Re: [Ietf-message-headers] Provisional registration of 5 X-Device-* HTTP Header fields for use in content transformation guidelines

[SM <sm at resistor.net>]

Hi Francois,
At 08:34 10-07-2009, Francois Daoust wrote:
> The situation the Mobile Web Best Practices working group is trying 
> to address is one where different vendors used different HTTP header 
> field conventions (all starting with 'X-') for the same use in their 
> proxies, requiring content providers to support the different names 
> when they wanted to do things properly. What the working group would 
> like to do is to shrink the list of existing conventions to one and 
> only one convention, but would prefer not to introduce any new 
> convention (be it the last one) for that to happen. The 
> 'X-Device-foo' format is the most commonly used format in the list 
> based on the group's experience, and thus the chosen one.

I'll encourage you to adopt Graham's suggestion of moving to 
"Device-foo" or any other valid name that does not start with 
"X-".  These "X-" headers are generally for private use where the 
interpretation is by private agreement.  There are people outside the 
group that might use that header field for other purposes and they 
are unlikely to look it up in the registry.  If your working group 
already agrees on having one convention, it is better to do away with the "X-".

> That said, I have nothing against provisional registration of both 
> forms. I add the topic to the working group's agenda for discussion. 
> I suppose that, in any case, if we register the 
> 'X-Device-User-Agent' header field, the 'Device-User-Agent' header 
> field de facto would become "unavailable", not to trigger any confusion.

You'll have to register that header field as well.

Regards,
-sm