1.  I believe that one can get all the si hits.  All the optical signals =
can be split.  The number of VRB's gets large if we do every si read out =
fiber.  

2. I see no problem in sending the cluster information to L3.  One will ned to 
do it for mark and pass checking anyway. 

3. I am continuing to make progress on a readout design for all tracks to 
1.5 GeV/c.  I now believe that we have a good way of reading them out.  
This means that the L1 roads should be available at L1 time so that the 
STT can be prepared to select the hits as the SVX is read out.  The 
details are not completely done so there could still be problems.  I hope 
to do this before the NIU workshop.
We will send 6 tracks per sector (480 total, 80 per 60 degree sector). 
 This is not easy to increase.  The Pt min is 1.5 GeV/c.

Mapping raw hits to physical space is easy as is getting the rough z 
position.

I do not know how fast we can compute a chi square.  It depends on the         
details of the DSP.         

4. I have talked to Meena a few weeks ago about algorithms for z_vertex finding.
What she wants to do is a variant of the hough transform which is well described
in the 
literature.  I gave her copies of several papers.