Re: [hybi] Call for Consensus: draft-tyoshino-hybi-websocket-perframe-deflate as WG Item

[John Tamplin <[e-mail]>]

On Thu, Mar 1, 2012 at 2:59 PM, Takeshi Yoshino <[e-mail]> wrote:

> On Tue, Feb 28, 2012 at 09:05, John Tamplin <[e-mail]> wrote:
>>
>> I think it needs some significant changes (in particular, if it is going
>> to allocate a reserved frame bit it needs to be extensible to other
>> compression algorithms, which also implies renaming it to compress-frame or
>> similar), but I think it is a fine starting point.
>>
>
> Do you think drastic abstraction is better? I.e. should we put texts
> describing generic per-frame compression spec and texts describing how to
> use it with DEFLATE into separate sections, or even into separate specs?
>

I think it is fine to have it in the same spec.  It can describe the
general use of per-frame compression, and then supply the specific case of
using deflate as the algorithm.  I don't think any more is required than
reordering pieces of the doc and some name changes, as well as adding some
way of selecting the compression algorithm (and supplying parameters to it
if necessary).


> If it's sufficient to make it clear that the bit allocated in this spec is
> also available for any other per-frame compression method, I'd replace
> "COMP" bit in the spec with, say, "Per-frame compressed" bit, and change
> the registration section to:
>
> 7.2.  Registration of the "Per-frame compressed" WebSocket Framing Header
> Bit
>
>    This section describes a WebSocket framing header bit registration in
>    the WebSocket Framing Header Bits Registry.  [RFC6455]
>
>    Header Bit
>       RSV1
>
>    Common Name
>       Per-frame compressed
>
>    Meaning
>       Per-frame compression is applied to the frame or not.
>
>    Reference
>       Section 4 of this document.
>
>    The "Per-frame compressed" framing header bit is used to indicate
> whether
>    any negotiated per-frame compression extension is applied to the frame
> or not.
>
>    When this bit is used by this per-frame DEFLATE extension, it means
> whether
>    "Application data" part of the frame contains octets generated using
>    DEFLATE or not.
>
>    Any other compression extension MAY define
>    its own use of this bit.  Note that this per-frame DEFLATE
>    extension and that extension will be incompatible in such case.
>

I would prefer a bit more generic definition of the bit, and rather than
saying that this is the per-frame deflate extension it is a general-purpose
per-frame compression extension, and deflate is merely a required option
among potentially many different compression algorithms.

-- 
John A. Tamplin
Software Engineer (GWT), Google