RUN 2B L1CAL MEETING MINUTES
			     10 July, 2002
			 (Oklahoma Workshop)


Jovan Mitrevski		 Simulation Results from Nevis
---------------
o See Jovan's slides linked off of the agenda for details
  - justification for the baseline algorithm proposal can be found in
    his talk

Baseline Sliding Windows Algorithm Proposal
-------------------------------------------
o Based on recent simulation results we are proposing the following as
  the algorithm to be pursued in the TABs
  - Jet Algorithm: 2,1,1 (Emmanuelle's notation)
    . RoI = 2x2 TTs
    . Local Maxima found be comparing RoIs in a 5x5 region
    . Jet Cluster Energy from 4x4 region
  - Electron Algorithm
    . RoI = 2x2 TTs
    . Local Maxima found be comparing RoIs in a 5x5 region
    . EM Cluster Energy from 2x2 region
    . EM Isolation from ring of TTs directly touching RoI
    . Had Isolation from 4x4 TTs behind RoI and EM isolation ring
  - Tau Algorithm
    . still under discussion
o We are also proposing the following architecture for the TABs
  - Each TAB considers:
    . input TTs:          40x9 in etaxphi
    . output Local Max's: 31x4 (because of ICR and edge effects)
o Some of the advantages of this scheme are
  - performance in terms of efficiency vs background rate curves is as
    good or better than all other algorithms considered
  - 2x2 RoI ==> better position resolution than 3x3 RoI on jets and EM
    clusters 
  - 2x2 RoI ==> less "double-counting" of single electrons as 2 jets
    than for 3x3 RoI's
  - sending all 40 eta TTs to a single board ==> ICR energies can be
    included in jets and EM clusters
  - sending all 40 eta TTs to a single board ==> more efficient for
    transfer of data to Cal-Trk match board
o The main consequence of this decision is that the cable mapping
  between the ADFs and TABs will change
  - we are working on a scheme that will minimize the impact of this
    change, but there will be some effect on how the ADF sends out its
    data.