For a description of OAI-PMH see here.
We issue a "ListRecords" request in marcxml format and obtain a "file" with a list of records. For our purposes, one of these records has a structure similar to the following:
|<marc:datafield tag="773" ind1=" " ind2=" ">||wrapper for the following fields|
|<marc:subfield code="p">Phys.Rev.Lett.</marc:subfield>||journal name||Phys.Rev.Lett.|
|<marc:subfield code="y">1973</marc:subfield>||journal year||1973|
|<marc:subfield code="v">31</marc:subfield>||journal volume||31|
For a more detailed description of the structure, please see here.
From these fields, we know that a paper (identifier) has been published by "journal" in "year". The "volume" is also important because, for historical reasons, certain journals have a part of their name contained in the "volume" field (e.g. "Phys.Rev.D").
The code that we use to query the OAI interface, to download the XML data and to parse it can be provided upon request.