The "doi" URI Scheme

Introduction A DOI name is a global unique identifier of a referent, which can be any digital, physical or abstract entity, including inventions, literary and artistic works, ideas, symbols, names, images, designs, etc. DOI names are, for example, widely used to identify academic publications. The DOI system is specified in and , with the former offering regular formal snapshot of the latter. EXAMPLE 1: The DOI name "10.1103/PhysRevLett.59.381" refers to the article Per Bak, Chao Tang, and Kurt Wiesenfeld, "Self-organized criticality: An explanation of the 1/f noise", Phys. Rev. Lett. 59, 381. A DOI name is persistent over time. This persistence is provided by the independence of the DOI name from the referent itself and its descriptive elements. These descriptive elements of a referent, including location and ownership, can change over time, and their current values are retrieved by resolving the DOI name. The set of elements retrieved by resolving a DOI name is called the DOI record. The DOI name resolution process uses the Handle System specified at , and , as updated by . This document specifies a URI scheme for DOI names. This scheme formalizes the notation "doi:<DOI name>", which is in widespread use. When resolved, the URI corresponding to a DOI name yields the DOI record associated with the name. EXAMPLE 2: "doi:10.1103/PhysRevLett.59.381" is the URI corresponding to the DOI name above. This document intended to satisfy the guidelines and registration procedures specified at .

Syntax As specified at , a DOI name consists of an ordered sequence of Unicode code points of the Graphic type. The scheme-specific-part of the URI corresponding to a given DOI name, i.e., the DOI Name URI, SHALL be equal to the result of the following ordered sequence of steps:

express the ordered sequence of Unicode code points that comprise the DOI name as a UTF-8 String, as defined at , without the byte order mark and without any normalization;
percent-encode any byte in the UTF-8 String that is neither unreserved nor equal to "/".

The URI shall contain neither a query component nor a fragment component. EXAMPLE 1: The DOI name "10.5594/SMPTE.ST2067-21.2020" corresponds to the URI <doi:10.5594/SMPTE.ST2067-21.2020>. EXAMPLE 2: The DOI name "10.26321/Á.GUTIÉRREZ.ZARZA.02.2018.03" with the code point sequence <U+0031, U+0030, U+002E, U+0032, U+0036, U+0033, U+0032, U+0031, U+002F, U+00C1, U+002E, U+0047, U+0055, U+0054, U+0049, U+00C9, U+0052, U+0052, U+0045, U+005A, U+002E, U+005A, U+0041, U+0052, U+005A, U+0041, U+002E, U+0030, U+0032, U+002E, U+0032, U+0030, U+0031, U+0038, U+002E, U+0030, U+0033> corresponds to the URI <doi:10.26321/%C3%81.GUTI%C3%89RREZ.ZARZA.02.2018.03>. NOTE 1: The sequence of code points comprising a DOI name is not normalized and equivalence between DOI names is based on code points. For example, two DOI names that differ only in the abstract character "Á" being encoded as <U+00C1> in the first and as <U+0041, U+0301> in the second are not identical. NOTE 2: Presenting a DOI name by rendering its sequence of code points to glyphs can be ambiguous since multiple code points or sequences of code points can result in the same glyphs. For example, U+002D HYPHEN-MINUS, U+2212 MINUS SIGN and U+2013 EN DASH are rendered as similar glyphs. As another example, the abstract character "á" can be represented by either the code point U+00E1 or the sequence of code points <U+0061, U+0301>. Presenting a DOI name in its URI form resolves this ambiguity. NOTE 3: DOI names are case-insensitive only when testing for equivalence and only with respect to the Basic Latin Unicode block.

Resolution Resolving a DOI name means retrieving its DOI record, which contains the descriptive elements associated with the referent identified by the DOI name. A DOI name URI can be be resolved by performing an HTTP GET request at the following URL (expressed using ABNF syntax as defined at ): "https://doi.org/api/handles/" scheme-specific-part where scheme-specific-part is the scheme-specific-part of the DOI name URI, as defined at , and the "https" scheme is specified at . The body of the response is a JSON object, as defined at , that contains the following members:

responseCode

The property is a Number. The following values are defined:

1: The resolution completed successfully. The HTTP response status code is 200.
2: The resolution did not complete successfully because of a server error. The HTTP response status code is 500.
100: The DOI name was not found. The HTTP response status code is 404.
200: No descriptive elements were found for the requested DOI name. The HTTP response status code is 200.

handle

The property is a String. It is equal to the DOI name for which resolution was requested.

values

The property is an Object. It contains the descriptive elements for the referent identified by the DOI name. The contents of the property are specified at .

illustrates the DOI record, at the time of this writing, for the DOI name corresponding to the URI <doi:10.1000/182>. The DOI record was retrieved by performing an HTTP GET request to <https://doi.org/api/handles/10.1000/182>.

DOI record for the DOI name "10.1000/182" (at the time of this writing)

Security Considerations A DOI name is an opaque string, which does not have a discernible meaning on its own and is for use by humans and machines alike. It consists of a sequence of Unicode codepoints and the security considerations at apply. The DOI name resolution process is conducted using the Hypertext Transfer Protocol Secure, which ensures condifentiality and integrity of the transaction, and he security considerations at apply. The results of the DOI name resolution process is a JSON object and the security considerations at apply.