about compute msd

Dear all,

Following the (not so) recent replacement of "fix msd" by "compute
msd", I wished to confirm that the 2 lines below are equivalent to
what did "fix 1 all msd 100 msd.foo":

compute 1 all msd
fix 1 all ave/time 100 1 100 c_1 file msd.foo mode vector start 0

In this case, it would mean that a lot of information is lost as the
msd for a given "Delta t" is computed once and for all. Would it be
possible, in the spirit of what is done in fix ave/correlate, to
average for each "Delta t" over every possible initial times during
the simulation?


Are you saying the old fix msd did something different?
I don't think so. It is not feasible to compute MSDs
on-the-fly from many starting points, b/c it would require storing
many snapshots of old atom coords. It would be
easy for the user to specify intervals that blew out