On Sun, 30 May 2010 11:14:54 -0700 Abraham Williams <[email protected]> wrote:
> On Sun, May 30, 2010 at 11:01, Bernd Stramm <[email protected]> > wrote: > > > The user does trust the app, otherwise they would not be using it. > > The problem with the scheme of using the app *and* a browser is > > that the user has to trust *both* of them. > > > > And if they don't trust the app, why are they using it to post their > > tweets? > > > > Trust is not a boolean value. There are levels of it. I trust my > mobile browser to not take over my Twitter account but I only trust > random new Twitter client to not post spam. If the Twitter client > breaks my trust it is easy to revoke access to it. Is it easy for most users? The authentication token doesn't expire, so an application (any application, not just desktop/mobile client) can do what it wants for quite a while. You should be careful with trusting browsers, mobile or otherwise. They are very leaky. And my point remains, when using a browser *and* a standalone client, the user trusts both of them. From the point of view of the honest application developer, I do not want to assume that another application (the browser) which is completely unknown to me, is trustworthy. Hence I prefer the solution with the small integrated webview in my application. But also, what do people to with their twitter account that needs protecting, other than posting messages and media content? And of course giving away great data to the data miners. A lot of this authentication business misses the easiest intrusion vector anyway. People will steal your phone, and have access to everything you store on it. Including any authentication for serious business. Nevermind trusting the standalone app or the browser, the entire system is easily compromised if its stolen. -- Bernd Stramm <[email protected]>
