Hi Dianne,

Thanks for your response.  I'd like to address some of your points.

> users should not worry about the apps they install doing unexpected harmful 
> things

I believe that limiting the functionality of the phone/API enough to
achieve complete safety for any application would be very difficult.
If it is even possible, then I think it would result in a neutered
platform.  So why not require digital signatures or user approval for
risky application behaviors instead of removing the functionality in
question?

> the security side is just as critical but a lot more subtle ...
> just saying that platform X does it so it is okay to do it on Android as well 
> is not enough.

I don't deny the importance and subtly of the security issue.  That is
why I thought it was important to note what the rest of the industry
has done.  Companies like Nokia share the concerns of Google about
developing a healthy application ecosystem but they have gone much
further then I am suggesting and made this bluetooth functionality
wide open - I think this observation is very relevant.  There is a
huge amount of empirical data for you to draw upon now: hundreds of
millions of phones, from all brands except RIM and Apple, that allow
applications the functionality in question.

Tom.

On Jan 6, 10:26 pm, "Dianne Hackborn" <hack...@android.com> wrote:
> I think you are being overly dismissive of the security repercussions.  One
> of our goals with Android is to create an open and thriving third party
> application market.  Two cornerstones of doing this is that it should be
> dead easy for a user to find and install an application, and users should
> not worry about the apps they install doing unexpected harmful things.  The
> former is a fairly clear goal, but the security side is just as critical but
> a lot more subtle -- each little piece you take out of the security picture
> is a growing, long-term detriment to the user's trust.
>
> From the very start, our approach with Android security has been as
> conservative as possible: if there is an application feature that seems
> dangerous, we'll try to either take the time needed to think about and
> address of its repercussions, or wait on making it available.  The same
> approach needs to be taken here, everything thought through before making it
> available.  This sounds like it requires enough thought that it probably
> doesn't make sense to have in a 1.0 API (though Nick would know better than
> I).
>
> At any rate, just saying that platform X does it so it is okay to do it on
> Android as well is not enough.  PalmOS lets apps patch the core OS to insert
> their tendrils into everything it does; should that also be allowed on
> Android?  I wouldn't think so.  That isn't to say a feature shouldn't be
> there, but it should be done with thought and consideration to all of the
> repercussions it has.
>
>
>
> On Sun, Dec 21, 2008 at 10:01 PM, Qwavel <qwa...@gmail.com> wrote:
>
> > Nick,
>
> > Thanks for participating in this open conversation about the bluetooth
> > API - this is the first time that I'm aware of that outside developers
> > have had the opportunity to express themselves at this stage in the
> > development of a phone OS/API.
>
> > As I'm sure you are aware, Bluetooth data connection between apps are
> > supported by JSR82.  To the best of my knowledge, the only platform on
> > which pairing is required for these connections is the Blackberry.  As
> > far as I can tell, this was done for the pretense of security since
> > the platform was originally only targeted at the enterprise market.
> > On the Blackberry dev forums I regularly see confusion and surprise
> > about this restriction.
>
> > The only other platform (beside the Blackberry) which really limits
> > bluetooth is the iPhone, but this is expected of Apple.
>
> > I am being dismissive about the security advantages of the blackberry
> > approach for these reasons:
>
> > - The majority of phones available now (in Europe but not in the US)
> > allow full access to JSR82, without requiring pairing, and without
> > even requiring that the midlet be signed.
>
> > - More importantly, I've not encountered any regret about this, or any
> > sense that it is a mistake.  Instead, easy access to JSR82 is
> > spreading: now, even LG and Samsung are starting to provide this.
>
> > - Security concerns like this should not be addressed by limiting the
> > functionality of the system, when they can be addressed at the
> > application security level.  I can't comment on the difficulty of
> > implementing this, but certainly it would be better to produce an OS
> > that is not limited in the way that the BB and iPhone are.
>
> > If you really believe that bluetooth communication without pairing is
> > a security hole - and I believe that Nokia and SE have shown that it
> > isn't - then I think it would be better handled by the application
> > level security mechanisms.
>
> > Thanks,
> > Tom.
>
> > On Dec 3, 12:22 pm, Nick Pelly <npe...@google.com> wrote:
> > > We are likely to prevent Bluetooth data connections (RFCOMM) from apps
> > > unless the two phones have been paired. It's really hard to make
> > > security work any other way.
>
> > > Nick
> > > Android Systems Engineer
>
> > > On Wed, Dec 3, 2008 at 1:37 AM, whitemice <markbr...@zedray.co.uk>
> > wrote:
>
> > > > Hi Nick
> > > > While we are on the subject, I am looking for Android *Ad-hoc*
> > > >Bluetoothsupport.
>
> > > > Example: Alice and Bob both have my client running on their phones,
> > > > and walk withinBluetoothrange of each other in a social setting.  I
> > > > want the application to:
> > > > (a) Be able to detect the otherBluetoothphone in the room
> > > > (b) Detect that the same application is running on the other phone
> > > > (c) Create a data connection between the two phones without asking for
> > > > the user's permission (permission is granted beforehand).
>
> > > > Is this considered a security problem, or will this kind of thing be
> > > > allowed in the new API?
>
> > > > Some more info on what I am doing….
> > > >http://blog.zedray.com/snowball/
>
> > > > Regards
> > > > Mark
>
> --
> Dianne Hackborn
> Android framework engineer
> hack...@android.com
>
> Note: please don't send private questions to me, as I don't have time to
> provide private support.  All such questions should be posted on public
> forums, where I and others can see and answer them.
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google
Groups "Android Developers" group.
To post to this group, send email to android-developers@googlegroups.com
To unsubscribe from this group, send email to
android-developers-unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/android-developers?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to