Doctype sniffing and the W3C Markup Validation Service

• From: Eric Bednarz <lists@bednarz.nl>
• Date: Tue, 23 May 2006 22:25:05 +0200
• To: "www-validator@w3.org" <www-validator@w3.org>
• Message-ID: <op.s90tn3nvhsy0uf@localhost.localdomain>

Hi,

the W3C Markup Validation Service front-end employs doctype sniffing at  
its worst, as is evident from the primary result message.

The odds are that this is a deliberate misfeature; please document it.   
The output is confusing to people who don't know that HTML is not supposed  
to be ISO 8879 compliant (especially if their OS ships with tools that  
are).

Two simple test-cases:

<http://validator.w3.org/check?uri=http%3A%2F%2Fbednarz.nl%2F%2B%2Fsgml%2Fdoctype%2Fw3c-validator%2Fnamecase-general;ss=1>

<http://validator.w3.org/check?uri=http%3A%2F%2Fbednarz.nl%2F%2B%2Fsgml%2Fdoctype%2Fw3c-validator%2Fpi-in-prolog;ss=1>

(apparently PIC is expected to be the XML-way)

TIA,
Eric

Received on Tuesday, 23 May 2006 20:25:28 UTC