[urn] Re: Request for registration of formal URN namespace "dnp"
Director - Stakeholder Engagement <[email protected]>
| Newsgroups | gmane.ietf.urn |
|---|---|
| Message-ID | <DX1P273MB0903F36B3960A47DCFCD84DDC1DE2@DX1P273MB0903.AREP273.PROD.OUTLOOK.COM> |
draft-sohail-urn-dnp-01 is now posted at https://datatracker.ietf.org/doc/draft-sohail-urn-dnp/, incorporating the correction to the "class" production. The revised registration template follows and supersedes the one sent on 5 August. ----------------------- URN NAMESPACE REGISTRATION TEMPLATE Formal Namespace Identifier: dnp Submitted under Section 6.2 of RFC 8141 (Expert Review) Revised 10 August 2026. This version superseds the template submitted on 5 August 2026. The only substantive change is to the "class" production in the Syntax field, correcting the inconsistency with the URN-equivalence rule identified by Dale R. Worley on the [email protected] list. This is the text PDA asks IANA to register. Namespace Identifier: dnp Version: 1 Date: 2026-08-10 Registrant: Pakistan Digital Authority (PDA) Government of Pakistan 4th Floor, 5-A Constitution Avenue, Sector F-5/1 Islamabad, Pakistan Designated contact: Hira Sohail Director of Partnerships & Stakeholder Engagement Email: [email protected] Telephone: +92 302 516 1211 Web: https://pda.gov.pk/ PDA is a federal statutory body established under the Digital Nation Pakistan Act, 2025, with statutory responsibility for national digital standards. PDA is not requesting the fast-track registration procedure of Section 6.3 of RFC 8141; this registration is submitted under the Expert Review procedure of Section 6.2. Purpose: Names in this namespace identify normative and authoritative instruments published under the Digital Nation Pakistan programme: national digital policies, frameworks, technical standards, reference architectures, specifications, implementation guidelines, data schemas, application programming interface contracts, public registries, published datasets, the National Digital Masterplan and its sectoral plans, and formal directives issued by PDA. The primary community of use is the Government of Pakistan -- federal ministries and divisions, provincial governments and their attached departments and autonomous bodies -- together with the private-sector entities that build on or must conform to Pakistan's digital public infrastructure. Secondary communities include international standards bodies, multilateral development institutions with programmes in Pakistan, comparative policy researchers, the archival and library sector, and implementers in other jurisdictions who reference Pakistani specifications when designing interoperable systems. The Internet community at large benefits in three respects. First, the corpus is published openly and is of direct interest to the growing body of practice on digital public infrastructure, in which Pakistan is both an adopter and a contributor; stable citation makes that corpus usable in the scholarly and technical record. Second, Pakistani specifications describe interfaces with which non-Pakistani systems interoperate, so unambiguous identification of a specific edition of a specification has engineering value beyond Pakistan. Third, government publications are among the most frequent victims of reference decay; a managed namespace with an explicit non-reassignment commitment reduces that decay for a corpus that would otherwise be cited by ordinary HTTP URI alone. This namespace complements rather than duplicates existing namespaces. The "lex" namespace (RFC 9676) identifies legal norms and is the appropriate identifier system for Pakistani legislation, which this namespace does not cover. The "iso" and "ietf" namespaces identify the outputs of those bodies; where a DNP instrument profiles or adopts such an output it cites it in the native namespace rather than reassigning it. Where an instrument also bears a Digital Object Identifier or, in future, a Pakistani National Bibliography Number, those assignments coexist with the URN and are recorded as equivalences in the Register. Software that can make use of these names includes citation managers and reference-checking tools; document management and records systems within government; conformance and procurement tooling that must assert which edition of a standard a system was built against; long-term digital preservation systems in the national archival and library sector; and standards-registry software operated by PDA. Resolution services are available and are described under Resolution below. Neither this namespace nor its definition is expected to become a constituent part of a standard developed in the IETF. The namespace is expected to be referenced normatively by standards issued by PDA itself. Syntax: Names conform to the URN syntax of Section 2 of RFC 8141. The Namespace Specific String is further constrained as follows, using the ABNF of RFC 5234. ALPHA and DIGIT are imported from RFC 5234. NSS = class ":" local-id [ ":" edition ] class = 2*24( ALPHA ) local-id = id-start *62( id-char ) id-end id-start = ALPHA / DIGIT id-char = ALPHA / DIGIT / "-" / "." id-end = ALPHA / DIGIT edition = ver-form / date-form ver-form = "v" 1*3DIGIT [ "." 1*3DIGIT ] date-form = 4DIGIT "-" 2DIGIT "-" 2DIGIT The grammar above defines the set of well-formed names. It is wider than the set of names PDA assigns: the assignment practice for local-id and the canonical form given under URN-equivalence below are narrower, and a name presented in a form the grammar admits but PDA would not have assigned is still well-formed and still URN-equivalent to the assigned name. The NSS contains at most two colon characters and consists of exactly two or three fields. No structure is implied by the colon beyond that field division. class states the kind of instrument. The initially registered classes are: policy, framework, standard, architecture, specification, guideline, directive, schema, api, registry, dataset, masterplan, publication. The set is closed at any given time and is extended only by amendment to the PDA specification cited under Documentation. A registered class is never withdrawn or redefined; a class no longer used for new assignments is marked closed and the names already assigned within it remain valid. local-id carries the identifier already assigned to the instrument under PDA's document nomenclature framework, transformed to lower case, or a mnemonic assigned by PDA where no such identifier exists. It is unique within its class and is never reassigned. It does not encode subject matter, issuing unit, date or status, and consumers must not parse it for such information. The optional edition field distinguishes two things that are deliberately identified separately. A name without an edition field identifies the instrument as a continuing work -- the standard or policy as an institutional object persisting across revisions -- and is appropriate where the citing party means "this instrument, whichever edition is current". A name with an edition field identifies one fixed, immutable published state, and is appropriate where the citing party means "this exact text". Both forms are permanent; neither is ever reassigned; the work-level name survives withdrawal of the instrument from current effect, withdrawal being a status change rather than a change of identity. The two forms are distinct names for distinct resources and are not URN-equivalent to one another. The ver-form of edition is used where the instrument carries an editorial version number; the date-form, a Gregorian calendar date as YYYY-MM-DD, where the instrument is identified by date of issue. An instrument uses one form consistently throughout its life. Character handling. Every character permitted by the grammar is a member of the "pchar" production of RFC 3986 and requires no percent-encoding. Names therefore never contain percent-encoded octets; a candidate name containing a percent character is malformed and must be rejected rather than decoded. Characters outside the ASCII range must not appear. PDA publishes in Urdu and English and carries titles in both, but those titles are metadata held in the Register, not identifiers. URN-equivalence. Two names are URN-equivalent if they are URN-equivalent under Section 3.1 of RFC 8141 after additionally applying ASCII case folding to the entire NSS. That is, the NSS in this namespace is case-insensitive, so that urn:dnp:standard:dnp-d.002:v1 urn:dnp:standard:DNP-D.002:v1 URN:DNP:STANDARD:DNP-D.002:V1 are URN-equivalent. Because names never contain percent-encoded octets, the percent-encoding provisions of Section 3.1 have no effect here. No further equivalence rules are defined; hyphens and full stops within local-id are significant and are not elided. This rule only eliminates false negatives relative to the base procedure and does not cause any two names to be treated as distinct that the base procedure treats as URN-equivalent. PDA assigns in lower case and the lower-case form is canonical. Components. This namespace defines no r-component semantics; consistent with Section 2.3.1 of RFC 8141, PDA does not use them and will not do so before their semantics are standardised. A PDA resolver presented with an r-component ignores it and resolves the assigned-name. No q-component semantics of its own are defined; where a name resolves to a URI that is a locator, a q-component is handled per Section 2.3.2, and PDA resolution does not depend on q-component information. f-components are interpreted per Section 2.3.3, according to the media type of the retrieved representation; where PDA publishes an instrument in HTML it provides stable fragment identifiers corresponding to the numbered clauses, so that a clause-level citation such as urn:dnp:standard:dnp-x.001:v2#clause-4.3 is durable. Examples (illustrative, not assignments): urn:dnp:standard:dnp-x.001 urn:dnp:standard:dnp-x.001:v2 urn:dnp:architecture:dnp-d.002:v1 urn:dnp:masterplan:ndmp-2026-2035 urn:dnp:masterplan:ndmp-2026-2035:v1 urn:dnp:directive:pda-2026-014:2026-03-11 urn:dnp:schema:person-name:v1 urn:dnp:api:consent:v2 urn:dnp:dataset:connectivity-index:2026-06-30 Assignment: Assignment is closed. Names are assigned solely by PDA, or by an entity acting under a written instrument of delegation issued by PDA. Assignment is not open to application by third parties and there is no procedure by which an external party may request a name for a resource of its own. Delegation is contemplated because instruments within scope may be issued by federal or provincial entities under PDA's coordinating mandate. A delegation instrument specifies the classes and local-id ranges within which the delegate may assign and binds the delegate to the persistence commitment below. Delegations in force are listed in the Register. A delegation may be withdrawn; withdrawal has no effect on names already assigned under it, which remain valid and remain the responsibility of PDA. Uniqueness is enforced by construction. PDA maintains a single authoritative Register of all assigned names. A name comes into existence only by an entry in the Register, and the Register rejects any entry whose assigned-name is URN-equivalent to an existing entry. Delegated assignment operates against the same Register within non-overlapping allocations, so a delegate cannot create a collision. The Register records, for each name: the assigned-name; the instrument or edition identified; date of assignment; assigning entity; current status; current authoritative location(s); a SHA-256 digest of each edition-level resource as published, with the algorithm named in the record so it can be migrated without invalidating existing entries; titles and descriptive metadata in English and Urdu; and any equivalent identifiers in other identifier systems. Persistence. An assigned name is never reassigned to a different resource and is never withdrawn or deleted, under any circumstance. This holds where the instrument is superseded, repealed, withdrawn from effect or found to have been issued in error; where the issuing unit is abolished or merged; and where the instrument is no longer published. Each such fact is recorded as a status change in the Register and the name continues to identify what it has always identified. An error in assignment is corrected by recording the error and, if necessary, assigning a new name -- never by reusing the erroneous name. Institutional continuity. PDA is created by primary legislation rather than administrative order, and its dissolution or reconstitution would require an act of the legislature. Pakistani legislative practice on the reconstitution of statutory bodies provides for devolution of the functions, assets and records of a dissolved body upon a successor or upon the Division concerned; stewardship of this namespace is a function within the meaning of that practice. In addition, and not dependent on PDA's continued existence: the Register is published in full as a machine-readable open dataset on a regular cycle, so a complete copy exists outside PDA's systems at all times; PDA deposits the Register and the published corpus with Pakistan's national depository institutions and offers the same deposit to international web-archiving initiatives; and if stewardship is transferred, PDA or its successor will submit a revised registration template recording the change. Transfer of stewardship does not affect names already assigned. Security and Privacy: Absence of personal data. Names are assigned only to published institutional instruments. A name must not be assigned to a natural person, to an identifier of a natural person, to an individual transaction or record, or to any resource whose representation contains personal data. This is stated normatively because PDA's statutory mandate includes national digital identity and data exchange infrastructure, and because the persistence properties that make URNs attractive for institutional documents make them actively harmful for identifiers of people: a URN cannot be revoked, and the non-reassignment commitment above is incompatible with any right of erasure. Extending this namespace to personal or transactional data would require a separate registration with a materially different persistence model, and PDA does not propose one. The Register records the names of officers only in their institutional capacity as approving authority, information already disclosed on the face of the published instrument. Comparison and confusion. The case-insensitivity rule introduces the general risks discussed in RFC 6943 for identifiers compared under case folding. Those risks are bounded here: the NSS is restricted to ASCII letters, digits, hyphen and full stop, so ASCII case folding is locale-independent, involves no expansion or contraction of characters, and admits no Unicode confusables; the Turkish dotless-i problem does not arise because folding is defined over the ASCII range only. Implementations must perform case folding over the ASCII range only and must not apply locale-sensitive case mapping. A false positive could cause a system to treat a conformance assertion against one instrument as an assertion against another; because the identified resources are public and the Register is published, such an error is detectable by inspection. Separately, because the work-level and edition-level forms are similar in appearance but identify different resources, an implementation that truncates or elides the edition field silently substitutes a work-level citation for an edition-level one, changing its meaning in conformance and contractual contexts. Implementations must not truncate names, and applications displaying names should display them in full, consistent with Section 4.4 of RFC 8141. Authenticity and spoofing. A name asserts nothing about the authenticity of any document it may be attached to; a third party can place the string "urn:dnp:standard:..." on an unauthorised document as easily as any other string. The Register, retrieved over an authenticated channel, is the sole authority for what a name identifies, and the cryptographic digests it records for edition-level resources allow a retrieved document to be checked against the published text. Consistent with Section 8 of RFC 8141, the information in this registration is a declaration and should be treated as advisory. Resolver operation. Queries to a PDA resolver reveal to PDA which instruments a client is interested in and, absent transport security, reveal the same to network observers. Because the corpus includes policy and regulatory instruments, that interest may be sensitive: a pattern of queries could indicate the direction of an entity's compliance work or a researcher's enquiry. Accordingly the PDA resolver is offered over HTTPS, does not require authentication, and does not require or set cookies. PDA retains resolver request logs only in aggregate form and only for capacity planning, and does not disclose individual query records. Directory harvesting is not a concern: the Register is published in full and complete enumeration of the namespace is an intended feature. Interoperability: The string "DNP". "DNP" is used as a corporate brand by at least one substantial commercial enterprise unconnected with Pakistan, which holds trademark registrations in that string in several jurisdictions and operates a brand top-level domain under it. "DNP" is also an established abbreviation in chemistry and in professional sport. PDA claims no association with any of these, and registration of this NID neither asserts nor implies any right in the string outside the URN namespace registry. PDA has selected the string because it is the statutory short form of Digital Nation Pakistan, the programme established by the Act of the same name, and not for any associative value. PDA notes that a URN NID and a trademark occupy different registries with different scopes, that no resource identified in this namespace is a commercial good or service, and that Section 5.1 of RFC 8141 leaves disputes over strings to the parties concerned. PDA will engage in good faith with any objection raised during Expert Review. Legal norms. Names in this namespace are not assigned to primary legislation of the Islamic Republic of Pakistan; authority over the authoritative text of Acts of Parliament rests with organs of the State other than PDA, and the "lex" namespace (RFC 9676) is the appropriate vehicle. Instruments in this namespace frequently cite Pakistani legislation; such citations use "lex" or the Gazette reference. An instrument issued by PDA under a statutory power is identified here as a directive; the statutory power itself is not. Bibliographic identifiers. Pakistan has no registered National Bibliography Number sub-namespace under RFC 8458 at the time of writing. Should the national library register one and assign numbers within this corpus, the resulting names would coexist with names in this namespace and be recorded as equivalences in the Register. The same applies to Digital Object Identifiers, which PDA assigns to some publications through a DOI registration agency. Internal document codes. PDA's internal document nomenclature codes appear within local-id in transformed form. Those codes circulate outside the URN context and a reader may encounter one in isolation. A bare code is not a URN and must not be treated as one. The transformation from code to local-id is defined in the PDA specification cited under Documentation; it is not in general reversible by inspection and implementations must not attempt to reverse it algorithmically. Character handling. Because names never contain percent-encoded octets or non-ASCII characters, the encoding pitfalls that arise when pre-existing identifier systems are mapped into URN syntax do not arise here. Protocol slots. Consistent with Section 4.1 of RFC 8141, a name in this namespace is not a locator and should not be placed in a URI protocol slot whose defined semantics require dereferencing to a representation. It is appropriate in citation fields, metadata records, conformance declarations and provenance assertions. Resolution: Resolution is intended. PDA operates and undertakes to continue to operate a resolution service for names in this namespace. The service is offered over HTTPS at the endpoint https://urn.dnp.gov.pk/, operated by PDA under a Government of Pakistan domain. A client resolves a name by requesting the assigned-name as a path component of that endpoint, for example https://urn.dnp.gov.pk/urn:dnp:standard:dnp-x.001:v2. The endpoint is also recorded in the PDA specification cited under Documentation; should it ever change, the specification and this registration are revised, and names already assigned are unaffected. Resolution follows the work/edition distinction. A work-level name resolves to the current authoritative edition of the instrument where one is in force, and otherwise to a record describing the instrument and its editions. An edition-level name resolves to that fixed edition and continues to do so after it has been superseded; a superseded edition remains retrievable, marked as superseded with a reference to the edition that superseded it. PDA does not remove superseded editions from publication. A client requesting a representation of the identified resource receives it, or a redirection to its current location. A client requesting metadata by content negotiation receives the Register record, including status, provenance, digests and equivalent identifiers. Resolution of a syntactically valid but unassigned name is distinguishable from resolution of a name assigned to an instrument no longer in force; the two are different conditions and are reported differently. Resolution is a convenience and is not constitutive. A name remains valid, and continues to identify what the Register says it identifies, irrespective of whether any resolver is reachable and irrespective of whether the identified resource is currently retrievable. PDA does not at present operate a registration process for third-party publicly advertised resolution services for this namespace and does not at present recommend any resolver other than its own. Should PDA establish such a process, the requirements for being publicly advertised will be published alongside the specification cited under Documentation and this registration will be revised. PDA defines no r-components and its resolver ignores any that are presented. Documentation: [1] Sohail, H., Ed., "A Uniform Resource Name (URN) Namespace for Digital Nation Pakistan (DNP)", Work in Progress, Internet-Draft, draft-sohail-urn-dnp-01, 10 August 2026, <https://datatracker.ietf.org/doc/draft-sohail-urn-dnp/>. A stable specification maintained by PDA is published at https://standards.dnp.gov.pk/ under the PDA document nomenclature framework and is kept synchronised with this registration. Additional Information: The Digital Nation Pakistan Act, 2025 establishes PDA and sets out its functions, including the specification of national digital standards and the coordination of digitalisation across government. The Act is the source of PDA's authority to issue the instruments identified in this namespace and of the institutional continuity described under Assignment. Registrations by national government bodies in this registry include those of the Federal Chancellery of the Republic of Austria and of New Zealand (RFC 4350), and by national memory institutions including the National Archives of Finland. This registration follows the same pattern: a national public authority seeking durable identification of a defined corpus of official material, without any claim to a country-code-derived namespace. PDA notes that Section 5.1 of RFC 8141 reserves strings of the form ALPHA ALPHA "-" for possible future country-code-based registration; this registration does not use, anticipate or depend on any such reservation. Revision Information: None. This is the initial registration. ________________________________ From: Director - Stakeholder Engagement <[email protected]> Sent: 06 August 2026 11:40 To: Dale R. Worley <[email protected]> Cc: [email protected] <[email protected]> Subject: Re: [urn] Request for registration of formal URN namespace "dnp" Dale, Thank you for the review, and for the kind words. You are right, and I accept the change. As drafted, the grammar restricted "class" to lower-case letters while the URN-equivalence rule declared the whole NSS case-insensitive and offered an all-upper-case name as an example of an equivalent form. On the strict reading, the third example was not a well-formed name at all, so the equivalence rule was ranging over a string the grammar rejected. Adopting your production: class = 2*24( ALPHA ) and removing the now-unused "lowalpha" production. The intent was that the grammar define well-formedness while assignment practice, which is narrower, keep everything in lower case. Encoding the narrower practice in the grammar was the error. Resolving it in favour of the equivalence rule changes neither which names are URN-equivalent to which, nor PDA's practice of assigning in lower case, which the canonicality statement still records. I have also added a sentence to the Syntax field making the distinction explicit, so the same question is not raised against a later revision: The grammar above defines the set of well-formed names. It is wider than the set of names PDA assigns: the assignment practice for local-id and the canonical form given under URN-equivalence below are narrower, and a name presented in a form the grammar admits but PDA would not have assigned is still well-formed and still URN-equivalent to the assigned name. Both changes will appear in draft-sohail-urn-dnp-01, together with any further comments the list has over the next few days. I will post a revised registration template alongside it, clearly marked as superseding the one sent on 5 August, so there is no ambiguity about which text is being put to IANA. Regards, Hira Sohail Director of Partnerships & Stakeholder Engagement [cid:b364b46c-ef91-491b-b49d-42f7b10bb827] Phone +92 302 516 1211 Email [email protected] Website https://pda.gov.pk/ Address Pakistan Digital Authority - 4th Floor, Daftarkhwan Vanguard, 5 Constitution Ave. F-5/1, Islamabad, Pakistan. Disclaimer: This email and any attachments may contain confidential or privileged information and are intended only for the individual(s) to whom they are addressed. If you are not the intended recipient, please notify the sender immediately and delete this email from your system. Any unauthorized copying, disclosure, or distribution of the contents of this email is strictly prohibited. ________________________________ From: Dale R. Worley <[email protected]> Sent: 06 August 2026 06:54 To: Director - Stakeholder Engagement <[email protected]> Cc: [email protected] <[email protected]> Subject: Re: [urn] Request for registration of formal URN namespace "dnp" Director - Stakeholder Engagement <[email protected]> writes: > URN NAMESPACE REGISTRATION TEMPLATE > Formal Namespace Identifier: dnp > Submitted under Section 6.2 of RFC 8141 (Expert Review) This is easily the best registration template I've seen. I heartily approve of it, although I suggest one minor change. The syntax is given as: > NSS = class ":" local-id [ ":" edition ] > > class = 2*24( lowalpha ) > > local-id = id-start *62( id-char ) id-end > id-start = ALPHA / DIGIT > id-char = ALPHA / DIGIT / "-" / "." > id-end = ALPHA / DIGIT > > edition = ver-form / date-form > ver-form = "v" 1*3DIGIT [ "." 1*3DIGIT ] > date-form = 4DIGIT "-" 2DIGIT "-" 2DIGIT > > lowalpha = %x61-7A ; a-z Taken literally, this means that the "class" must be composed of lower-case letters. But the section on equivalence says that lower case and upper case are equivalent: > URN-equivalence. Two names are URN-equivalent if they are > URN-equivalent under Section 3.1 of RFC 8141 after additionally > applying ASCII case folding to the entire NSS. That is, the NSS > in this namespace is case-insensitive, so that > > urn:dnp:standard:dnp-d.002:v1 > urn:dnp:standard:DNP-D.002:v1 > URN:DNP:STANDARD:DNP-D.002:V1 Taking the syntax description strictly, the "class" must be lower-case (although the local-id may be mixed case), which means that the third example is not a valid DNP URN. I recommend replacing the "class" production with > class = 2*24( ALPHA ) and eliminating the "lowalpha" production. Dale Thanks Best Regards, Director - Stakeholder Engagement ________________________________ [Pakistan Digital Authority] [Phone] [Mobile] [Email] +92 51 920 5027 [email protected] [Location] 4th Floor, 5-A Constitution Avenue Sector F-5/1, Islamabad 44000, Pakistan [Digital Nation Pakistan] [Transforming Pakistan into a Digital Nation, enabling a Digital Society, Digital Economy, and Digital Governance] _______________________________________________ urn mailing list -- [email protected] To unsubscribe send an email to [email protected]
image.png
(image/png, 11.9 KB) - not displayed