(updated 2023-06-28) A few days ago, we discussed the new dataset publications in dblp. As a preparation for more and more detailed datasets we slightly modify the DTD that defines the structure of our XML data export.
A quick reminder: you can download the dblp dataset as a single XML file. For more details please see our FAQ page.
All modifications are additions or slight changes of data type. They should not affect most data imports. The new DTD can be used for older releases of the XML file. We will not add the new elements/attributes before May 29, 2023.
All changes can be seen in our change log. If you have any questions please contact us.
- Added <stream> element to publication types.
<stream> will contain a reference to a stream, e.g., a specific journal or conference series in which the publication appeared.
This is a supplement to the information contained in the <booktitle>, <journal> and <url> elements in more structured form.
- Added <rel> element to publication types.
<rel> will contain the key of a different dblp publication, indicating that a publication is in some kind of relation to the referenced publication.
Example: the relations of different versions of a data set to the abstract concept of the data set itself or the relation of a preprint to the peer reviewed article.
Optional attribute type can contain specifics of that relation.
- Type change: all publtypes and type attributes are now NMTOKENS instead of CDATA.
All publtype and type values conformed to this before, but we now make it a requirement.
Update on June 28, 2023
We did a further late change to the DTD:
- Added a sort attribute to <rel> elements.
This attribute contains a numeric value which gives a relative ordering (ascending) between <rel> elements in the same context (e.g., versions pointing to the same parent entity). Note that the value of the attribute does not have any concrete meaning, only the ordering is relevant. For a meaningful designator see the label attribute of the <rel> element.
Example: Record A contains
<rel type="versionOf" uri="..." sort="3" label="1.1"></rel>
Record B contains
<rel type="versionOf" uri="..." sort="1" label="1.0"></rel>
This indicates that Record A (sort=”3″) is sorted after Record B (record=”1″).