Use of BGP label block

I have a question from an operators perspective regarding the use of the
label block in the BGP signalling drafts. The recommended algorithm uses a
CE or VE identifier as an index to determine the correct label a receiving
PE should use to send packets destined for a particular VPN to a particular
PE. The drafts state that the CE/VE IDs must be unique within a particular
VPN but not across VPNs.

This doesn't seem too bad in the intra-provider case as the operator can
manage his own CE/VE ID space and simply allocate the next CE to join a VPN
the next available ID for that particular VPN. However, in the inter-AS case
how are operators to co-ordinate their use of CE/VE IDs? For large operators
that might peer with several other large operators, keeping track of which
CE/VE IDs belong to which operator, and mapping the CE/VE IDs to the correct
VPNs doesn't sound appealing.

Obviously when choosing a signalling protocol there are many things to take
into consideration, and as a result an operator may decide that having to
manage CE/VE IDs is a minor consequence. However, generally speaking the use
of a label block to overcome BGPs inability to signal labels on an
individual point-to-point basis will lead to an increase in operational
costs (compared to signalling mechanisms that don't use a label block and
therefore CE/VE IDs do not need to be provisioned and managed).

Are there any plans to produce guidelines on how CE/VE IDs should be
provisioned and managed, particularly in the inter-provider case? 

Thanks,

Richard