The "Definition of the S-Coordinate for graphical plots along the LCLS beamline", is at http://www-ssrl.slac.stanford.edu/lcls/prd/1.1-103-r0.pdf
The SUML is the beam path length (as the "crow" flies), including all
curves and bumps along the way.
The "LINAC" Z-coordinate is a line along the linac axis, and parallel to
it.
The "UNDULATOR" or "LCLS" Z-coordinate (also called Z') is a line along
the undulator axis, and parallel to it.
Note the linac axis is pitched downhill about 0.3 degrees while the
undulator is not, so this is why we have two Z coordinates (linac-Z and
undulator-Z).
These coordinates are described in an LCLS tech note at:
http://www-ssrl.slac.stanford.edu/lcls/technotes/lcls-tn-03-8.pdf
The length of a device is included in the EXCEL MAD-output file...
http://www-ssrl.slac.stanford.edu/lcls/linac/optics/lcls.xls
...under column "F", called "L_EFF". For magnets, this is the magnetic
length, not the physical length.
Note that the Z-coordinate in this EXCEL file is labeled as "X Coor" at
the top of the column. I'm' sorry for this, but the mechanical
designers wanted to call it this to fit with their Solid-Edge program.
You can look at the bottom of the EXCEL column to see the "z" label
shown there.
The "LCLS" Z-coordinate is used in the "LTU" EXCEL file at:
http://www-ssrl.slac.stanford.edu/lcls/linac/optics/ltu.xls
The index page with links to these EXCEL files notes which coordinates
are used ("linac" or "LCLS"):
See http://www-ssrl.slac.stanford.edu/lcls/linac/optics/.
I know it's a bit much, but that's the story.