- From: Dirk Balfanz <balfanz@google.com>
- Date: Tue, 08 Mar 2016 06:07:33 +0000
- To: "Hodges, Jeff" <jeff.hodges@paypal.com>, W3C WebAuthn WG <public-webauthn@w3.org>
- Message-ID: <CADHfa2C-3ay6M75s=zeoHAUo3q4WDNNVPz9LSgmi8Z9aJY5T=Q@mail.gmail.com>
"FIDO" is vendor-neutral. Why do they need to be standards-org-neutral? Maybe something along the lines of "cryptographic authentication credential"? Dirk. On Mon, Mar 7, 2016 at 3:57 PM Hodges, Jeff <jeff.hodges@paypal.com> wrote: > Beyond a simple cut-n-paste-and-jam-em-all-into-one-file approach to > merging the three source specs (web-api, signature-format, key-attestation) > info a single spec file, there's the issue of figuring out how to > de-FIDO-ize the text therein. > > There's terms such as "FIDO 2.0 credential", "FIDO assertion", etc strewn > throughout. > > The key, it seems to me, as we'd briefly chatted about in the #webauthn > irc channel during the meeting last Fri, is figuring out how to refer to > what is presently termed "FIDO Credentials" in the web-api and > key-attestation specs.. > > > grep -li "fido cred" ./*/Overview.html > > ./webauthn-key-attestation/Overview.html > > ./webauthn-web-api/Overview.html > > I took at look at the SiteBoundCredential term in the Creds Mgmt spec < > http://w3c.github.io/webappsec-credential-management/#siteboundcredential> > and that doesn't actually map to FIDO Creds because the former are bound > to a web origin [RFC6454] and the latter are bound to a Relying Party's > domain name reduced (aka "domain lowered") to eTLD+1 (eTLD = effective Top > Level Domain, aka Public Suffix), which is also known as "Relying Party > Identity (RPID)" in the submitted fido specs. > > So we ought to figure out what to rename "FIDO Credentials" to, in a > vendor-neutral, standards-org-neutral manner. > > some ideas I've heard or thought of.. > > Origin-bound strong creds (OBSCreds) [won't work because not > binding to origin] > > Scoped strong creds / scoped creds (SSCreds) > > RPID-bound strong creds (RBSCreds) > > > Basically, in looking through the specs, it seems that if we nail down the > name for the credentials, then the names of the other things (e.g., > assertions, extensions, etc) will follow fairly easily. > > WDYT? > > =JeffH > > > >
Received on Tuesday, 8 March 2016 06:08:13 UTC