- From: Alex Russell <slightlyoff@google.com>
- Date: Thu, 14 Mar 2013 14:17:01 -0700
- To: Mountie Lee <mountie.lee@mw2.or.kr>
- Cc: Matthias Dugué <mdugue@clever-age.com>, "public-webcrypto-comments@w3.org" <public-webcrypto-comments@w3.org>, Nicolas Hoizey <nhoizey@clever-age.com>
- Message-ID: <CANr5HFUMdExf_V-TvD9z5kbReZ9CSZFO5FJ-74g1OWM-j8LALQ@mail.gmail.com>
On Wed, Mar 13, 2013 at 10:24 PM, Mountie Lee <mountie.lee@mw2.or.kr> wrote: > Hi. > let me add my comment for your questions. > > > On Thu, Mar 14, 2013 at 1:53 PM, Matthias Dugué <mdugue@clever-age.com>wrote: > >> Dear WebCrypto WG members, >> >> We are web developers, working in a web agency for various clients and >> various projects, from institutional to e-commerce websites. >> >> We looked very closely into the WebCrypto API. Here are some thoughts, >> from our point of view: >> >> In our opinion, nowadays, the low-level API requires some skills that the >> average developer doesn't have, which makes us think the high-level API is >> even more relevant and should be delivered quickly. Admittedly, it is a >> low-level API, but this makes it hard to use for a broad spectrum of web >> developers, mostly because of its permissive nature with many security >> choices, namely algorithms. A lot of us don't understand the implications >> of these choices, and think the learning curve is too steep. >> > currently no consensus for high-level API. > True. > some member need it > but > some member (like me) think do not need it. > That should not block it if other members find value. > because the higher-level API can be contributed by 3rd party community > (jQuery team will be good example) > As a general rule, this argument does not work. The idea that "users will consume a higher-level wrapper" most of the time means that your group is either committing design malpractice by miscalculating the costs of the wrappers or by excluding your largest motivating use-cases. It is design failure not to deliver a feature that meets the needs of most of your potential users. > While recommending a set of algorithms to implement, the low-level API >> does not guarantee a consistent implementation across browsers, which may >> generate reluctance in adoption. We fear that some browser editors will >> refuse to implement some algorithms due to their policies or technical >> difficulty, which will create a fragmented support, like the one we see >> today with the <video> codecs. >> > Considering the key support and access, the fact that low-level API >> allows to retrieve a key which is not present in the browser environment >> (and thus create an error) will be confusing. We believe it would be >> preferable not to present a key whose material is not available, rather >> than offering to use it but throw an error when trying to access it. This >> reminds us of the <video> support and its canPlayType() method ( >> http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#dom-navigator-canplaytype), >> that can return *probably* if the media is playable, which is better >> than *maybe* but worse than *yes*… Please, we need strong support when >> we need to deal with APIs, not *probably* support! (-*is the key >> available ?* -*probably*…) >> > similar issues (http://www.w3.org/2012/webcrypto/track/actions/61) were > also raised in previous F2F meeting. > and my answer is some testing code is enough in lower level. > your example "canPlayType()" also belong to high-level. > > >> Finally, the use case for certificate management is missing (as a >> simple and attractive means to implement the user/application >> authentication). We're very impatient to see the outcome of your efforts to >> bring us a robust API to build crypto methods. But, as web applications >> developers, our first emergency is a simple to use and robust API, to deal >> with certificates/authentication, in order to prevent the security hole >> that is currently the login/password couple. >> > certificate management issue is one of secondary issues. > I expect a draft version of certificate as a different document from > current API spec will be suggested working group soon. > My team is preparing the proposals and you can review and add your comment. > > >> Thanks a lot for your work and your interest in reading this email, >> we'll be closely following your drafts! >> >> >> Matthias Dugué (@madsgraphics) >> >> Nicolas Hoizey (@nhoizey) >> >> >> -- >> Matthias DUGUÉ - http://www.clever-age.com/ >> Agence de Paris >> Mobile : +33 6 13 12 73 75 >> Tél. +33 1 53 34 66 10 >> >> Clever Age - *Digital Architecture* >> *Clever Garden - Digital Landscape* >> Clever Presence - *Digital Running* >> > > > > -- > Mountie Lee > > PayGate > CTO, CISSP > Tel : +82 2 2140 2700 > E-Mail : mountie@paygate.net > > ======================================= > PayGate Inc. > THE STANDARD FOR ONLINE PAYMENT > for Korea, Japan, China, and the World > > >
Received on Thursday, 14 March 2013 21:17:59 UTC