Re: [ietf-outcomes] what's massive?

[Peter Saint-Andre <[e-mail]>]

On 2/3/10 6:16 PM, Dave CROCKER wrote:
> 
> 
> On 2/3/2010 2:14 PM, Peter Saint-Andre wrote:
>> What counts as massive adoption? In discussing this wiki with some XMPP
>> developers just now, we decided that XMPP adoption is somewhere between
>> "some" and "massive". I'd say the same for technologies like SIP.
>> Perhaps we could have:
>>
>> + = some adoption
>> ++ = significant adoption
>> +++ = massive adoption
> 
> 
> FWIW, here's some background that went into the current design of the
> rating scale...
> 
> This all falls into the category of survey research, which ultimately
> calls for a subjective assessment by a person.  Here, we're trying to
> use community rough consensus to validate the assignments.
> 
> I've been calling the current rating model as '5-points with a tail'. 
> The tail is the '++>' extra value, that refers to work which is so
> successful that it prompts follow-on work.

That was another question I had: what is follow-on work?

> What you are suggesting is that it be a 7-point scale.  There is always
> a desire to add gradations to a scale.  Absent careful training for the
> folks assigning values, having more resolution to the scale actually
> makes things more ambiguous, less consistent, and more variable to the
> respondent.
> 
> In addition, it's not clear how much utility there would be in making
> finer-grained assignments, even if they could be made clearly and
> consistently.
> 
> 
> As for where to rank XMPP using the current scale...
> 
> If there is rough consensus that XMPP is a home run, then it should get
> ++.  If there is rough consensus that XMPP is successful, but not quite
> yet massively in use, then it probably warrants a single +.
> 
> As much as some of us use jabber/xmpp, my own impression is that its
> Internet-scale adoption is significant, but is still limited.

That seems reasonable. I tend to be one of those people who is quite
reluctant to select the extreme options on point scales...

Peter

-- 
Peter Saint-Andre
https://stpeter.im/