Re: [rtcweb] Alternative decision process in RTCWeb

[Harald Alvestrand <[e-mail]>]

On 12/02/2013 10:04 PM, Dave Crocker wrote:
> On 11/29/2013 12:22 AM, Roberto Peon wrote:
>> The only reason to specify a must-implement is to increase interop; if
>> mandating a codec does not increase the amount of interop, why do it?
>
>
> If the base specification does not provide enough information for 
> basic interoperability, what is the benefit in standardizing it?
>
> An alternative that I believe has already been mentioned is to 
> standardize /both/, but separately.
>
> Under two different names, which then lets the market decide on 
> whether either will succeed.
>
> Letting the market decide amongst competing choices used to be 
> something the IETF did more commonly.  One of the more colorful 
> examples was SNMP vs. CMOT.

I prefer to remember (not fondly) SNMPv2U vs SNMPv2*.

What the market told us at that time, *very* loudly, was that there 
would be no massive deployment of a properly secured version of SNMP 
until the IETF reached a decision.

The IETF eventually reached a decision (SNMPv3). I don't know if it 
eventually mattered that much; SNMP's grand ambitions had been overtaken 
by events.