Re: [stir] JWT/JSON (was - Re: Review of: draft-ietf-stir-passport-05)

[Christer Holmberg <[e-mail]>]

Hi Dave,

I don't know whether the WG has yet decided whether there will be options and alternatives (i.e. whether sending of claims/headers will be mandatory, optional or forbidden) - so we don't yet know whether there will be an "if" :)

But, if there will be an "if", 4424bis needs to describe how to correctly handle the different options and alternatives.

Regards,

Christer

Sent from my Windows Phone
________________________________
From: Dave Crocker<mailto:[e-mail]>
Sent: ‎05/‎08/‎2016 17:20
To: Christer Holmberg<mailto:[e-mail]>; Paul Kyzivat<mailto:[e-mail]>
Cc: [e-mail]<mailto:[e-mail]>
Subject: Re: [stir] JWT/JSON (was - Re: Review of: draft-ietf-stir-passport-05)

Christer,


On 8/4/2016 11:45 PM, Christer Holmberg wrote:
>
> The receiver needs to be able to parse JSON if the sender includes the
> claims, in order to verify the signature.

Alternatives and options make specifications more complex and often
introduce potential non-determinacies.  So the 'if' that you cite is not
a small matter.

In practical terms, the 'if' means that verifiers must be able to parse
json as well as encode it.


> Also, whatever headers we include, I assume the receiver should be able
> to parse them.
>
> But, parsing JSON is not a difficult thing to do, and there are
> available libraries for those who don't want to implement the parser
> themselves.

Just to make sure this sub-thread retains its context:  I did not
comment on the choice of JSON/JWT in the actual review.  It's not an
irrational or horrible choice.

But it does add overhead.  It adds it to the effort needed to understand
the specifications.  And it adds it to the software.  (It might also add
it to the execution of the software, but I suspect that is, at worst, a
negligible difference here.)

One of the more deceptive parts of writing standards is the seduction of
"is not a difficult thing to do".  In most cases where that sort of
comment is offered, it is quite true.  The problem is with incremental
complexities.  A not-difficult here; a not-difficult there... They
really do mount up.

By way of example, having to send the reader off to become proficient in
two additional specifications is not a small increment in developmental
overhead, especially when those specification have no natural -- ie,
pre-occurring -- relevance to the current work.

d/
--

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net