I finally managed to put some time on my OAuth application... I
released a proof of concept and would like your opinion on the matter.
Basically, it is a Android OAuth broker as discussed above.

On another note, I red about the OAuth AccessorSecret which could be
interesting in situation similar to the one discussed. However, I
noticed that not many service providers support the AccessorSecret
describe here http://wiki.oauth.net/AccessorSecret

The application is located on GitHub and can be found here:
http://github.com/novoda/oauth_for_android

I still need to work on the documentation a bit for easy install -
brightkite is not working due to the lack of support for
oauth_Callback on their side. Jaiku does not work either as they don't
support callback. MySpace and twitter should work.

Any comments are welcome.

Thanks
Carl

On Aug 26, 7:29 pm, "Carl H." <charr...@gmail.com> wrote:
> Sorry for the delay on this, I had deadlines to meet which made me
> drift away from my original OAuth application. Comments inline:
>
> << Yes, a service provider should be able to unauthorize a consumer
> easily. >>
> All devices will use my consumer keys to connect to service providers
> (I wish I could do this differently). The user will choose the
> services they want to use and authorise my consumer key adequately.
> This is where I am a bit twitchy as if one device end up being
> pirated, the service provider might cancel the key instead of the key
> for that specific user.
>
> << Android provide some way for one software module to identify
> another, before cooperating with it? >>
> Yes and it is what I am currently doing. Each call to my application
> will ask the user if he authorises application X to use MyOAuth to do
> an OAuth call on his behalf. He can answer by "always", "yes", "no" or
> "never". I keep track of all applications calling each service
> provider and let user pro-actively choose how to handle malicious
> calls. Application X does not require any OAuth information. It calls
> MyOauth application and receives the output of that call. As such it
> does not care about any OAuth signing and so forth. I also intend to
> have each application register itself with its own consumer key.
>
> Further to the above fine grained security, I am also taking a similar
> approach to Google in regard on its way to access OAuth from Android.
> It is apparently possible to access Google's OAuth infrastructure from
> within your application if the following is part of your application's
> manifest (the manifest defines which permissions/services/activities
> your application runs):
> <uses-permission
> android:name="com.google.android.googleapps.permission.GOOGLE_AUTH">
>
> This is not very well documented and unsure how well it works but the
> idea is similar. If an application uses OAuth calls from MyOAuth, the
> application is required to have my own OAuth permission in its
> manifest. As such when installing the app onto the device, the user is
> requested to accept the installation while knowing it will access
> OAuth info.
>
> Throughout my experience within Android, several applications access
> OAuth based service providers. I just want to have it centralized in
> one provider in order for the user to have better control of which app
> does what.
>
> I will get back to my OAuth app next week and hopefully will have it
> documented online shortly. This will give a better idea of what I am
> currently doing and happy to accommodate any comments.
>
> Cheers,
> Carl
>
> On Aug 21, 6:08 pm, John Kristian <jmkrist...@gmail.com> wrote:
>
>
>
> > Rather than assign a different consumer secret to each device, I
> > suggest each device get a unique access token secret using OAuth; that
> > is by asking the user to authenticate to each service provider and
> > authorize the device. (The token secret and consumer secret are
> > separate elements in OAuth.)
>
> > Yes, a determined user can find his own token secret. That's OK. The
> > problem is allowing someone else (an attacker) to use the secret. To
> > prevent that, applications should discourage the user from finding his
> > token secret, and hide the secret from other, possibly malicious
> > applications running on the same mobile device. The latter needs some
> > infrastructure, to prevent applications from seeing each other's data.
>
> > For a consumer copied to users' mobile devices (or desktops), a
> > conventional consumer secret isn't very useful because it isn't really
> > secret. Assume attackers will know it. You might as well publish it,
> > as Google did. Look elsewhere for assurance that consumer software is
> > trustworthy. I expect any healthy software market will have some way
> > to help users trust the software they use.
>
> > Yes, OAuth doesn't help fight piracy. (I mean the use of unauthorized
> > copies or imitations of licensed software.)
>
> > On Aug 20, 6:38 pm, Sunir Shah <su...@freshbooks.com> wrote:
>
> > > Sorry, the original proposal as I understood it was that every  
> > > application gets a consumer secret for each mobile device. Presumably  
> > > you could arrange this by cooking the binary every download with a  
> > > different consumer secret. This is impossible in mobile environments  
> > > with centralized distribution (iTunes, AppWorld). However, in those  
> > > environments, this isn't really a problem.
>
> > > My first reaction was that there is no way to avoid piracy, since if  
> > > Pirate Alice downloaded the app, she could post it on the Internet and  
> > > anyone (i.e. Bob) could use it. However, I realize now that isn't the  
> > > case, since you can monitor the use of the consumer secret, and if one  
> > > secret seems overly used, you can destroy it disabling all pirates.  
> > > Hello Windows Genuine Advantage.
>
> > > However, I would add in response to your suggestion, you should not  
> > > rely on the premise that it is hard for the user to find their token  
> > > secret since it is so easy to retrieve with common reverse engineering  
> > > tools.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"OAuth" group.
To post to this group, send email to oauth@googlegroups.com
To unsubscribe from this group, send email to oauth+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/oauth?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to