On Mon, Sep 17, 2012 at 3:11 PM, Carsten Bormann <[email protected]> wrote:

> On Sep 17, 2012, at 22:09, Dick Hardt <[email protected]> wrote:
>
> > Is there any reason to NOT use URL safe base64 encoding?
>
> No.
>
> I think Jim's arguments are very confused about the reasons for providing
> choice in a protocol.
> Choice is fundamentally bad for interoperability.  Always.
> However, you do provide choice when there is something you win from this
> choice.
> (E.g., crypto agility is demonstrably necessary.)
> If you can't demonstrate such a win, you eliminate choice as much as
> possible.
>
> Since there is zero advantage for base64 over base64url, this is a
> no-brainer: nail it down.
>
> For those that would like another example for this: In the IETF, we chose
> network byte order.  Once.  For everything.
> Other protocols have a hodgepodge, or, worse, actively negotiate byte
> order.
> (There are some theoretical advantages for doing this between two systems
> that are one of the other byte orders.
> In reality, it means you have to provide two code paths and neither is
> optimized as well as it could be if the choice had been nailed down.
> And you get a nice helping of unexpected interop failures on top.)
>
> Again: nail it down.  Once.  For everything in this protocol.
>

Well said. Agree with everything in the argument above.


>
> Grüße, Carsten
>
> _______________________________________________
> jose mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/jose
>



-- 
--Breno
_______________________________________________
jose mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/jose

Reply via email to