Re: Change proposal for issue 103, was: ISSUE-103 change proposal

On Wed, 24 Mar 2010 07:00:34 +0100, Julian Reschke <>  
> On 24.03.2010 04:45, Maciej Stachowiak wrote:
>> Julian & Philip, how confident are you that the full set of characters
>> that need escaping is U+003C, U+000D, U+000A, U+0009 and U+0020? Does &
>> need to be escaped?
> I'm 99% confident. Philip already pointed out one oversight, but there  
> may be more. That's why we have WG to review this.
> An alternative point of view is: why do we introduce a new attribute  
> that is so hard to get right that we don't dare ourselves to describe  
> how?

It is only hard in XML and that is because XML itself is hard. For HTML it  
is trivial to get right.

>> Speaking in my non-chairing capacity, I think it is better to have
>> correct advice than no advice, but worse to have incorrect advice than
>> no advice. Is there anything we can do to review what characters may be
>> special in an attribute value?
>> ...
> Testing?
> How is this different from the advice for text/html? Why give advice for  
> one format but not the other?

a) The dominant use will be in HTML. b) The advice for HTMl is trivial  
compared with what you need to know for XML.

Anne van Kesteren

Received on Wednesday, 24 March 2010 15:58:03 UTC