[sugar] [Post-trial2] Buddy identification: security and UI requirements?

[Bert Freudenberg]

Just to be sure: I assume you are referring to a petname sytem, Eben?

http://www.skyhunter.com/marcs/petnames/IntroPetNames.html

I think this is indeed a very nice and simple solution to a hard  
problem.

- Bert -

On Jul 21, 2007, at 23:21 , Eben Eliason wrote:

> I, too will need to think in much more depth about this problem.  I
> know I discussed a few of the implications with Dan as he began
> implementing the PS early on.
>
> From an interaction design perspective, I don't think that it's
> necessary to allow name changes (or at least, we don't need to make it
> trivial, like changing colors will be).  The idea is that children
> will be identified by their given name, and as such I'd rather not
> refer to a "nickname", unless we actually make the distinction between
> a nickname (changeable) and real name (much less so).  This doesn't
> guarantee security, of course, but it does make it harder to
> impersonate someone.
>
> Also, we discussed possibilities for what might happen when a child
> changes colors, photos, etc.  One possibility is that these children
> are rendered slightly differently in the UI, perhaps with a badge.  By
> clicking on them (or some interaction not yet determined...perhaps for
> friends you actually get a notification message) you will be able to
> see the former colors/photo (as stored on your machine) and the new
> colors/photo (as provided by them).  This would make it rather
> difficult for Bob to spoof your friend Alice, assuming that you have
> ever in the past received Bob's identity info, since he would always
> have to advertize "I'm Alice, formerly known to you as Bob."
>
> I suppose this would actually be sufficient if we also specially
> identified friends in the UI as well, as you mention, perhaps by
> applying a star badge to them.  In that case, any friend of yours who
> attempts to spoof another friend of yours will be rendered the same,
> but will also initiate an "identity change" message.  Any non-friend
> of yours - even those you've never seen before - can try as they might
> to spoof a friend of yours, but can never have the star badge.
>
> As another note, when these changes in identity occur, its probably
> beneficial for the children to view the "identity change" message
> before the interface adjusts the rendering of the XO whose identity
> changed.  That way you can find Alice rendered in pink/purple as
> you're used to, without being temporarily confused by the fact that
> she changed overnight to green/yellow.
>
> As far as unique naming goes, as I mentioned, I'd really prefer to use
> real names.  Real names, of course, aren't always unique.  I'd like to
> use first names only by default, adding the last initial in the
> display when first names conflict.  Of course, when the last initial
> conflicts, we could use full last names.  Beyond that, we're mostly
> out of luck, but perhaps that doesn't matter in the general case.  If,
> however, you want to make two people with identical names your
> friends, the "make friend" process could check the list of current
> friends, find the match, and alert you to it, offering to let you
> modify the name that you locally use to represent the new friend.
> This has the added bonus that, if an outsider you've never before seen
> tries to spoof a friend, and you try to make the outsider your friend
> thinking that you accidentally de-friended the person they're
> spoofing, you'll get the duplicate friend message and think twice
> about continuing.
>
> - Eben
>
>
> On 7/21/07, Ivan Krsti? <krstic at solarsail.hcs.harvard.edu> wrote:
>> On Jul 21, 2007, at 4:07 PM, Simon McVittie wrote:
>>> (Mailing this now while I remember, because this is something we
>>> need to
>>> think about relatively soon after Trial 2, I think. Daf and I will
>>> be in
>>> Boston on Monday, Tuesday and Wednesday trying to debug
>>> collaboration, so it'd
>>> be great if we could talk to Ivan and Eben about this while we're
>>> there.)
>>
>> I'll read the rest of the message in a bit; can you meet with me at
>> 2PM Monday? I'm extremely busy the first three days of next week.
>>
>> --
>> Ivan Krsti? <krstic at solarsail.hcs.harvard.edu> | http://radian.org