Present : M. Aiba, E. Benedetto, O. Berrig, H. Burkhardt, R. Calaga, A. Franchi, M. Giovannozzi, W. Herr, JB.Jeanneret, J. Jowett, M. Meddahi, E. Metral, JL.Nougaret, S. Redaelli, F. Roncarolo, T. Risselada, F. Schmidt, F. Tecker, R. Tomas (list not complete)

1) MAD-X Version 4 & Plan (FS) (slides in MADX_27.03.2009.pdf.)
Version MADX 4.00.09 has now become the default version. A large effort was made to debug the program core. The treatment of LHC Beam 2 had to be modified (see below). The offset files used by the Aperture module are now read in TFS format. PTC is integrated inside the new madx.

The previous production version MADX 3.04.53 (November 2008) will be frozen (see next point) and may be accessed as "madx_old". Equally, after the introduction of version 4.00.09 the development of new production versions will be slowed down as the program must be as stable as possible during the LHC commissioning phase.

It is also worth mentioning that there is a change concerning "Offsetelem file syntax" which is documented on the MAD-X webpage in the "Aperture, Geometric" section.

2) LHC Beam 2 Model (TR) (slides in Meeting-22/LHC_beam_2_model.ppt).
The treatment of the LHC beam 2 in MADX contained too many bugs and it was finally decided to eliminate the individual BV attribute from the element and class definitions. From now on the same sign convention will be applied to all elements of beam 2. On the other hand, the effect of the global BV attribute in the Beam command is unchanged.

The new version is therefore not compatible with the previous versions. To avoid mismatches between program and machine versions MADX will no longer accept element or class definitions with "bv" attributes, which now generate a fatal error. New sequences have been installed in the LHC AFS database for the most recent LHC versions, and this change will thus be transparent to the users. The old sequences have been hidden in a special directory, and are still available to be used with the "madx_old" program version.

There should be no real need to use the individual BV flag for other machines at CERN and in case it is this should be taken care of.

It was mentioned after the meeting that the BV flag in MAD-X might be used for instance for RHIC at BNL. Colleagues from BNL have been notified.

IMPORTANT!!! It should be mentioned that the code but also the LHC input files have been modified according to the "book" as outlind in: beam2-beam4.doc

3) MADX, PTC and CLIC (JBJ) (slides in madx_ptc_clic.pdf).
The calculation of chromatic quantities is often incorrect in the madx Twiss, particularly in the case of betatron coupling, but also in other cases. Making this completely correct would require a too large effort. The new madx estimates the deltap/p derivatives by performing a second Twiss with a slightly different momentum value if the keyword "chrom" is present in the Twiss (or Match) command.

In cases where this estimation is not sufficiently accurate PTC has to be used. Particularly for small machines (speed being less critical) this solution is recommended.

4) PTC in the TT2-TT10 Transfer Line (EB) (slides in .pdf).
The presence of strong skew quadrupoles in beam lines or rings makes it very difficult to calculate (and interpret) the lattice functions.

In the PS-SPS transfer line the transverse emittances of the SPS fixed target beam are completely exchanged between the two planes. The Twiss calculation of madx shows negative beta functions for this case. Here again PTC has to be used.