Items of IMU software development that remain to be completed:
- Test BmuCosmic requiring a seed. DONE 2001.04.17
- Create test data for all of the roads used by BmuCosmic, and
test them. These are stored in the code itself, by the way. There are
about 34 different cases: 34 west have been tested, most satisfactorily.
The critical 0,0,0,0 cases are OK.
Most of the east side is similar, but there are points of difference.
A lot of pencil and paper and plot work.
- Decide on and implement an output bank for BmuCosmic. This
involves translating to standard chi-squared, etc.
MuonStub has no provision for the t0 info, and isn't likely to be useful.
2-3 days? Maybe no bank, just write to log file as currently?
- Check if we can use calibration runs to get at least ASD by ASD
relative t0's for the wires. BOB SAYS THE PULSE
LOOKS TOO CRUMMY TO USE. HE IS RIGHT. THE RMS ON THE T0 IS TOO
LARGE TO BE USEFUL. HALTING 2001.04.24
Well, maybe not now that the CCU's aren't oscillating. Assigned Piyabut
to look at this.
- Empty the imu directory, putting code in Muon and MuonMods and
perhaps MuonGeometry as well. 1 day prep, 1 day tests
- Test seed-requirements in IMU_Stub. In progress... About ready to
run tests w/ and w/out. Stalled due to higher priority projects. Will
have to restart.
- Change BmuManipulation.cc to allow a cell 0/3 with no jumper to call
itself a seed. May need to fiddle with IMU_Hit as well.
- Integrate 2-hit stub finding in IMU_Stub. Requires a different
algorithm (not hard) and tunable matching criteria and conditions
under which to accept a 2-hit stub.
- Run 2-hit stub studies on data.
- Why is IMU not included in the new simulation? Check Lena's work
and try to integrate the new IMU geometry. The geometry is now
in place and functional. 2001.06.14
Next step, integrating the digitization.
- Fix bug in digitization; use own IMU random number stream rather than
- Fix overlap with CMX. Needs new volumes.
- Compare new geometry output with old FORTRAN output. Understand the
differences (there will be some).
- Revamp IMU_Stub to take parameters, and do away with IMU_Control and
- Make sure IMU_Candidate can be called in such a way as to refit
itself, with appropriate parameters. Determine which fit parameters we
care to support (theta and vertex are two).
- Compare IMU_Stub with MuonStubFinder. Must wait until the latter
is extended to Bmu.
- Assemble the database entries for all the stack alignments. Use
the text auxilliary database facility for this.
We need info about how to create tables, fill from
the text file, and read info from tables. Bloom doesn't want us to
proceed too far until the alignment database decisions are final.
- Assemble database entries for each wire's t0.
- Test all the database accesses.
- Load the real database with all the above, and test retrieval.
Check retrieval speed. Is this in memory, or do we have to do a fetch
- Create a set of validation plots.
- Create a validation scheme for testing stub-finding. Create a small
data file and a set of expected answers.
- Review IMU_Hit. Is this getting a bit over-the-top? (Default is to
leave it as is.)
- Make sure event display has right geometry. It has been using
the old FORTRAN version.
- Review geometry in event display. What form of hits are appropriate? The
current picture is sort-of-OK.
- Study timing cuts for BSU from the data.
DONE 2001.08.06 The answer is 165-210 nsec, though this includes the
tracks back-scattered from the quads.
- We need a drift model based on data from the hall. This will
likely include an angle correction, and may include a near-the-wire
correction as well.
- Create calibration validation plots for BSU and TSU. Maybe somebody
else will do this based on the more central work.
- Use real run and event numbers. Eg: event->Number() in IMUStubModule. DONE 2001.04.25 They aren't needed, so ignore them.
- Go through code and use ErrorLogger to flag errors rather than
cout and #ifdef. DONE. 2001.05.03
- Check BSU location and orientation in simulation. Oddity needs to
be followed up on. Abandoned with
advent of new simulation. 2001.06.08
- Run a simulation with COT/SVX and MuonLinkerModule and see if the
FakeEv muon tracks are linked. DONE.
It works. 2001.05.02
- Further question on linking. Is the inefficiency (26%) due to the
COT not tracking well in this region, or pilot error, or to a
mismatch in the simulation/reconstruction code?
- Review crosstalk. Is it actually suppressed? Bob and I find
different rates. In particular, look at /cdf/data10/CA3229.0/1/2/3 and
/cdf/data10/CA3230.0/1/2, using (myron!=0&&event->Number()&3==1)
- Fix print bug in IMU_Candidate that ignores the side (-999).
- Check that BSU's are planted appropriately in new sim.
- Fix the YMON plots.
- Make the ImuHistos plots standard rather than root-based. Not a
high priority--it works, and isn't of general interest.
- Add the other half-cell tracking to ImuHistos. Important for
studying the scintillator efficiencies, esp at boundaries.
if you have trouble accessing the information on this page.