Identification of objects with Uniform Resource Identifier (URI): recommendations for application in plant phenotyping.

Phenotyping experiments are made in various installations from greenhouses to lean field. These different experiments tend to answer agricultural challenges such as food security. The plant phenotyping handles a multitude of different objects, from biological and genetical material to weather data by phenotypic traits measures. This multitude of objects from their diversity but also from their large number need to be handled properly (each installation is generating up to 10 terabytes of photos in a month for one experiment). Identifying these objects is more and more important in the scope of shared and open data: the aim is  to correctly understand the mechanisms at work behind the observations, making the links between events and effects.
In order to fill their purpose, identifiers must have some properties, namely non-ambiguousness, unicity, persistence, resolvability, stability.

These properties will ensure that the identifier will stay active long-enough and allow the data associated to it to be used in a different domain where it was originally created.

In this document we present rules and good practices for object identification in the plant phenomics research domain and propose an adaptation of the rules designed by computer and web scientists. Some key characteristics of this schema are i) the ability to use semantics within the identifier and ii) making it resolvable through browser. Semantics is a desired quality for the technical staff of phenotyping platforms to manipulate the identifiers. However, it is also a big issue when designing a persistent identifier (e.g. usable in more than twenty years) and an important drawback if people put too much semantics in it. The first recommendation is to use only the minimum information in identifiers, and the date of creation is a good one.