Why does a proceedings year sometimes differ from the year of the conference?

In the context of a paper published in a conference's proceedings, a dblp record's ‹year› field has a different meaning than in all other contexts:

Hence, if the proceedings volume of a conference is not published in the same year as the conference is hosted, this will lead to a discrepancy between the ‹year› entries of the inproceedings and the proceedings records. The following example shows the situation for a typical post-proceedings volume:

<inproceedings key="conf/naa/Xiang08">
<title>Numerical Quadrature for Bessel Transformations ...</title>
<year>2008</year>
<booktitle>NAA</booktitle>
<crossref>conf/naa/2008</crossref>
...
</inproceedings>

<proceedings key="conf/naa/2008">
<title>Numerical Analysis and Its Applications 2008, ...</title>
<year>2009</year>
<booktitle>NAA</booktitle>
...
</proceedings>

We retain this (maybe unexpected) practice since a paper by an author named J. Smith that has been presented at a conference in the year, say, 2000, is commonly referred to as the "Smith, 2000" paper, regardless of when the actual post-proceedings volume appeared as a published product. A paper's year of publication can still be retrieved from the ‹year› field of the referenced (see ‹crossref›) proceedings record.

maintained by Schloss Dagstuhl LZI at University of Trier