RE: Comments on Section 6.1 (Persistent Connections) of HTTPbis Part 1, version 17

Hi Dan,

You wrote:
>So I've also been reviewing the -17 specs, and was going to
>suggest that 6.1.1 is no longer relevant (except for the "SHOULD
>implement" part at the end) and should just be removed.

I tend to agree with the following provisos:
1. The history of why persistent connections were introduced is valuable to newcomers to HTTP, and it would be nice to retain this information somewhere. Perhaps Appendix A is the right place for it.
2. It would also be helpful to provide a succinct definition of persistent connections. Apart from the table of contents, the term appears twice (firstly in section 3.4 Handling Incomplete Messages and then in Section 5.1.1 (Chunking)) before section 6.1, but it is not defined, and there is no forward reference to section 6.1 in section 3.4 (or 5.1.1) for a definition. The current 6.1 provides a feel for what they are, but a more succinct definition would be useful. Perhaps this definition should appear in section 2 (Architecture), due to it being an important part of the rationale behind version 1.1. This would then mean that the term was defined before it is first used in section 3.4.

Jonathan

-----Original Message-----
From: Dan Winship [mailto:dan.winship@gmail.com]
Sent: Wednesday, 21 December 2011 1:53 AM
To: Jonathan Billington
Cc: ietf-http-wg@w3.org; Sonya Arnold
Subject: Re: Comments on Section 6.1 (Persistent Connections) of HTTPbis Part 1, version 17

On 12/19/2011 09:22 PM, Jonathan Billington wrote:
> 1.  Section 6.1.1 Purpose

So I've also been reviewing the -17 specs, and was going to suggest that
6.1.1 is no longer relevant (except for the "SHOULD implement" part at
the end) and should just be removed. Everyone agrees that persistent
connections are a good thing now; we don't need to spend a whole page
citing 15-year-old research to convince the reader.

-- Dan

Received on Tuesday, 20 December 2011 22:48:09 UTC