adverts. This means that there is  
> no opportunity to see the "final" page to check for syntax errors.  
> Instead one has to hope that one's publishing system is  
> sufficiently bug free to reject ill-formed content automatically.
>
> As a data point on the success of this approach, out of the 3  
> application xhtml+xml sites I visit regularly, I have personally  
> observed catastrophic XML parse errors on 2/3 and I know the third  
> did have an incident in which well-formedness was not maintained.  
> This is despite all three authors being highly competent and all  
> sites taking extensive precautions such as validating user  
> comments. One has now moved back to HTML 4 (the others make use of  
> SVG and MathML so their hands are somewhat tied).
>
> If, instead of being blogs, these sites had been, say Ebay, how  
> much business would have been lost? Why would the business ever  
> introduce the additional risk of a technology that insists on  
> enforcing well-formedness at the client end -- fundamentally the  
> wrong place to impose such a constraint since it inconveniences the  
> people least able to deal with the problem. If HTML5 were to take  
> the path of ensuring well formedness, I would expect HTML4,  
> presumably with all the same interoperability problems we have  
> today, to remain the defacto current HTML for much of the web.
>

it doesn't ignore it - these sytems will *not* overnight suddenly  
declare themselves to be sending html5 compliant code.
I don't agree that using content from differing sources is like black  
magic, and the problems that you mention originate from the poor  
error handling of previous versions of html anyway.

Gareth

Received on Wednesday, 2 May 2007 09:03:39 UTC