Re: WebFinger compromises

On Oct 31, 2012, at 11:43 AM, Evan Prodromou <evan@status.net> wrote:

> On 12-10-31 11:48 AM, Dick Hardt wrote:
>> +1 on everything. 
>> 
>> A simple, easy to understand spec that solves the major use cases released soon is far superior to kitchen sink spec that solves all use cases that is released in a year.
> I think that RFC 6415 does that:
> 
> http://tools.ietf.org/html/rfc6415

I reread it. It looks complicated, specifically supporting both XML and JSON.

> 
> It describes pretty clearly what most of us have meant by "Webfinger".
>> JSON only (if that is not obvious, you need to write some code this decade)
> I don't think this is reasonable.

"reasonable" is an opinion. You are being a polite Canadian. I know one when I see one. :)

> We're lucky to have had some quick implementations of Webfinger and it's not clear how or when they'd be updated to JSON. For Open Source implementations like StatusNet, we just don't have the ability to force people to roll out new versions.

Are these deployments being heavily used? If not, I don't see it as a significant issue.

> 
> JSON preferred, XML optional is probably the only way to go forward.

Both are effectively required if that is the case. I would prefer to deal with just XML rather than support both. Given a choice between XML and JSON, JSON wins EASILY!

Supporting both adds MASSIVE complexity to understanding and working with the specification.

>> 1 round trip vs 2 round. Pick one that is simple to implement. Let's not get caught up in optimization. Brad's comments below seem sane (as usual)
>> 
> The current spec is here:
> 
> http://tools.ietf.org/html/draft-ietf-appsawg-webfinger-02
> 
> It describes a 1-round-trip extension to RFC 6415.
> 

I prefer the two round trips so that servers don't have to serve dynamic pages, but that is only my opinion.

I STRONGLY prefer there be just one way to do it. Either it is 1-round trip per the ID or it is 2 round trips.

> Here's what I think is important: we need to build applications that use Webfinger. There's a virtuous cycle we haven't yet kickstarted. Things like Simple Web Discovery are a step in the wrong direction.

Often, people start things like Simple Web Discovery because the alternative was too complicated or did not solve their problem and the comity around the alternative did not solve their problem.