W3C home > Mailing lists > Public > www-validator@w3.org > March 2007

Re: Feature request: validation by snippet

From: Stephanie Troeth <steph@unadorned.org>
Date: Thu, 22 Mar 2007 21:40:28 -0400
Message-ID: <4603300C.2030600@unadorned.org>
To: www-validator@w3.org

Hi David,

David Håsäther wrote:

> This doesn't make sense at all in my opinion. You can't just fit 
> together  some snippets and expect the document to be valid. Consider 
> for example  that you validate the following surrounded by the dummy 
> content
> 
>   <div>...</div>
> 
> It validates fine of course, so you put it somewhere in the document
> 
>   <p>
>      ...
>      <div>...</div>
>      ...
>   </p>
> 
> Oops. Not valid anymore. You're then back to square one, trying to 
> figure  out why the _whole_ document doesn't validate.
> 
> I also suspect that this will lead to questions such as "these two  
> snippets validate fine, but when I put them together it doesn't 
> validate  anymore. Is the validator broken??!?1".

The validator parses and treats information in a similar way to that of 
a software compiler. Often a simple mistake at the top of a document 
cause a cascade of errors because the parser cannot tell the 
relationship of different areas of code apart.

Template development, or markup development for CMS has a tendency to be 
modular - so one often writes code that is only a part of a document. If 
we have the means to make sure this small chunk validates (and if 
everyone does the same with other chunks), there is definitely an 
increased likelihood that there will be less errors by the time a whole 
document is generated, and it also becomes easier for developers to 
isolate the remaining errors.

What I'm trying to say is, it's not a black and white situation - it 
would just improve the conditions for which commercial developers can 
more readily generate valid code.

Think of it as a step towards divide and conquer - for generated markup 
of better quality.

regards,
-steph
Received on Friday, 23 March 2007 01:40:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 25 April 2012 12:14:24 GMT