Thanks for your feedback.
Happy to share information about the project.
I'm keen to explore decentralised forms of personal data management
(which contrast with the Ralph Lauren, LexisNexis, Amazon and Loyalty
Card centralised approaches).
The approach is predicated on the following...
* Exhaustive Profiling (storing structured information about user
behaviour, information they have seen/created)
* Secure Decentralised Networking (as offered by i3, hip and uip)
* User Hosting (typical of peer to peer applications, but a perfect
complement to exhaustive profiling - the information never leaves
your desktop unless you say so)
The vision is one of users themselves being the primary repository of
information in a way which is flexible and powerful enough that it
would be feasible for centralised services to cease storing profiling
information about users. To take an example built around Piggybank,
if Amazon purchase pages can be scraped for metadata which is then
stored and retained by me concerning my purchases, then
* Amazon doesn't need to store the information (since they can
retrieve that information at any time, assuming I allow them -
reinforcing my ownership of the information)
* I can tout the information to Barnes and Noble (since they have an
equivalent access to Amazon, if I allow them - the cost of switching
providers can be very low)
* I can use the information to populate my "Delicious Library"
application
* (etc. etc.)
...but for the "exhaustive profiling" part, I need to auto-launch
piggybank scrapers and autosave their results, otherwise users may
forget to push the button at the key page.
The Amazon example may not be the best one, but it gives an idea of
the starting point for this work - that in principle, the centralised
provider can do everything the same without storing the information,
but that users managing their own information means that they can
engaged in the decision process, and get to legislate on an ad hoc
basis how their personal information is used.
More information will appear on the BT MIT collaborative website
linked below as the project develops.
Cefn
http://cefn.com
BT MIT Collaborator
http://that.labs.bt.com/ezpublish/
On 25 May 2005, at 20:56, David Huynh wrote:
> Cefn Hoile wrote:
>
>
>> Is there a mechanism to auto-launch Piggybank scripts when a URL
>> match is made (a la Greasemonkey) as opposed to requiring users
>> to notice the data coin and manually activate it?
>>
>
> No. I believe both modes are desirable.
>
>
>> Equally is there a mechanism to auto-save the model which a
>> Piggybank script builds, as opposed to requiring users to notice
>> the save button and manually activate it?
>>
>
> You can bookmark anything viewed in a temporary model and PB will
> try to re-create it later when you browse to that URL. But no,
> there is no mechanism to auto-save.
>
>
>> A combination of these two mechanisms would be ideal for the
>> application I am imagining at the moment. I could try some kind
>> of hack to execute this (perhaps using Greasemonkey) but
>> wondering if I've missed something in Piggybank.
>>
>
> Perhaps you could talk more about the application you're imagining?
> Maybe other people here are interested in similar applications.
>
>
>> Where should I be looking for this kind of information?
>> (aside from going through the source of course, which is the next
>> step)
>>
>
> PB actually has a very small feature set. Everything you can do you
> see in the user's how-to list on the web site. :-)
>
> As for source code, I'd suggest looking at
>
> piggy-bank/firefox/chrome/content/overlay.js
>
> That's where things start. The Developer's Guide on the website
> also has pointers to where things are in the codebase.
>
> David
>
>
>> Cefn
>> http://cefn.com
>>
>
>
Received on Thu May 26 2005 - 15:07:57 EDT