Re: Last work item - ACL2?

[Mark Crispin <MRC@xxxxxxxxxxxxxx>]

IMHO, ACL2 should be dropped.

As currently constituted, ACL2 does not address my issues.

There needs to be a means to export a simpler form of access control in 
IMAP, such as provided by a filesystem.  The IMAP ACL framework does not 
offer this capability, and it's become painfully apparent that attempts to 
make the ACL framework do so only make matters worse.

I think that a more productive course, *outside* of IMAPEXT, would be the 
following:

I define a generic specification for access control that sets "read" and 
"write" access for "owner", "group", and "world"; along with a way to "set 
owner", "set/add group".

Alexey, Chris, and any other interested parties define an informational 
RFC that describes how a server that implements both ACL and the simpler 
access control specification should map the two.

Alexey, Chris, and any other interested parties consider a simplified 
version of ACL that deletes tied rights and other features that attempted 
to accomodate filesystem semantics.  The result is that a server that 
implements ACL is required to implement the One True Way of ACL.

The long term view is to make the simpler form mandatory to implement for 
all servers, at least to the point of locking in read-write for owner and 
no-access for world.  ACL would not be mandatory to implement, but with a 
single True Form for ACL would be more likely to be implemented.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.