RUN 2B L1CAL UNPACKER DISCUSSION
                   9 May, 2005

Present 
o Alberta       R.Moore (tried valiantly)
o Fermilab      A.Boehnlein, Md.Naimuddin, J.Stark, M.Verzocchi
o FSU           T.Adams
o McGill        B.Vachon
o MSU           P.Laurens
o Nevis         H.Evans
o Northeastern  E.Barberis, D.Wood
o Notre Dame    M.Hildreth
o York          W.Taylor


Unpacker Options
----------------
1) IOGen
2) SMT-like unpacker
3) Custom unpacker
   - Modification of current L1Cal unpacker (Marco)
   - Continuation of Todd's hex-dump unpacker
   - these are essentially identical


IOGen used in MC
----------------
o Input from L0CalTower IOGen object

o L1Cal Data output in MC
  - L1Cal-Track = an IOGen object
  - L2 data = IOGen object
  - raw data = nothing yet

o MC vs Data formats
  - could want different MC and data formats


Discussion of Options
---------------------
o What happens if data format is wrong
  - IOGen crashes if e.g. header data is wrong
  - Custom (Marco's) code returns an error if data is wrong

o Advantages of using IOGen
  - handles VRB byte-swapping
  - same unpacker in data as MC

o Custom Solution
  - could be written in ~1 day
    . very easy to do
  - easily modifiable

o Unpacker for L2
  - this needs to be IOGen
  - Roger is working on this

o Types of unpackers
  a) L1 --> L2 (data and MC): has to be IOGen
  b) used by reco/L3
  c) general MC

o Possible Solution
  - custom code for now
  - revisit IOGen unpacker later for possible use in the final system

o How to proceed
  - Todd and Nayeem will work together on this
    . will look over Marco's code as well
  - Schedule
    . code by Friday ?
    . need data in L3 format
    . some issues writing out more than a few events in hardware
      (VRB/SBC issue?)
  - Need to discuss with Roger about integration issues with IOGen/MC
    in the longer term