Get BibTeX info from DOI may be ambiguous

When I retrieve the bib data to the publication

Kloeden, Peter E. ; Kozyakin, Victor S. Uniform nonautonomous attractors under discretization. Partial differential equations and applications. Discrete Contin. Dyn. Syst. 10 (2004), no. 1-2, 423–433.

via MR Lookup, zbMath or directly from the article’s site by the link

I got correct bib data indicated above. I would like to mention that, for this publication, DOI=10.3934/dcds.2004.10.423

However, when in JabRef I am trying to get info for this publication using DOI (via menu BibTex -> New Entry by entering there DOI) I got the following record

Article (Kozyakin2003)
Kozyakin, V. & Kloeden, P.
Uniform nonautonomous attractors under discretization
Discrete and Continuous Dynamical Systems, American Institute of Mathematical Sciences (AIMS), 2003 , 10 , 423-433

where the YEAR is DIFFERENT !!!

It seems the problem is not due to JabRef, since the same year I got by getting bibliography data using other services for which the PRIMARY data format is RIS (Crossref, Mendeley, Zotero etc.) My opinion is that the “year” record in BibTeX is not the same as the “PY” record in RIS from which the year of publication is retrieved: the “year” in BibTex means the year of “official, paper” publication of the article, while “PY” in RIS reflects “first public appearance” of the article which may be different in the case when a journal admits “first online” publications.

In any case it would be useful to make somewhere in JabRef announcement that retrieving bib data by DOI might be ambiguous.

We use CrossRef to get the bibliographic metadata for a given DOI. In your case, CrossRef indeed claims that the article is published in 2003:
As we rely on the correctness of the data provided by crossref (or mathscinet etc), there is nothing we can do. You might contact CrossRef and tell them about this discrepancy.

Dear Tobiaz,

thank you for confirming my doubts about the use of Internet services in JabRef for obtaining bibliographic metadata. The prospect of at least occasionally receiving incorrect data substantially nullifies all the advantages of such services, no matter who is to blame for the erroneousness of these data.

When you enter a new entry by doi we use the official service to get the bibtex data.
For any existing entry, when you do lookup metadata /bibtexda you will see a diff/merge dialog which shows the current values and the retrieved values, so it allows you to compare the entries.

As @tobiasdiez already mentioned, we can do nothing about the correctness of the data.

Sorry, Christoph,

but namely entering NEW entry by doi I am getting wrong year in the record

author = {Victor Kozyakin and P.E. Kloeden},
title = {Uniform nonautonomous attractors under discretization},
journal = {Discrete and Continuous Dynamical Systems},
year = {2003},
volume = {10},
number = {1-2},
pages = {423–433},
month = {oct},
doi = {10.3934/dcds.2004.10.423},
publisher = {American Institute of Mathematical Sciences ({AIMS})},

And, at this stage I am getting NO warnings or other HINTS that the returned data may be incorrect.