[sugar] Sharing data between activities.

Samuel Klein meta.sj
Fri Aug 24 00:00:01 EDT 2007

On 8/23/07, Ivan Krsti? <krstic at solarsail.hcs.harvard.edu> wrote:
> On Aug 23, 2007, at 11:11 PM, Samuel Klein wrote:
> > Except for the common situation of a program that wants to read from
> > image, midi, media, text, game, or other files that the user has
> > explicitly chosen to share with the world.
> Right. I'm unopposed to allowing programmatic enumeration and access
> to such files without user interaction as long as the user
> interaction to deem files public in this way is obvious and trusted
> (e.g. from the journal). Eben?
> > The case at hand seems to be a set of data explicitly flagged for
> > public consumption, as a property of the library it is part of [or by
> > its author].  A dedicated space for 'published' material produced
> > while running a given activity could be made world-readable, and
> > programmatically readable by other activities, without loss of privacy
> > or risk of data destruction.
> You don't want an activity being able to deem a user document world-
> readable; you want the user to do it herself.

This depends on what constitutes an activity deeming things
world-readable.  My understanding is that an activity will have an
ambient state of publishing, just as it has an ambient state of
active-sharing, at each point in time -- so that if sharing is set to
"everyone", published files do the right thing for the duration of
that session.

> What about files an
> activity ships? There's an argument to be made that the activity
> ought to be able to deem some of them world-readable in the manifest.
> This doesn't make me _happy_, but I could potentially be convinced
> it's a good idea. So far, no one has put forward a really compelling
> argument.

Why not make this the default state of sharable files that ship with
an activity, and encourage activities to explicitly store sharable
files in an appropriate directory?  I'm not clear on why we want to
make it difficult for activities to flag their contents as
world-readable.  Is there a counter-use case?


More information about the Sugar-devel mailing list