RFC 9290 | CoRE Problem Details | August 2022 |
Fossati & Bormann | Standards Track | [Page] |
This document defines a concise "problem detail" as a way to carry machine-readable details of errors in a Representational State Transfer (REST) response to avoid the need to define new error response formats for REST APIs for constrained environments. The format is inspired by, but intended to be more concise than, the problem details for HTTP APIs defined in RFC 7807.¶
This is an Internet Standards Track document.¶
This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Further information on Internet Standards is available in Section 2 of RFC 7841.¶
Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at https://www.rfc-editor.org/info/rfc9290.¶
Copyright (c) 2022 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
REST response status information such as Constrained Application Protocol (CoAP) response codes (Section 5.9 of [RFC7252]) is sometimes not sufficient to convey enough information about an error to be helpful. This specification defines a simple and extensible framework to define Concise Binary Object Representation (CBOR) [STD94] data items to suit this purpose. This framework is designed to be reused by REST APIs, which can identify distinct "shapes" of these data items specific to their needs. Thus, API clients can be informed of both the high-level error class (using the response code) and the finer-grained details of the problem (using the vocabulary defined here). This pattern of communication is illustrated in Figure 1.¶
The framework presented is largely inspired by the problem details for HTTP APIs defined in [RFC7807]. Appendix B discusses applications where interworking with [RFC7807] is required.¶
The terminology from [RFC7252], [STD94], and [RFC8610] applies; in particular, CBOR diagnostic notation is defined in Section 8 of RFC 8949 [STD94] and Appendix G of [RFC8610]. Readers are also expected to be familiar with the terminology from [RFC7807].¶
In this document, the structure of data is specified in Concise Data Definition Language (CDDL) [RFC8610] [RFC9165].¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
A Concise Problem Details data item is a CBOR data item with the following
structure (rules named starting with tag38
are defined in Appendix A):¶
(Examples of elaborated Concise Problem Details data items can be found later in the document, e.g., Figure 3.)¶
A number of problem detail entries, the Standard Problem Detail entries, are predefined (more predefined details can be registered, see Section 3.1).¶
Note that, unlike [RFC7807], Concise Problem Details data items have no explicit "problem type". Instead, the category (or, one could say, Gestalt) of the problem can be understood from the shape of the problem details offered. We talk of a "problem shape" for short.¶
A short, human-readable summary of the problem shape. Beyond the shape of the problem, it is not intended to summarize all the specific information given with the problem details. For instance, the summary might include that an account does not have enough money for a transaction to succeed but not the detailed information such as the account number, how much money that account has, and how much would be needed.¶
A human-readable explanation specific to this occurrence of the problem.¶
A URI reference that identifies the specific occurrence of the problem. It may or may not yield further information if dereferenced.¶
The CoAP response code (Sections 5.9 and 12.1.2 of [RFC7252]) generated by the origin server for this occurrence of the problem.¶
The base URI (see Section 5.1 of RFC 3986 [STD66]) that should be used to resolve relative URI references embedded in this Concise Problem Details data item.¶
The language-tag (tag38-ltag) that applies to the presentation of unadorned text strings (not using tag 38) in this Concise Problem Details data item; see Appendix A.¶
The writing-direction (tag38-direction) that applies to the presentation of unadorned text strings (not using tag 38) in this Concise Problem Details data item; see Appendix A.¶
Both "title" and "detail" can use either an unadorned CBOR text string
(text
) or a language-tagged text string (tag38
); see Appendix A for
the definition of the latter.
Language tag and writing direction information for unadorned text
strings is intended to be obtained from context; if that context
needs to be saved or forwarded with a Concise Problem Details data
item, "base-lang" and "base-rtl" can be used.
If no such (explicitly saved or implicit) context information is
available, unadorned text is interpreted with language-tag "en" and
writing-direction "false" (ltr).¶
The "title" string is advisory and included to give consumers a shorthand for the category (problem shape) of the error encountered.¶
The "detail" member, if present, ought to focus on helping the client correct the problem rather than giving extensive server-side debugging information. Consumers SHOULD NOT parse the "detail" member for information; extensions (see Section 3) are more suitable and less error-prone ways to obtain such information. Note that the "instance" URI reference may be relative; this means that it must be resolved relative to the representation's base URI, as per Section 5 of RFC 3986 [STD66].¶
The "response-code" member, if present, is only advisory; it conveys
the CoAP response code used for the convenience of the consumer.
Generators MUST use the same response code here as in the actual CoAP
response; the latter is needed to assure that generic CoAP software that
does not understand the problem-details format still behaves
correctly.
Consumers can use the "response-code" member to determine what the
original response code used by the generator was, in cases where it
has been changed (e.g., by an intermediary or cache), and when message
bodies persist without CoAP information (e.g., in an events log or analytics
database).
Generic CoAP software will still use the CoAP response code.
To support the use case of message-body persistence without support by
the problem-details generator, the entity that persists the Concise
Problem Details data item can copy over the CoAP response code
that it received on the CoAP level.
Note that the "response-code" value is a numeric representation of the
actual code (see Section 3 of [RFC7252]), so it does not take the usual
presentation form that resembles an
HTTP status code: 4.04 Not Found
is represented by the number 132.¶
The "base-uri" member is usually not present in the initial request-response communication as it can be inferred as per Section 5.1.3 of RFC 3986 [STD66]. An entity that stores a Concise Problem Details data item or otherwise makes it available for consumers without this context might add in a "base-uri" member to allow those consumers to perform resolution of any relative URI references embedded in the data item.¶
This specification defines a generic problem-details container with only a minimal set of attributes to make it usable.¶
It is expected that applications will extend the base format by defining new attributes.¶
These new attributes fall into two categories: generic and application specific.¶
Generic attributes will be allocated in the standard-problem-detail-entries
slot according to the registration procedure defined in Section 3.1.¶
Application-specific attributes will be allocated in the
custom-problem-detail-entries
slot according to the procedure described in
Section 3.2.¶
Consumers of a Concise Problem Details data item MUST ignore any Standard Problem Detail entries or Custom Problem Detail entries, or keys inside the Custom Problem Detail entries, that they do not recognize ("ignore-unknown rule"); this allows problem details to evolve. When storing the data item for future use or forwarding it to other consumers, it is strongly RECOMMENDED to retain the unrecognized entries; exceptions might be when storage or forwarding occurs in a different format/protocol that cannot accommodate them or when the storage or forwarding function needs to filter out privacy-sensitive information and for that needs to assume unrecognized entries might be privacy-sensitive.¶
Beyond the Standard Problem Detail keys defined in Figure 2, additional
Standard Problem Detail keys can be registered for use in the
standard-problem-detail-entries
slot (see Section 6.1).¶
Standard Problem Detail keys are negative integers, so they can never conflict with Custom Problem Detail keys defined for a specific application domain (which are unsigned integers or URIs.)¶
In summary, the keys for Standard Problem Detail entries are in a global namespace that is not specific to a particular application domain.¶
Section 2 provides a number of generally applicable Standard Problem Detail entries. The present section both registers another useful Standard Problem Detail entry and serves as an example of a Standard Problem Detail Entry registration, in the registration template format that would be ready for registration.¶
-8¶
unprocessed-coap-option¶
one-or-more<uint>
, where¶
one-or-more<T> = T / [ 2* T ]¶
Option number(s) of CoAP option(s) that were not understood¶
Section 3.1.1 of RFC 9290¶
The specification of the Standard Problem Detail entry referenced by the above registration template follows:¶
The Standard Problem Detail entry unprocessed-coap-option
provides
the option number or numbers of any CoAP options present in the request that
could not be processed by the server.¶
This may be a critical option that the server is unaware of, or an option the server is aware of but could not process (and chose not to, or was not allowed to, ignore it).¶
The Concise Problem Details data item including this Standard Problem Detail Entry can be used in fulfillment of the "SHOULD" requirement in Section 5.4.1 of [RFC7252].¶
Several option numbers may be given in a list (in no particular order), without any guarantee that the list is a complete representation of all the problems in the request (as the server might have stopped processing already at one of the problematic options). If an option with the given number was repeated, there is no indication which of the values caused the error.¶
Clients need to expect to see options in the list that they did not send in the request; this can happen if the request traversed a proxy that added the option but did not act on the problem-details response being returned by the origin server.¶
For a few special values of unprocessed CoAP options (such as Accept or Proxy-Uri), note that there are special response codes (4.06 Not Acceptable, 5.05 Proxying Not Supported, respectively) to be sent instead of 4.02 Bad Option.¶
Applications may extend the Concise Problem Details data item with additional entries to convey additional, application-specific information.¶
Such new entries are allocated in the custom-problem-detail-entries
slot and
carry a nested map specific to that application. The map key can be either
an (absolute!) URI (under control of the entity defining this extension)
or an unsigned integer.
Only the latter needs to be registered (Section 6.2).¶
Within the nested map, any number of attributes can be given for a single extension. The semantics of each custom attribute MUST be described in the documentation for the extension; for extensions that are registered (i.e., are identified by an unsigned int), that documentation goes along with the registration.¶
The unsigned integer form allows a more compact representation. In exchange, authors are expected to comply with the required registration and documentation process. In comparison, the URI form is less space efficient but requires no registration. Therefore, it is useful for experimenting during the development cycle and for applications deployed in environments where producers and consumers of Concise Problem Details are more tightly integrated. (Thus, the URI form covers the potential need we might otherwise have for a "Private Use" range for the unsigned integers.)¶
Note that the URI given for the extension is for identification purposes only and, even if dereferenceable in principle, it MUST NOT be dereferenced in the normal course of handling problem details (i.e., outside diagnostic or debugging procedures involving humans).¶
Figure 3 shows an example (in CBOR diagnostic notation)
of a custom extension using a (made-up) URI as the custom-problem-detail-entries
key.¶
Obviously, a Standards Development Organization (SDO) like 3GPP can also easily register such a Custom
Problem Detail entry to receive a more efficient unsigned integer key;
Figure 4 shows how
the same example would look using a (made-up) registered unsigned int as the
custom-problem-detail-entries
key:¶
In summary, the keys for the maps used inside Custom Problem Detail entries are defined specifically for use with the identifier of that Custom Problem Detail entry, the documentation of which defines these internal entries, typically chosen to address a given application domain.¶
When there is a need to evolve a Custom Problem Detail entry definition, the "ignore-unknown rule" discussed in Section 3 provides an easy way to include additional information. The assumption is that this is done in a backward- and forward-compatible way. Sometimes, Custom Problem Detail entries may need to evolve in a way where forward compatibility by applying the "ignore-unknown rule" would not be appropriate: for example, when adding a "must-understand" member, which can only be ignored at the peril of misunderstanding the Concise Problem Details data item ("false interoperability"). In this case, a new Custom Problem Detail key can simply be registered for this case, keeping the old key backward and forward compatible.¶
Problem details may unintentionally disclose information. This can lead to both privacy and security problems. See Section 5 for more details that apply to both domains; particular attention needs to be given to unintentionally disclosing Personally Identifiable Information (PII).¶
Concise Problem Details can contain URIs that are not intended to be dereferenced (Section 3.2, Paragraph 5). One reason is that dereferencing these can lead to information disclosure (tracking). Information disclosure can also be caused by URIs in problem details that are intended for dereferencing, e.g., the "instance" URI. Implementations need to consider which component of a client should perform the dereferencing and which servers are trusted with serving them. In any case, the security considerations of Section 7 of RFC 3986 [STD66] apply.¶
The security and privacy considerations outlined in Section 5 of [RFC7807] apply in full. While these are phrased in terms of security considerations for new RFC 7807 problem types, they equally apply to the problem detail entry definitions used here (Section 3). In summary, both when defining new detail entries and when actually using them to generate a Concise Problem Details data item, care needs to be taken that they do not leak sensitive information. Entities storing or forwarding Concise Problem Details data items need to consider whether this leads to information being transferred out of the context within which access to sensitive information was acceptable. See also Section 3, Paragraph 6 (the last paragraph of the introduction to that section). Privacy-sensitive information in the problem details SHOULD NOT be obscured in ways that might lead to misclassification as non-sensitive (e.g., by base64-encoding).¶
This specification defines a new subregistry titled "Standard Problem Detail Keys" in the "Constrained RESTful Environments (CoRE) Parameters" registry [IANA.core-parameters], with "Specification Required" as the Registration Procedure (Section 4.6 of [RFC8126]).¶
Each entry in the registry must include:¶
a negative integer to be used as the value of the key¶
a name that could be used in implementations for the key¶
type of the data associated with the key in CDDL notation¶
a brief description¶
a reference document¶
(see Section 2.3 of [RFC8126])¶
The designated expert is requested to assign the shortest key values (1+0 and 1+1 encoding) to registrations that are likely to enjoy wide use and can benefit from short encodings.¶
To be immediately useful in CDDL and programming-language contexts, a
name consists of a lowercase ASCII letter (a-z) and zero or more
additional ASCII characters that are either lowercase letters,
digits, or a hyphen-minus, i.e., it matches [a-z][-a-z0-9]*
.
As with the key values, names need to be unique.¶
The specification in the reference document needs to provide a description of the Standard Problem Detail entry, replicating the CDDL description in "CDDL Type", and describing the semantics of the presence of this entry and the semantics of the value given with it.¶
Initial entries in this subregistry are as follows:¶
Key Value | Name | CDDL Type | Brief Description | Reference | Change Controller |
---|---|---|---|---|---|
-1 | title |
text / tag38
|
Short, human-readable summary of the problem shape | RFC 9290 | IETF |
-2 | detail |
text / tag38
|
Human-readable explanation specific to this occurrence of the problem | RFC 9290 | IETF |
-3 | instance |
~uri
|
URI reference identifying specific occurrence of the problem | RFC 9290 | IETF |
-4 | response-code |
uint .size 1
|
CoAP response code | RFC 9290 | IETF |
-5 | base-uri |
~uri
|
Base URI | RFC 9290 | IETF |
-6 | base-lang |
tag38-ltag
|
Base language tag (see Appendix A) | RFC 9290 | IETF |
-7 | base-rtl |
tag38-direction
|
Base writing direction (see Appendix A) | RFC 9290 | IETF |
-8 | unprocessed-coap-option |
one-or-more<uint>
|
Option number(s) of CoAP option(s) that were not understood | RFC 9290, Section 3.1.1 | IETF |
This specification defines a new subregistry titled "Custom Problem Detail Keys" in the "Constrained RESTful Environments (CoRE) Parameters" registry [IANA.core-parameters], with as "Expert Review" as the Registration Procedure (Section 4.5 of [RFC8126]).¶
The designated expert is instructed to attempt making the registration experience as close to First Come First Served as reasonably achievable, but checking that the reference document does provide a description as set out below. (This requirement is a relaxed version of "Specification Required" as defined in Section 4.6 of [RFC8126].)¶
Each entry in the registry must include:¶
an unsigned integer to be used as the value of the key¶
a name that could be used in implementations for the key¶
a brief description¶
a reference document that provides a description of the map, including a CDDL description, that describes all inside keys and values¶
(see Section 2.3 of [RFC8126])¶
The designated expert is requested to assign the shortest key values (1+0 and 1+1 encoding) to registrations that are likely to enjoy wide use and can benefit from short encodings.¶
To be immediately useful in CDDL and programming-language contexts, a
name consists of a lowercase ASCII letter (a-z) and zero or more
additional ASCII characters that are either lowercase letters,
digits, or a hyphen-minus, i.e., it matches [a-z][-a-z0-9]*
.
As with the key values, names need to be unique.¶
Initial entries in this subregistry are as follows:¶
Key Value | Name | Brief Description | Reference | Change Controller |
---|---|---|---|---|
7807 | tunnel-7807 | Carry RFC 7807 problem details in a Concise Problem Details data item | RFC 9290, Appendix B | IETF |
IANA has added the following media type to the "Media Types" registry [IANA.media-types].¶
Name | Template | Reference |
---|---|---|
concise-problem-details+cbor | application/concise-problem-details+cbor | RFC 9290, Section 6.3 |
application¶
concise-problem-details+cbor¶
N/A¶
N/A¶
binary (CBOR data item)¶
none¶
Section 6.3 of RFC 9290¶
Clients and servers in the Internet of Things¶
The syntax and semantics of fragment identifiers is as specified for "application/cbor". (At publication of RFC 9290, there is no fragment identification syntax defined for "application/cbor".)¶
CoRE WG mailing list (core@ietf.org) or IETF Applications and Real-Time Area (art@ietf.org)¶
COMMON¶
none¶
IETF¶
no¶
IANA has registered a Content-Format number in the "CoAP Content-Formats" subregistry, within the "Constrained RESTful Environments (CoRE) Parameters" registry [IANA.core-parameters], as follows:¶
Media Type | Encoding | ID | Reference |
---|---|---|---|
application/concise-problem-details+cbor | - | 257 | RFC 9290 |
In the registry "CBOR Tags" [IANA.cbor-tags], IANA has registered CBOR tag 38. IANA has updated the reference for CBOR tag 38 to point to RFC 9290, Appendix A.¶
This appendix serves as the archival documentation for CBOR tag 38, a tag for serializing language-tagged text strings in CBOR. The text of this appendix is adapted from the specification text supplied for its initial registration. It has been extended to allow supplementing the language tag by a direction indication.¶
As with any IANA-registered item, a specification that further updates this registration needs to update the reference column of the IANA registry (see Section 6.5). Future specifications may update this appendix, other parts of this document, or both. (When updating this appendix, keep in mind that applications beyond Concise Problem Details data items may adopt the tag defined here.) Users of this tag are advised to consult the registry to obtain the most recent update for this appendix.¶
In some cases, it is useful to specify the natural language of a text string. This specification defines a tag that does just that. One technology that supports language-tagged strings is the Resource Description Framework (RDF) [RDF].¶
A language-tagged text string in CBOR has the tag 38 and consists of an array with a length of 2 or 3.¶
The first element is a well-formed language tag described in BCP 47 ([RFC5646] and [RFC4647]), represented as a UTF-8 text string (major type 3).¶
The second element is an arbitrary UTF-8 text string (major type 3). Both the language tag and the arbitrary string can optionally be annotated with CBOR tags; this is not shown in the CDDL below.¶
The optional third element, if present, represents a ternary value that indicates a direction, as follows:¶
false
: left-to-right direction ("ltr").
The text is expected to be displayed with left-to-right base
direction if standalone and isolated with left-to-right direction
(as if enclosed in LRI ... PDI or equivalent, see [Unicode-14.0.0-bidi]) in the
context of a longer string or text.¶
true
: right-to-left direction ("rtl").
The text is expected to be displayed with right-to-left base
direction if standalone and isolated with right-to-left direction
(as if enclosed in RLI ... PDI or equivalent, see [Unicode-14.0.0-bidi]) in the context
of a longer string or text.¶
null:
indicates that no indication is made about the direction
("auto"), enabling an internationalization library to make an auto-detection
decision such as treating the string as if enclosed in FSI ... PDI
or equivalent, see [Unicode-14.0.0-bidi].¶
If the third element is absent, directionality context may be applied
(e.g., base-directionality information for an entire CBOR message or
part thereof). If there is no directionality context applied, the
default interpretation is the same as for null
("auto").¶
In CDDL:¶
tag38 = #6.38([tag38-ltag, text, ?tag38-direction]) tag38-ltag = text .regexp "[a-zA-Z]{1,8}(-[a-zA-Z0-9]{1,8})*" tag38-direction = &(ltr: false, rtl: true, auto: null)¶
NOTE: Language tags of any combination of case are allowed. But Section 2.1.1 of [RFC5646], part of BCP 47, recommends a case combination for language tags that encoders that support tag 38 may wish to follow when generating language tags.¶
Data items with tag 38 that do not meet the criteria above are not valid (see Section 5.3.2 of RFC 8949 [STD94]).¶
NOTE: The Unicode Standard [Unicode-14.0.0] includes a set of characters designed for tagging text (including language tagging) in the range U+E0000 to U+E007F. Although many applications, including RDF, do not disallow these characters in text strings, the Unicode Consortium has deprecated these characters and recommends annotating language via a higher-level protocol instead. See the section "Deprecated Tag Characters" in Section 23.9 of [Unicode-14.0.0] as well as [RFC6082].¶
NOTE: while this document references a version of Unicode that was recent at the time of writing, the statements made based on this version are expected to remain valid for future versions.¶
Examples in this section are given in CBOR diagnostic notation first and then as a pretty-printed hexadecimal representation of the encoded item.¶
The following example shows how the English-language string "Hello" is represented.¶
38(["en", "Hello"])¶
D8 26 # tag(38) 82 # array(2) 62 # text(2) 656E # "en" 65 # text(5) 48656C6C6F # "Hello"¶
The following example shows how the French-language string "Bonjour" is represented.¶
38(["fr", "Bonjour"])¶
D8 26 # tag(38) 82 # array(2) 62 # text(2) 6672 # "fr" 67 # text(7) 426F6E6A6F7572 # "Bonjour"¶
The following example shows how the Hebrew-language string
is represented, where in the rtl
direction, each character is:
"ם" (HEBREW LETTER FINAL MEM, U+05DD),
"ו" (HEBREW LETTER VAV, U+05D5),
"ל" (HEBREW LETTER LAMED, U+05DC),
"ש" (HEBREW LETTER SHIN, U+05E9).
Note the rtl
direction expressed by setting the third element in the array to "true".¶
38(["he", "שלום", true])¶
D8 26 # tag(38) 83 # array(3) 62 # text(2) 6865 # "he" 68 # text(8) D7A9D79CD795D79D # "שלום" F5 # primitive(21)¶
On certain occasions, it will be necessary to carry ("tunnel") [RFC7807] problem details in a Concise Problem Details data item.¶
This appendix defines a Custom Problem Detail entry for that purpose. This is assigned Custom Problem Detail key 7807 in Section 6.2. Its structure is:¶
tunnel-7807 = { ? &(type: 0) => ~uri ? &(status: 1) => 0..999 * text => any }¶
To carry an [RFC7807] problem details JSON object in a Concise Problem Details data item, first convert the JSON object to CBOR as per Section 6.2 of RFC 8949 [STD94]. Create an empty Concise Problem Details data item.¶
Move the values for "title", "detail", and "instance", if present, from the [RFC7807] problem details to the equivalent Standard Problem Detail entries. Create a Custom Problem Detail entry with key 7807. Move the values for "type" and "status", if present, to the equivalent keys 0 and 1 of the Custom Problem Detail entry. Move all remaining key/value pairs (additional members as per Section 3.2 of [RFC7807]) in the converted [RFC7807] problem details object to the Custom Problem Detail map unchanged.¶
The inverse direction, carrying Concise Problem Details in an RFC 7807 problem details JSON object requires the additional support provided by [HTTPAPI], which is planned to create the HTTP Problem Types Registry. An HTTP Problem Type can then be registered that extracts top-level items from the Concise Problem Details data item in a similar way to the conversion described above and that carries the rest of the Concise Problem Details data item in an additional member via base64url encoding without padding (Section 5 of [RFC4648]). Details can be defined in a separate document when the work on [HTTPAPI] is completed.¶
The authors would like to thank Mark Nottingham and Erik Wilde, the authors of RFC 7807; Klaus Hartke and Jaime Jiménez, the coauthors of an earlier draft version of this specification; Christian Amsüss, Marco Tiloca, Ari Keränen, and Michael Richardson for review and comments on this document. Francesca Palombini for her review (and support) as responsible AD, and Joel Jaeggli for his OPSDIR review, both of which brought significant additional considerations to this document.¶
For Appendix A, John Cowan and Doug Ewell are also to be acknowledged. The content of an earlier draft version of this appendix was also discussed in the "apps-discuss@ietf.org" and "ltru@ietf.org" mailing lists. More recently, the authors initiated a discussion about the handling of writing direction information in conjunction with language tags. That led to discussions within the W3C Internationalization Core Working Group. The authors would like to acknowledge that cross-organization cooperation and particular contributions from John Klensin and Addison Phillips and specific text proposals by Martin Dürst.¶
Peter defined CBOR tag 38, basis of Appendix A.¶
Christian contributed what became Section 3.1.1.¶