W3C home > Mailing lists > Public > public-html-xml@w3.org > January 2011

text/html fallback using the Same Bytes (was: Re: Use cases)

From: Henri Sivonen <hsivonen@iki.fi>
Date: Fri, 7 Jan 2011 11:48:26 +0200
Message-Id: <AA4708F2-54CA-4F36-9162-0A56B47E145D@iki.fi>
To: public-html-xml@w3.org
On Jan 6, 2011, at 18:32, Sam Ruby wrote:

> The contents of my weblog are served as application/xhtml+xml to user agents that support such, and as text/html to everybody else.  In the first case, the content is valid XHTML5.  In the latter, valid HTML5. In both cases, the response bodies are byte for byte identical.
> 
> Despite being byte for byte identical response bodies, the responses are two different things, and result in DOMs with detectable differences.

Is there a reason why Planet Venus and your blog software don't write two different byte streams for Apache httpd to choose from? That is, when the final bytes don't come from a text editor but come from some kind of automated processing pipeline anyway, why don't you make the pipeline write one serialization for application/xhtml+xml and another for text/html? As far as HTTP intermediaries are concerned, your identical byte streams are different responses anyway.

Considering how confusing and brittle polyglot is, even if you can perform the stunt of serving one byte stream as both application/xhtml+xml and text/html, is it worth encouraging (by your example and by the Polyglot publication of the HTML WG) the kids at home to try the stunt and shoot themselves in the foot?

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Friday, 7 January 2011 09:49:01 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 7 January 2011 09:49:02 GMT