Proposal: Change HTTP transport header names

[Ben Poor]

Dear All,

Rather a late juncture to be proposing this, as I wish to get this done by the end of the week, but its a somewhat important question and one that has just become known to me.

The header names currently in use by the HTTP transport are taken from their equivalents used in the previous specification (v1.0.0), and use the well-known 'X-' prefix for user-defined HTTP headers.

As Paolo has pointed out elsewhere, this behaviour is now becoming deprecated, and since we have moved to containing the response fully within a JSON response it now seems largely unecessary.

I'm somewhat surprised it hasn't been spotted before, but can I propose that we remove the 'X-' prefix from all JSON header names in the RadioVIS frame?

Therefore, the following:

• X-RadioVIS-Message-ID
  
• X-RadioVIS-Destination
  
• X-RadioVIS-Link
  
• X-RadioVIS-Destination
  

Would become:

• RadioVIS-Message-ID
  
• RadioVIS-Destination
  
• RadioVIS-Link
  
• RadioVIS-Destination

Its still worth namespacing these values with the RadioVIS application.

Any thoughts on this? It would be advantageous to get this correct before ratification to avoid any unpleasantness. Any opinions from people already prototyping this transport?

Ben

[Ben Poor]

FYI - I've received messages of support for this move, so I'm also proposing we do the same for Section 6.4, Slide Sizes. 

This is where a receiver can optionally send device dimensions in the HTTP header for content negotiation on slide size. Therefore:

* X-DeviceWidth

* X-DeviceHeight

* X-DevicePPI

Would become

* DeviceWidth

* DeviceHeight

* DevicePPI

I'm rather late to the game with best practices on this topic, but we may as well get this right for the next spec. Again, any objections welcome, but I'll go ahead and amend the specification for distribution in the meantime.

Ben

[Byrion Smith]

Hi,

Yes, removing the X makes sense to me. I think it would be worth hyphenating the header names as that tends to be the norm (eg. Content-Length/User-Agent etc.).

The other thing is it might be worth replacing the word Device with Screen or Display for that tiny bit more clarity. For example:

* Display-Width

* Display-Height

* Display-PPI

Byrion.

[Ben Poor]

Sure, I'd agree with that.

Changing the header name also makes more sense - as somebody pointed out that 'Device Width' a slightly ambiguous (device != display). I'm also going through the document and rationalising how I refer to a receiver/device/client, so that fits in with that too (I picked 'receiver').

Anyone think this is a bad idea?

Ben