Re: Validation error frequencies

Kornel Lesinski <kornel@geekhood.net> hodd gsachd:
>
> On Mon, 11 Feb 2008 16:28:38 -0000, Henri Sivonen <hsivonen@iki.fi> wrote:
>
>>> ...
>> <img style='border: 0;'> is not an improvement over <img border='0'>.


style="border:0" on numerous img elements isn't an improvement, but

   <style> /*a*/ img {border:none} </style>

is. A validator should give this hint. I bet a lot of authors would be  
glad to know this "fancy trick".

> ...

>>
>> <nobr> has been around forever and must continue to be supported by  
>>  browsers. What's the harm in making it conforming, too?
>
> I think the harm is in making HTML5 less clean and simple for authors.
> If element is conforming, it's more likely to be used and more likely
> to be taught. <nobr> and <a name> have good replacements and are
> unnecessary additions/exceptions for someone who doesn't have to deal
> with legacy code.


Yes, in the end authors will profit from a cleaner and simpler HTML.  
There are enough unavoidable dirty hacks inside.


>>> <wbr> ...

quirksmode.org isn't up to date here, I think. &#8203; breaks the line in IE7.
<http://de.selfhtml.org/html/text/zeilenumbruch.htm#erlauben>

<bugreport microsoft edgecase>
   <p style="width:2em"> xxxxxxx&#8203;yyyyyyy </p>

   If both, "xxxxxxx" and "yyyyyyy" overflow the container,
   IE7 adds an additional blank line.
</bugreport>

Note the interoperability issues with <wbr>.
<http://lists.w3.org/Archives/Public/public-html/2007Dec/0080.html>

Yes, ok, yes, <wbr> is better than &#8203; from author's view.  
Following our current Design Principles, it's a "must have". But the  
validation error freqency isn't the best argument here and elsewhere.

j.j.

Received on Tuesday, 12 February 2008 03:42:40 UTC