The most important argument has been given by Attila Krasznahorka as an
answer to question asked on
hn-atlas-pathelp@cernNOSPAMPLEASE.ch (ALFA data xAOD reprocessing):
(...)
All in all, any run 1 analysis that you actually want to
publish, has to happen using D3PDs. The xAODs with Run 1 data are only
meant as a technical exercise. To get ourselves ready for run 2. No
public result will ever be allowed to be made from these
datasets.
(...)
Part of the run1 data is already available in
D3PD format. There is
support form minimum bias
D3PD maker team to produce more data. This is
the easiest and fastest way to start ALFA diffractive analysis.