Re: [rtcweb] RTCWEB Data Channel: message interleaving

[Paul Kyzivat <[e-mail]>]

On 2/5/14 5:35 PM, Makaraju, Maridi Raju (Raju) wrote:
>> Will NDATA support be required for rtcweb?
> [Raju] Yes. It is mentioned in http://tools.ietf.org/html/draft-ietf-rtcweb-data-channel-06#page-3 which says the following:
>
>     The SCTP base protocol specified in [RFC4960] does not support the
>     interleaving of user messages.  Therefore sending a large user
>     message can monopolize the SCTP association.  To overcome this
>     limitation, [I-D.stewart-tsvwg-sctp-ndata]  defines an extension to
>     support message interleaving.  Once such an extension is available,
>     it SHOULD be used.
>

I guess that begs the question of what "Once such an extension is 
available" means, and what "it SHOULD be used" means. :-(

"available" could mean:

- once it is defined in a published RFC
- once it is defined in a WG draft or RFC
- once it is defined in an individual draft
   (e.g., draft-stewart-tsvwg-sctp-ndata-03)
- once it it is implemented in the browser you are using

"SHOULD be used" could apply to:
- browser implementers
- application implementers

And then there is the question of what "SHOULD" means. Specifically, 
what are the conditions where it would be acceptable to *not* follow this?

I *hope* this is intended to mean that browser implementers MUST 
*implement* it, and *use* it in their implementation of DataChannel once 
draft-stewart-tsvwg-sctp-ndata becomes an RFC. And SHOULD do so now 
based on draft-stewart-tsvwg-sctp-ndata-03.

	Thanks,
	Paul