RDF: Persistent Client Store
The stores used by Aurora are all for the user and are used to store personal information such as bookmarks and history. A website cannot easily gain access to these datastores (and most likely, the user would not like the website to be able to read his/her bookmarks etc.)
This technology can be repurposed to provide the website a client side persistent store. The website can read/write to this store using a javascript reflection of the RDF APIs.
The user sets size and longevity constraints on these databases on a per domain basis. Each domain can ask for its own database. A domain cannot read the database of another domain. However, a domain can explicitly give read/write permission (of its PCS) to another domain. E.g., cnet.com might give access to snap.com.
This feature can also be used to implement off-line forms and other such applications. The database is stored as an RDF file (one per domain). So, it can easily be uploaded, etc. At this point, there is no facility to store binary data into PCS, but this can easily be added.
Most of the work to enable this feature has already been done. The big remaining part is javascript reflection of the RDF APIs.
Last updated: $Id: pcs.html,v 1.2 1999/03/18 09:07:58 daniel.brickley%bristol.ac.uk Exp $