Handling of MIF in clients

[bpindelski]

Multi-image filesets (MIFs) are logically grouped into filesets during import time. The clients (Insigh, Web) need to provide the user with feedback when certain operations are performed on them.

Current state

The 2 primary operations under consideration are chgrp and delete.

Chrgrp is an issue since it is not currently possible to have the same underlying binary data (the original filesets) contained in 2 separate groups. Therefore if one image from an MIF is moved to another group, it will become unviewable. Therefore, for demo 2, the chgrp operation was completely disallowed.

Delete is an issue because a user may accidentally delete more data than was intended. One image of an MIF present in a dataset that is being deleted, does not guarantee that the user intended to delete all the other images in the MIF in all datasets. The current state of delete as of demo 3 is that dangerous deletes are permitted.

Goals for demo 4

Primarily, this story is about getting these permission-related operations for MIFs into a generally releasable state. There are further, more extensive solutions possible (including deep copies and an abstracted IO system) but these will only be considered once base functionality is in a "safe" state. The changes to bring us there include:

• graph definitions on the server should be conservative by default, preventing users from doing things they don't expect;
• clients need to be aware of this conservative behavior and inform users of possible corrective actions;
• and where possible, clients should provide users feedback (and "training") about what MIFs via subtle hints in the UI.

[agilo]

Updated status, related task in progress

[jamoore]

All tasks closed.