HEPDATA wishlist

HEPDATA is seen as crucial for long-term data curation, and for ease of access, for example for Rivet.

Several desirable extensions were discussed, some of which are essential.

Data types

There is a need to end the assumption that all data are in the form of histograms (with x/y axes). Non-histogram data which needs to be stored include:

• cross sections w/o artificial bin, co
• correlation/error matrices
• +/- excursions for multiple systematics (this should be output for rivet, and is the preferred format for storing/expressing correlations)
• link to Rivet analysis
• link to paper and other documentation (arXiv/inpsire)

Search facility

Needs upgrading, better keywords, more user-friendly.

Data entry

Formats should be specified, and auto-entry enabled with the possibility of test uploads and authorisation by experimental contacts.

Export to & import by Rivet

• plot titles and captions
• data year, experiment, luminosity

Rivet to add these to plots. Rivet should also be able to get the MC name and versions from HepMC and add these to legend.

Auxilliary data

The possibility of adding more data sets to a main data set could be useful. Use case examples include backgrounds, theory curves, MC curves compared to the data in the publication.

There should also be recognition and encouragement for experiments to upload very large numbers of plots if they are available at publication time but not in the publication (for example with slightly different cuts, different jet algorithms etc)