Re: [SVGMobile12] more on data types

On Jan 12, 2006, at 12:08 AM, David Woolley wrote:

>
>> It isn't the job of a renderer to inform users of errors in the
>> content. That's the job of compliance verifying tools. UAs aren't
>
> But experience with the HTML authoring community has shown that only
> a very small number use such tools (and they tend to be those who
> actually understand the language in the first place and don't make
> mistakes - or more likely, cut and paste others' mistakes.)
>
> Internet Explorer is the only compliance verification tool used by
> most small web sites and some big ones, for HTML.
>
> One of the problems for SVG, and for HTML as commonly misinterpreted
> as a page description language, is that the results are judged on the
> purely subjective grounds of do they look right.
>
> Requests to have user agent tools reject invalid content arise from  
> the
> fact that they are the only tools used to validate most content.  To
> be effective, that has to be part of the standards.

This is reasonable in most cases, however I think specifically for  
the case of unrecognized elements, attributes and attribute values,  
mandating the silent ignore behavior is pretty good, and in some ways  
better. It's acceptable because it won't lead to harmful  
interoperability problems - if all UAs are mandated to ignore, then  
stray random gunk won't end up requiring weird quirky parsing rules.  
And it is better because it's more forward-compatible. Future  
versions of the spec will be able to add new elements, attributes and  
attribute value syntax, and fallback in downversion UAs will be  
simple and well-defined. This lets you use new features even before  
every single UA has adopted them, which seems pretty useful for web  
deployment.

Regards,
Maciej

Received on Thursday, 12 January 2006 09:00:44 UTC