idnits 2.17.1 draft-kunze-ark-11.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- ** It looks like you're using RFC 3978 boilerplate. You should update this to the boilerplate described in the IETF Trust License Policy document (see https://trustee.ietf.org/license-info), which is required now. -- Found old boilerplate from RFC 3978, Section 5.1 on line 17. -- Found old boilerplate from RFC 3978, Section 5.5 on line 2334. ** Found boilerplate matching RFC 3978, Section 5.4, paragraph 1 (on line 2326), which is fine, but *also* found old RFC 2026, Section 10.4C, paragraph 1 text on line 38. ** This document has an original RFC 3978 Section 5.4 Copyright Line, instead of the newer IETF Trust Copyright according to RFC 4748. ** This document has an original RFC 3978 Section 5.5 Disclaimer, instead of the newer disclaimer which includes the IETF Trust according to RFC 4748. ** The document seems to lack an RFC 3979 Section 5, para. 1 IPR Disclosure Acknowledgement. ** The document seems to lack an RFC 3979 Section 5, para. 2 IPR Disclosure Acknowledgement. ** The document seems to lack an RFC 3979 Section 5, para. 3 IPR Disclosure Invitation. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- ** The document seems to lack a 1id_guidelines paragraph about 6 months document validity -- however, there's a paragraph with a matching beginning. Boilerplate error? ** The document is more than 15 pages and seems to lack a Table of Contents. == No 'Intended status' indicated for this document; assuming Proposed Standard == The page length should not exceed 58 lines per page, but there was 49 longer pages, the longest (page 2) being 63 lines == It seems as if not all pages are separated by form feeds - found 0 form feeds but 50 pages Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** The document seems to lack an IANA Considerations section. (See Section 2.2 of https://www.ietf.org/id-info/checklist for how to handle the case when there are no actions for IANA.) ** The document seems to lack separate sections for Informative/Normative References. All references will be assumed normative when checking for downward references. ** There are 8 instances of too long lines in the document, the longest one being 19 characters in excess of 72. ** The abstract seems to contain references ([Qualifier]), which it shouldn't. Please replace those with straight textual mentions of the documents in question. == There are 8 instances of lines with non-RFC2606-compliant FQDNs in the document. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the RFC 3978 Section 5.4 Copyright Line does not match the current year == Line 1124 has weird spacing: '... regexp repla...' -- The document seems to lack a disclaimer for pre-RFC5378 work, but may have content which was first submitted before 10 November 2008. If you have contacted all the original authors and they are all willing to grant the BCP78 rights to the IETF Trust, then this is fine, and you can ignore this comment. If not, you may need to add the pre-RFC5378 disclaimer. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (23 February 2006) is 6637 days in the past. Is this intentional? -- Found something which looks like a code comment -- if you have code sections in the document, please surround them with '' and '' lines. Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Missing Reference: 'Qualifier' is mentioned on line 437, but not defined == Unused Reference: 'MD5' is defined on line 2129, but no explicit reference was found in the text -- Possible downref: Non-RFC (?) normative reference: ref. 'ANVL' -- Possible downref: Non-RFC (?) normative reference: ref. 'ARK' -- Possible downref: Non-RFC (?) normative reference: ref. 'DCORE' -- Possible downref: Non-RFC (?) normative reference: ref. 'DERC' -- Possible downref: Non-RFC (?) normative reference: ref. 'DOI' -- Possible downref: Non-RFC (?) normative reference: ref. 'ERC' -- Possible downref: Non-RFC (?) normative reference: ref. 'Handle' ** Obsolete normative reference: RFC 2616 (ref. 'HTTP') (Obsoleted by RFC 7230, RFC 7231, RFC 7232, RFC 7233, RFC 7234, RFC 7235) ** Downref: Normative reference to an Informational RFC: RFC 1321 (ref. 'MD5') ** Obsolete normative reference: RFC 2915 (ref. 'NAPTR') (Obsoleted by RFC 3401, RFC 3402, RFC 3403, RFC 3404) -- Possible downref: Non-RFC (?) normative reference: ref. 'NLMPerm' -- Possible downref: Non-RFC (?) normative reference: ref. 'NOID' -- Possible downref: Non-RFC (?) normative reference: ref. 'PURL' ** Obsolete normative reference: RFC 822 (Obsoleted by RFC 2822) -- Possible downref: Non-RFC (?) normative reference: ref. 'TEMPER' -- Possible downref: Non-RFC (?) normative reference: ref. 'THUMP' ** Obsolete normative reference: RFC 2396 (ref. 'URI') (Obsoleted by RFC 3986) ** Downref: Normative reference to an Informational RFC: RFC 2288 (ref. 'URNBIB') ** Obsolete normative reference: RFC 2141 (ref. 'URNSYN') (Obsoleted by RFC 8141) ** Obsolete normative reference: RFC 2611 (ref. 'URNNID') (Obsoleted by RFC 3406) Summary: 21 errors (**), 0 flaws (~~), 8 warnings (==), 17 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Internet-Draft: draft-kunze-ark-11.txt J. Kunze 3 ARK Identifier Scheme University of California (UCOP) 4 Expires 23 August 2006 R. P. C. Rodgers 5 US National Library of Medicine 6 23 February 2006 8 The ARK Persistent Identifier Scheme 10 (http://www.ietf.org/internet-drafts/draft-kunze-ark-11.txt) 12 Status of this Document 14 By submitting this Internet-Draft, each author represents that any 15 applicable patent or other IPR claims of which he or she is aware 16 have been or will be disclosed, and any of which he or she becomes 17 aware will be disclosed, in accordance with Section 6 of BCP 79. 19 Internet-Drafts are working documents of the Internet Engineering 20 Task Force (IETF), its areas, and its working groups. Note that 21 other groups may also distribute working documents as Internet- 22 Drafts. 24 Internet-Drafts are draft documents valid for a maximum of six months 25 and may be updated, replaced, or obsoleted by other documents at any 26 time. It is inappropriate to use Internet-Drafts as reference 27 material or to cite them other than as ``work in progress.'' 29 The list of current Internet-Drafts can be accessed at 30 http://www.ietf.org/ietf/1id-abstracts.txt 32 The list of Internet-Draft Shadow Directories can be accessed at 33 http://www.ietf.org/shadow.html. 35 Distribution of this document is unlimited. Please send comments to 36 jak@ucop.edu. 38 Copyright (C) The Internet Society (2006). All Rights Reserved. 40 Abstract 42 The ARK (Archival Resource Key) naming scheme is designed to 43 facilitate the high-quality and persistent identification of 44 information objects. A founding principle of the ARK is that 45 persistence is purely a matter of service and is neither inherent in 46 an object nor conferred on it by a particular naming syntax. The best 47 that an identifier can do is to lead users to the services that 48 support persistence. The term ARK itself refers both to the scheme 49 and to any single identifier that conforms to it. An ARK has five 50 components: 52 [http://NMAH/]ark:/NAAN/Name[Qualifier] 54 an optional and mutable Name Mapping Authority Hostport, the "ark:" 55 label, the Name Assigning Authority Number (NAAN), the assigned Name, 56 and an optional and possibly mutable Qualifier supported by the NMA. 57 The NAAN and Name together form the immutable persistent identifier 58 for the object. An ARK is a special kind of URL that connects users 59 to three things: the named object, its metadata, and the provider's 60 promise about its persistence. When entered into the location field 61 of a Web browser, the ARK leads the user to the named object. That 62 same ARK, followed by a single question mark ('?'), returns a brief 63 metadata record that is both human- and machine-readable. When the 64 ARK is followed by dual question marks ('??'), the returned metadata 65 contains a commitment statement from the current provider. Tools 66 exist for minting, binding, and resolving ARKs. 68 1. Introduction 70 This document describes a scheme for the high-quality naming of 71 information resources. The scheme, called the Archival Resource Key 72 (ARK), is well suited to long-term access and identification of any 73 information resources that accommodate reasonably regular electronic 74 description. This includes digital documents, databases, software, 75 and websites, as well as physical objects (books, bones, statues, 76 etc.) and intangible objects (chemicals, diseases, vocabulary terms, 77 performances). Hereafter the term "object" refers to an information 78 resource. The term ARK itself refers both to the scheme and to any 79 single identifier that conforms to it. A reasonably concise and 80 accessible overview and rationale for the scheme is available at 81 [ARK]. 83 Schemes for persistent identification of network-accessible objects 84 are not new. In the early 1990's, the design of the Uniform Resource 85 Name [URNSYN] responded to the observed failure rate of URLs by 86 articulating an indirect, non-hostname-based naming scheme and the 87 need for responsible name management. Meanwhile, promoters of the 88 Digital Object Identifier [DOI] succeeded in building a community of 89 providers around a mature software system [Handle] that supports name 90 management. The Persistent Uniform Resource Locator [PURL] was 91 another scheme that has the unique advantage of working with 92 unmodified web browsers. ARKs represent an approach that attempts to 93 build on the strengths and to avoid the weaknesses of the other 94 schemes. 96 A founding principle of the ARK is that persistence is purely a 97 matter of service. Persistence is neither inherent in an object nor 98 conferred on it by a particular naming syntax. Nor is the technique 99 of name indirection - upon which URNs, Handles, DOIs, and PURLs are 100 founded - of central importance. Name indirection is an ancient and 101 well-understood practice; new mechanisms for it keep appearing and 102 distracting practitioner attention, with the Domain Name System [DNS] 103 being a particularly dazzling and elegant example. What is often 104 forgotten is that maintenance of an indirection table is the 105 overwhelming and unavoidable cost to the organization providing 106 persistence, and the cost is equivalent across naming schemes. That 107 indirection has always been a native part the web while being so 108 lightly utilized for the persistence of web-based objects is an 109 indication of how unsuited most organizations are to the task of 110 table maintenance and to the overall challenge of digital permanence. 112 Persistence is achieved through a provider's successful stewardship 113 of objects and their identifiers. The highest level of persistence 114 will be reinforced by a provider's robust contingency, redundancy, 115 and succession strategies. It is further safeguarded to the extent 116 that a provider's mission is shielded from marketplace and political 117 instabilities. These are by far the major challenges confronting 118 persistence providers, and no identifier scheme has any direct impact 119 on them. In fact, some schemes appear to be actual liabilities for 120 persistence because they create short- and long-term dependencies for 121 every object access on complex, special-purpose local and global 122 infrastructures, parts of which are proprietary and all of which 123 increase the carry-forward burden for the preservation community. It 124 is for this reason that the ARK scheme relies only on educated name 125 assignment and light use of general-purpose infrastructures that the 126 entire internet community needs (the DNS, web servers, and web 127 browsers) and that one can reasonably expect many others to help 128 carry forward into the technologically evolving future. 130 1.1. Reasons to Use ARKs 132 If no persistent identifier scheme contributes directly to 133 persistence, why not just use URLs? A particular URL may be as 134 durable an identifier as it is possible to have, but nothing 135 distinguishes it from an ordinary URL to the recipient who is 136 wondering if it is suitable for long-term reference. An ARK is just 137 a URL, distinguished by its form, that provides some of the necessary 138 conditions for credible persistence. An ARK invites access to not 139 one, but to three things: to the object, to its metadata, and to a 140 nuanced statement of commitment from the provider regarding the 141 object. Existence of the two extra services can be probed 142 automatically by appending either `?' or `??' to the ARK. 144 The form of the ARK also supports the natural separation of naming 145 authorities into the original name assigning authority and the 146 diverse multiple name mapping (or servicing) authorities that in 147 succession and in parallel will take over custodial responsibilities 148 from the original assigner for the large majority of a long-term 149 object's archival lifetime. The mapping authority, indicated by the 150 hostname part of the URL that contains the ARK, serves to launch the 151 ARK into cyberspace. Should it ever fail (and there is no reason why 152 a well-chosen hostname of a 100-year-old cultural memory institution 153 shouldn't last as long as the DNS), that host name is considered 154 disposeable and replaceable. Again, the form of the ARK helps 155 because it defines exactly how to recover the core immutable object 156 identity, and several simple algorithms (based on the URN model) are 157 defined for locating another mapping authority. 159 There are tools to assist in generating ARKs and other identifiers, 160 such as [NOID] and "uuidgen", both of which rely for uniqueness on 161 human-maintained registries. This document also contains some 162 guidelines and considerations for managing namespaces and choosing 163 hostnames wisely. 165 1.2. Three Requirements of ARKs 167 The first requirement of an ARK is to give users a link from an 168 object to a promise of stewardship for it. That promise is a multi- 169 faceted covenant that binds the word of an identified service 170 provider to a specific set of responsibilities. No one can tell if 171 successful stewardship will take place because no one can predict the 172 future. Reasonable conjecture, however, may be based on past 173 performance. There must be a way to tie a promise of persistence to 174 a provider's demonstrated or perceived ability - its reputation - in 175 that arena. Provider reputations would then rise and fall as 176 promises are observed variously to be kept and broken. This is 177 perhaps the best way we have for gauging the strength of any 178 persistence promise. Note that over time, current providers have 179 nothing to do with the intentions of the original assigners of names. 181 The second requirement of an ARK is to give users a link from an 182 object to a description of it. The problem with a naked identifier 183 is that without a description real identification is incomplete. 184 Identifiers common today are relatively opaque, though some contain 185 ad hoc clues that reflect brief life cycle periods such as the 186 address of a short stay in a filesystem hierarchy. Possession of 187 both an identifier and an object is some improvement, but positive 188 identification may still be uncertain since the object itself might 189 not include a matching identifier or might not carry evidence obvious 190 enough to reveal its identity without significant research. In 191 either case, what is called for is a record bearing witness to the 192 identifier's association with the object, as supported by a recorded 193 set of object characteristics. This descriptive record is partly an 194 identification "receipt" with which users and archivists can verify 195 an object's identity after brief inspection and a plausible match 196 with recorded characteristics such as title and size. 198 The final requirement of an ARK is to give users a link to the object 199 itself (or to a copy) if at all possible. Persistent access is the 200 central duty of an ARK. Persistent identification plays a vital 201 supporting role but, strictly speaking, it can be construed as no 202 more than a record attesting to the original assignment of a never- 203 reassigned identifier. Object access may not be feasible for various 204 reasons, such as catastrophic loss of the object, a licensing 205 agreement that keeps an archive "dark" for a period of years, or when 206 an object's own lack of tangible existence confuses normal concepts 207 of access (e.g., a vocabulary term might be accessed through its 208 definition). In such cases the ARK's identification role assumes a 209 much higher profile. But attempts to simplify the persistence 210 problem by decoupling access from identification and concentrating 211 exclusively on the latter are of questionable utility. A perfect 212 system for assigning forever unique identifiers might be created, but 213 if it did so without reducing access failure rates, no one would be 214 interested. The central issue - which may be summed up as the "HTTP 215 404 Not Found" problem - would not have been addressed. 217 1.3. Organizing Support for ARKs: Our Stuff vs. Their Stuff 219 An organization and the user community it serves can often be seen to 220 struggle with two different areas of persistent identification: the 221 Our Stuff problem and the Their Stuff problem. In the Our Stuff 222 problem, we in the organization want our own objects to acquire 223 persistent names. Since we possess or control these objects, our 224 organization tackles the Our Stuff problem directly. Whether or not 225 the objects are named by ARKs, our organization is the responsible 226 party, so it can plan for, maintain, and make commitments about the 227 objects. 229 In the Their Stuff problem, we in the organization want others' 230 objects to acquire persistent names. These are objects that we do 231 not own or control, but some of which are critically important to us. 232 But because they are beyond our influence as far as support is 233 concerned, creating and maintaining persistent identifiers for Their 234 Stuff is not especially purposeful or feasible for us to do. There 235 is little that we can do about someone else's stuff except encourage 236 them to find or become providers of persistence services. 238 Co-location of persistent access and identification services is 239 natural. Any organization that undertakes ongoing support of true 240 persistent identification (which includes description) is well-served 241 if it controls, owns, or otherwise has clear internal access to the 242 identified objects, and this gives it an advantage if it wishes also 243 to support persistent access to outsiders. Conversely, persistent 244 access to outsiders requires orderly internal collection management 245 procedures that include monitoring, acquisition, verification, and 246 change control over objects, which in turn requires object 247 identifiers persistent enough to support auditable record keeping 248 practices. 250 Although, organizing ARK services under one roof thus tends to make 251 sense, object hosting can successfully be separated from name 252 mapping. An example is when a name mapping authority centrally 253 provides uniform resolution services via a protocol gateway on behalf 254 of organizations that host objects behind a variety of access 255 protocols. It is also reasonable to build value-added description 256 services that rely on the underlying services of a set of mapping 257 authorities. 259 Supporting ARKs is not for every organization. By requiring 260 specific, revealed commitments to preservation, to object access, and 261 to description, the bar for providing ARK services is higher than for 262 some other identifier schemes. On the other hand, it would be hard 263 to grant credence to a persistence promise from an organization that 264 could not muster the minimum ARK services. Not that there isn't a 265 business model for an ARK-like, description-only service built on top 266 of another organization's full complement of ARK services. For 267 example, there might be competition at the description level for 268 abstracting and indexing a body of scientific literature archived in 269 a combination of open and fee-based repositories. The description- 270 only service would have no direct commitment to the objects, but 271 would act as an intermediary, forwarding commitment statements from 272 object hosting services to requestors. 274 1.4. Definition of Identifier 276 An identifier is not a string of character data - an identifier is an 277 association between a string of data and an object. This abstraction 278 is necessary because without it a string is just data. It's nonsense 279 to talk about a string's breaking, or about its being strong, 280 maintained, and authentic. But as a representative of an 281 association, a string can do, metaphorically, the things that we 282 expect of it. 284 Without regard to whether an object is physical, digital, or 285 conceptual, to identify it is to claim an association between it and 286 a representative string, such as "Jane" or "ISBN 0596000278". What 287 gives a claim credibility is a set of verifiable assertions, or 288 metadata, about the object, such as age, height, title, or number of 289 pages. In other words, the association is made manifest by a record 290 (e.g., a cataloging or other metadata record) that vouches for it. 292 In the complete absence of any testimony (metadata) regarding an 293 association, a would-be identifier string is a meaningless sequence 294 of characters. To keep an externally visible but otherwise internal 295 string from being perceived as an identifier by outsiders, for 296 example, it suffices for an organization not to disclose the nature 297 of its association. For our immediate purpose, actual existence of 298 an association record is more important than its authenticity or 299 verifiability, which are outside the scope of this specification. 301 It is a gift to the identification process if an object carries its 302 own name as an inseparable part of itself, such as an identifier 303 imprinted on the first page of a document or embedded in a data 304 structure element of a digital document header. In cases where the 305 object is large, unwieldy, or unavailable (such as when licensing 306 restrictions are in effect), a metadata record that includes the 307 identifier string will usually suffice. That record becomes a 308 conveniently manipulable object surrogate, acting as both an 309 association "receipt" and "declaration". 311 Note that our definition of identifier extends the one in use for 312 Uniform Resource Identifiers [URI]. The present document still 313 sometimes (ab)uses the terms "ARK" and "identifier" as shorthand for 314 the string part of an identifier, but the context should make the 315 meaning clear. 317 2. ARK Anatomy 319 An ARK is represented by a sequence of characters (a string) that 320 contains the label, "ark:", optionally preceded by the beginning part 321 of a URL. Here is a diagrammed example. 323 http://foobar.zaf.org/ark:/12025/654xz321/s3/f8.05v.tiff 324 \___________________/ \__/ \___/ \______/ \____________/ 325 (replaceable) | | | Qualifier 326 | ARK Label | | (NMA-supported) 327 | | | 328 Name Mapping Authority | Name (NAA-assigned) 329 Hostport (NMAH) | 330 Name Assigning Authority Number (NAAN) 332 The ARK syntax can be summarized, 334 [http://NMAH/]ark:/NAAN/Name[Qualifier] 336 where the NMAH and Qualifier parts are in brackets to indicate that 337 they are optional. 339 2.1. The Name Mapping Authority Hostport (NMAH) 341 Before the "ark:" label may appear an optional Name Mapping Authority 342 Hostport (NMAH) that is a temporary address where ARK service 343 requests may be sent. It consists of "http://" (or any service 344 specification valid for a URL) followed by an Internet hostname or 345 hostport combination having the same format and semantics as the 346 hostport part of a URL. The most important thing about the NMAH is 347 that it is "identity inert" from the point of view of object 348 identification. In other words, ARKs that differ only in the 349 optional NMAH part identify the same object. Thus, for example, the 350 following three ARKs are synonyms for just one information object: 352 http://loc.gov/ark:/12025/654xz321 353 http://rutgers.edu/ark:/12025/654xz321 354 ark:/12025/654xz321 356 Strictly speaking, in the realm of digital objects, these ARKs may 357 lead over time to somewhat different or diverging instances of the 358 originally named object. In an ideal world, divergence of persistent 359 objects is not desirable, but it is widely believed that digital 360 preservation efforts will inevitably lead to alterations in some 361 original objects (e.g, a format migration in order to preserve the 362 ability to display a document). If any of those objects are held 363 redundantly in more than one organization (a common preservation 364 strategy), chances are small that all holding organizations will 365 perform the same precise transformations and all maintain the same 366 object metadata. More significant divergence would be expected when 367 the holding organizations serve different audiences or compete with 368 each other. 370 The NMAH part makes an ARK into an actionable URL. As with many 371 internet parameters, it is helpful to approach the NMAH being liberal 372 in what you accept and conservative in what you propose. From the 373 recipient's point of view, the NMAH part should be treated as 374 temporary, disposable, and replaceable. From the NMA's point of 375 view, it should be chosen with the greatest concern for longevity. A 376 carefully chosen NMAH should be at least as permanent as the 377 providing organization's own hostname. In the case of a national or 378 university library, for example, there is no reason why the NMAH 379 should not be considerably more permanent than soft-funded proxy 380 hostnames such as hdl.handle.net, dx.doi.org, and purl.org. In 381 general and over time, however, it is not unexpected for an NMAH 382 eventually to stop working and require replacement with the NMAH of a 383 currently active service provider. 385 This replacement relies on a mapping authority "resolver" discovery 386 process, of which two alternate methods are outlined in a later 387 section. The ARK, URN, Handle, and DOI schemes all use a resolver 388 discovery model that sooner or later requires matching the original 389 assigning authority with a current provider servicing that 390 authority's named objects; once found, the resolver at that provider 391 performs what amounts to a redirect to a place where the object is 392 currently held. All the schemes rely on the ongoing functionality of 393 currently mainstream technologies such as the Domain Name System 394 [DNS] and web browsers. The Handle and DOI schemes in addition 395 require that the Handle protocol layer and global server grid be 396 available at all times. 398 The practice of prepending "http://" and an NMAH to an ARK is a way 399 of creating an actionable identifier by a method that is itself 400 temporary. Assuming that infrastructure supporting [HTTP] 401 information retrieval will no longer be available one day, ARKs will 402 then have to be converted into new kinds of actionable identifiers. 403 By that time, if ARKs see widespread use, web browsers would 404 presumably evolve to perform this (currently simple) transformation 405 automatically. 407 2.2. The ARK Label Part - ark: 409 The label part distinguishes an ARK from an ordinary identifier. In 410 a URL found in the wild, the string, "ark:/", indicates that the URL 411 stands a reasonable chance of being an ARK. If the context warrants, 412 verification that it actually is an ARK can be done by testing it for 413 existence of the three ARK services. 415 Since nothing about an identifier syntax directly affects 416 persistence, the "ark:" label (like "urn:", "doi:", and "hdl:") 417 cannot tell you whether the identifier is persistent or whether the 418 object is available. It does tell you that the original Name 419 Assigning Authority (NAA) had some sort of hopes for it, but it 420 doesn't tell you whether that NAA is still in existence, or whether a 421 decade ago it ceased to have any responsibility for providing 422 persistence, or whether it ever had any responsibility beyond naming. 424 Only a current provider can say for certain what sort of commitment 425 it intends, and the ARK label suggests that you can query the NMAH 426 directly to find out exactly what kind of persistence is promised. 427 Even if what is promised is impersistence (i.e., a short-term 428 identifier), saying so is valuable information to the recipient. 429 Thus an ARK is a high-functioning identifier in the sense that it 430 provides access to the object, the metadata, and a commitment 431 statement, even if the commitment is explicitly very weak. 433 2.3. The Name Assigning Authority Number (NAAN) 435 Recalling that the general form of the ARK is, 437 [http://NMAH/]ark:/NAAN/Name[Qualifier] 439 the part of the ARK directly following the "ark:" is the Name 440 Assigning Authority Number (NAAN) enclosed in `/' (slash) characters. 441 This part is always required, as it identifies the organization that 442 originally assigned the Name of the object. It is used to discover a 443 currently valid NMAH and to provide top-level partitioning of the 444 space of all ARKs. NAANs are registered in a manner similar to URN 445 Namespaces, but they are pure numbers consisting of 5 digits or 9 446 digits. Thus, the first 100,000 registered NAAs fit compactly into 447 the 5 digits, and if growth warrants, the next billion fit into the 9 448 digit form. In either case the fixed odd numbers of digits helps 449 reduce the chances of finding a NAAN out of context and confusing it 450 with nearby quantities such as 4-digit dates. 452 The NAAN designates a top-level ARK namespace. Once registered for a 453 namespace, a NAAN is never re-registered. It is possible, however, 454 for there to be a succession of organizations that manage of an ARK 455 namespace one organization to succeed another 457 2.4. The Name Part 459 The part of the ARK just after the NAAN is the Name assigned by the 460 NAA, and it is also required. Semantic opaqueness in the Name part 461 is strongly encouraged in order to reduce an ARK's vulnerability to 462 era- and language-specific change. Identifier strings containing 463 linguistic fragments can create support difficulties down the road. 464 No matter how appropriate or even meaningless they are today, such 465 fragments may one day create confusion, give offense, or infringe on 466 a trademark as the semantic environment around us and our communities 467 evolves. 469 Names that look more or less like numbers avoid common problems that 470 defeat persistence and international acceptance. The use of digits 471 is highly recommended. Mixing in non-vowel alphabetic characters a 472 couple at a time is a relatively safe and easy way to achieve a 473 denser namespace (more possible names for a given length of the name 474 string). Such names have a chance of aging and traveling well. 475 Tools exists that mint, bind, and resolve opaque identifiers, with or 476 without check characters [NOID]. More on naming considerations is 477 given in a subsequent section. 479 2.5. The Qualifier Part 481 The part of the ARK following the NAA-assigned Name is an optional 482 Qualifier. It is a string that extends the base ARK in order to 483 create a kind of service entry point into the object named by the 484 NAA. At the discretion of the providing NMA, such a service entry 485 point permits an ARK to support access to individual hierarchical 486 components and subcomponents of an object, and to variants (versions, 487 languages, formats) of components. A Qualifier may be invented by 488 the NAA or by any NMA servicing the object. 490 In form, the Qualifier is a ComponentPath, or a VariantPath, or a 491 ComponentPath followed by a VariantPath. A VariantPath is introduced 492 and subdivided by the reserved character `.', and a ComponentPath is 493 introduced and subdivided by the reserved character `/'. In this 494 example, 496 http://foobar.zaf.org/ark:/12025/654xz321/s3/f8.05v.tiff 498 the string "/s3/f8" is a ComponentPath and the string ".05v.tiff" is 499 a VariantPath. The ARK Qualifier is a formalization of some 500 currently mainstream URL syntax conventions. This formalization 501 specifically reserves meanings that permit recipients to make strong 502 inferences about logical sub-object containment and equivalence based 503 only on the form of the received identifiers; there is great 504 efficiency in not having to inspect metadata records to discover such 505 relationships. NMAs are free not to disclose any of these 506 relationships merely by avoiding the reserved characters above. 507 Hierarchical components and variants are discussed further in the 508 next two sections. 510 The Qualifier, if present, differs from the Name in several important 511 respects. First, a Qualifier may have been assigned either by the 512 NAA or later by the NMA. The assignment of a Qualifier by an NMA 513 effectively amounts to an act of publishing a service entry point 514 within the conceptual object originally named by the NAA. For our 515 purposes, an ARK extended with a Qualifier assigned by an NMA will be 516 called an NMA-qualified ARK. 518 Second, a Qualifier assignment on the part of an NMA is made in 519 fulfillment of its service obligations and may reflect changing 520 service expectations and technology requirements. NMA-qualified ARKs 521 could therefore be transient, even if the base, unqualified ARK is 522 persistent. For example, it would be reasonable for an NMA to 523 support access to an image object through an actionable ARK that is 524 considered persistent even if the experience of that access changes 525 as linking, labeling, and presentation conventions evolve and as 526 format and security standards are updated. For an image "thumbnail", 527 that NMA could also support an NMA-qualified ARK that is considered 528 impersistent because the thumbnail will be replaced with higher 529 resolution images as network bandwidth and CPU speeds increase. At 530 the same time, for an originally scanned, high-resolution master, the 531 NMA could publish an NMA-qualfied ARK that is itself considered 532 persistent. Of course, the NMA must be able to return its separate 533 commitments to unqualified, NAA-assigned ARKs, to NMA-qualified ARKs, 534 and to any NAA-qualified ARKs that it supports. 536 A third difference between a Qualifier and a Name concerns the 537 semantic opaqueness constraint. When an NMA-qualified ARK is to be 538 used as a transient service entry point into a persistent object, the 539 priority given to semantic opaqueness observed by the NAA in the Name 540 part may be relaxed by the NMA in the Qualifier part. If service 541 priorities in the Qualifier take precedence over persistence, short- 542 term usability considerations may recommend somewhat semantically 543 laden Qualifier strings. 545 Finally, not only is the set of Qualifiers supported by an NMA 546 mutable, but different NMAs may support different Qualifier sets for 547 the same NAA-identified object. In this regard the NMAs act 548 independently of each other and of the NAA. 550 The next two sections describe how ARK syntax may be used to declare, 551 or to avoid declaring, certain kinds of relatedness among qualified 552 ARKs. 554 2.5.1. ARKs that Reveal Object Hierarchy 556 An NAA or NMA may choose to reveal the presence of a hierarchical 557 relationship between objects using the `/' (slash) character after 558 the Name part of an ARK. Some authorities will choose not to 559 disclose this information, while others will go ahead and disclose so 560 that manipulators of large sets of ARKs can infer object 561 relationships by simple identifier inspection; for example, this 562 makes it possible for a system to present a collapsed view of a large 563 search result set. 565 If the ARK contains an internal slash after the NAAN, the piece to 566 its left indicates a containing object. For example, publishing an 567 ARK of the form, 569 ark:/12025/654/xz/321 571 is equivalent to publishing three ARKs, 573 ark:/12025/654/xz/321 574 ark:/12025/654/xz 575 ark:/12025/654 577 together with a declaration that the first object is contained in the 578 second object, and that the second object is contained in the third. 580 Revealing the presence of hierarchy is completely up to the assigner 581 (NMA or NAA). It is hard enough to commit to one object's name, let 582 alone to three objects' names and to a specific, ongoing relatedness 583 among them. Thus, regardless of whether hierarchy was present 584 initially, the assigner, by not using slashes, reveals no shared 585 inferences about hierarchical or other inter-relatedness in the 586 following ARKs: 588 ark:/12025/654_xz_321 589 ark:/12025/654_xz 590 ark:/12025/654xz321 591 ark:/12025/654xz 592 ark:/12025/654 594 Note that slashes around the ARK's NAAN (/12025/ in these examples) 595 are not part of the ARK's Name and therefore do not indicate the 596 existence of some sort of NAAN super object containing all objects in 597 its namespace. A slash must have at least one non-structural 598 character (one that is neither a slash nor a period) on both sides in 599 order for it to separate recognizable structural components. So 600 initial or final slashes may be removed, and double slashes may be 601 converted into single slashes. 603 2.5.2. ARKs that Reveal Object Variants 605 An NAA or NMA may choose to reveal the possible presence of variant 606 objects or object components using the `.' (period) character after 607 the Name part of an ARK. Some authorities will choose not to 608 disclose this information, while others will go ahead and disclose so 609 that manipulators of large sets of ARKs can infer object 610 relationships by simple identifier inspection; for example, this 611 makes it possible for a system to present a collapsed view of a large 612 search result set. 614 If the ARK contains an internal period after Name, the piece to its 615 left is a base name and the piece to its right, and up to the end of 616 the ARK or to the next period is a suffix. A Name may have more than 617 one suffix, for example, 618 ark:/12025/654.24 619 ark:/12025/xz4/654.24 620 ark:/12025/654.20v.78g.f55 622 There are two main rules. First, if two ARKs share the same base 623 name but have different suffixes, the corresponding objects were 624 considered variants of each other (different formats, languages, 625 versions, etc.) by the assigner (NMA or NAA). Thus, the following 626 ARKs are variants of each other: 628 ark:/12025/654.20v.78g.f55 629 ark:/12025/654.321xz 630 ark:/12025/654.44 632 Second, publishing an ARK with a suffix implies the existence of at 633 least one variant identified by the ARK without its suffix. The ARK 634 otherwise permits no further assumptions about what variants might 635 exist. So publishing the ARK, 637 ark:/12025/654.20v.78g.f55 639 is equivalent to publishing the four ARKs, 641 ark:/12025/654.20v.78g.f55 642 ark:/12025/654.20v.78g 643 ark:/12025/654.20v 644 ark:/12025/654 646 Revealing the possibility of variants is completely up to the 647 assigner. It is hard enough to commit to one object's name, let 648 alone to multiple variants' names and to a specific, ongoing 649 relatedness among them. The assigner is the sole arbiter of what 650 constitutes a variant within its namespace, and whether to reveal 651 that kind of relatedness by using periods within its names. 653 A period must have at least one non-structural character (one that is 654 neither a slash nor a period) on both sides in order for it to 655 separate recognizable structural components. So initial or final 656 periods may be removed, and double periods may be converted into 657 single periods. Multiple suffixes should be arranged in sorted order 658 (pure ASCII collating sequence) at the end of an ARK. 660 2.6. Character Repertoires 662 The Name and Qualifier parts are strings of visible ASCII characters 663 and should be less than 128 bytes in length. The length restriction 664 keeps the ARK short enough to append ordinary ARK request strings 665 without running into transport restrictions (e.g., within HTTP GET 666 requests). Characters may be letters, digits, or any of these six 667 characters: 669 = # * + @ _ $ 671 The following characters may also be used, but their meanings are 672 reserved: 674 % - . / 676 The characters `/' and `.' are ignored if either appears as the last 677 character of an ARK. If used internally, they allow a name assigner 678 to reveal object hierarchy and object variants as previously 679 described. 681 Hyphens are considered to be insignificant and are always ignored in 682 ARKs. A `-' (hyphen) may appear in an ARK for readability, or it may 683 have crept in during the formatting and wrapping of text, but it must 684 be ignored in lexical comparisons. As in a telephone number, hyphens 685 have no meaning in an ARK. It is always safe for an NMA that 686 receives an ARK to remove any hyphens found in it. As a result, like 687 the NMAH, hyphens are "identity inert" in comparing ARKs for 688 equivalence. For example, the following ARKs are equivalent for 689 purposes of comparison and ARK service access: 691 ark:/12025/65-4-xz-321 692 http://sneezy.dopey.com/ark:/12025/654--xz32-1 693 ark:/12025/654xz321 695 The `%' character is reserved for %-encoding all other octets that 696 would appear in the ARK string, in the same manner as for URIs [URI]. 697 A %-encoded octet consists of a `%' followed by two hex digits; for 698 example, "%7d" stands in for `}'. Lower case hex digits are 699 preferred to reduce the chances of false acronym recognition; thus it 700 is better to use "%acT" instead of "%ACT". The character `%' itself 701 must be represented using "%25". As with URNs, %-encoding permits 702 ARKs to support legacy namespaces (e.g., ISBN, ISSN, SICI) that have 703 less restricted character repertoires [URNBIB]. 705 2.7. Normalization and Lexical Equivalence 707 To determine if two or more ARKs identify the same object, the ARKs 708 are compared for lexical equivalence after first being normalized. 709 Since ARK strings may appear in various forms (e.g., having different 710 NMAHs), normalizing them minimizes the chances that comparing two ARK 711 strings for equality will fail unless they actually identify 712 different objects. In a specified-host ARK (one having an NMAH), the 713 NMAH never participates in such comparisons. 715 Normalization of an ARK for the purpose of octet-by-octet equality 716 comparison with another ARK consists of four steps. First, any upper 717 case letters in the "ark:" label and the two characters following a 718 `%' are converted to lower case. The case of all other letters in 719 the ARK string must be preserved. Second, any NMAH part is removed 720 (everything from an initial "http://" up to the next slash) and all 721 hyphens are removed. 723 Third, structural characters (slash and period) are normalized. 724 Initial and final occurrences are removed, and two structural 725 characters in a row (e.g., // or ./) are replaced by the first 726 character, iterating until each occurrence has at least one non- 727 structural character on either side. Finally, if there are any 728 components with a period on the left and a slash on the right, either 729 the component and the preceding period must be moved to the end of 730 the Name part or the ARK must be thrown out as malformed. 732 The fourth and final step is to arrange the suffixes in ASCII 733 collating sequence (that is, to sort them) and to remove duplicate 734 suffixes, if any. It is also permissible to throw out ARKs for which 735 the suffixes are not sorted. 737 The resulting ARK string is now normalized. Comparisons between 738 normalized ARKs are case-sensitive, meaning that upper case letters 739 are considered different from their lower case counterparts. 741 To keep ARK string variation to a minimum, no reserved ARK characters 742 should be %-encoded unless it is deliberately to conceal their 743 reserved meanings. No non-reserved ARK characters should ever be 744 %-encoded. Finally, no %-encoded character should ever appear in an 745 ARK in its decoded form. 747 3. Naming Considerations 749 The most important threats faced by persistence providers include 750 such things as funding loss, natural disaster, political and social 751 upheaval, processing faults, and errors in human oversight. There is 752 nothing that an identifer scheme can do about such things. Still, a 753 few observed identifier failures and inconveniences can be traced 754 back to naming practices that we now know to be less than optimal for 755 persistence. 757 3.1. ARKS Embedded in Language 759 The ARK has different goals from the URI, so it has different 760 character set requirements. Because linguistic constructs imperil 761 persistence, for ARKs non-ASCII character support is unimportant. 762 ARKs and URIs share goals of transcribability and transportability 763 within web documents, so characters are required to be visible, non- 764 conflicting with HTML/XML syntax, and not subject to tampering during 765 transmission across common transport gateways. Add the goal of 766 making an undelimited ARK recognizable in running prose, as in 767 ark:/12025/=@_22*$, and certain punctuation characters (e.g., comma, 768 period) end up being excluded from the ARK lest the end of a phrase 769 or sentence be mistaken for part of the ARK. 771 This consideration has more direct effect on ARK usability in a 772 natural language context than it has on ARK persistence. The same is 773 true of the rule preventing hyphens from having lexical significance. 774 It is fine to publish ARKs with hyphens in them (e.g., such as the 775 output of UUID/GUID generators), but the uniform treatment of hyphens 776 as insignificant reduces the possibility of users transcribing 777 identifiers that will have been broken through unpredictable 778 hyphenation by word processors. Any measure that reduces user 779 irritation with an identifier will increase its chances of survival. 781 3.2. Objects Should Wear Their Identifiers 783 A valuable technique for provision of persistent objects is to try to 784 arrange for the complete identifier to appear on, with, or near its 785 retrieved object. An object encountered at a moment in time when its 786 discovery context has long since disappeared could then easily be 787 traced back to its metadata, to alternate versions, to updates, etc. 788 This has seen reasonable success, for example, in book publishing and 789 software distribution. An identifier string only has meaning when 790 its association is known, and this a very sure, simple, and low-tech 791 method of reminding everyone exactly what that association is. 793 3.3. Names are Political, not Technological 795 If persistence is the goal, a deliberate local strategy for 796 systematic name assignment is crucial. Names must be chosen with 797 great care. Poorly chosen and managed names will devastate any 798 persistence strategy, and they do not discriminate by identifier 799 scheme. Whether a mistakenly re-assigned name is a URN, DOI, PURL, 800 URL, or ARK, the damage - failed access and confusion - is not 801 mitigated more in one scheme than in another. Conversely, in-house 802 efforts to manage names responsibly will go much further towards 803 safeguarding persistence than any choice of naming scheme or name 804 resolution technology. 806 Branding (e.g., at the corporate or departmental level) is important 807 for funding and visibility, but substrings representing brands and 808 organizational names should be given a wide berth except when 809 absolutely necessary in the hostname (the identity-inert) part of the 810 ARK. These substrings are not only unstable because organizations 811 change frequently, but they are also dangerous because successor 812 organizations often have political or legal reasons to actively 813 suppress predecessor names and brands. Any measure that reduces the 814 chances of future political or legal pressure on an identifier will 815 decrease the chances that our descendants will be obliged to 816 deliberately break it. 818 3.4. Choosing a Hostname or NMA 820 Hostnames appearing in any identifier meant to be persistent must be 821 chosen with extra care. The tendency in hostname selection has 822 traditionally been to choose a token with recognizable attributes, 823 such as a corporate brand, but that tendency wreaks havoc with 824 persistence that is supposed to outlive brands, corporations, subject 825 classifications, and natural language semantics (e.g., what did the 826 three letters "gay" mean in 1958, 1978, and 1998?). Today's 827 recognized and correct attributes are tomorrow's stale or incorrect 828 attributes. In making hostnames (any names, actually) long-term 829 persistent, it helps to eliminate recognizable attributes to the 830 extent possible. This affects selection of any name based on URLs, 831 including PURLs and the explicitly disposable NMAHs. 833 There is no excuse for a provider that manages its internal names 834 impeccably not to exercise the same care in choosing what could be an 835 exceptionally durable hostname, especially if it would form the 836 prefix for all the provider's URL-based external names. Registering 837 an opaque hostname in the ".org" or ".net" domain would not be a bad 838 start. Another way is to publish your ARKs with an organizational 839 domain name that will be mapped by DNS to an appropriate NMA host. 840 This makes for shorter names with less branding vulnerability. 842 It is a mistake to think that hostnames are inherently unstable. If 843 you require brand visibility, that may be a fact of life. But things 844 are easier if yours is the brand of long-lived cultural memory 845 institution such as a national or university library or archive. 846 Well-chosen hostnames from organizations that are sheltered from the 847 direct effects of a volatile marketplace can easily provide longer- 848 lived global resolvers than the domain names explicitly or implicitly 849 used as starting points for global resolution by indirection-based 850 persistent identifier schemes. For example, it is hard to imagine 851 circumstances under which the Library of Congress' domain name would 852 disappear sooner than, say, "handle.net". 854 For smaller libraries, archives, and preservation organizations, 855 there is a natural concern about whether they will be able to keep 856 their web servers and domain names in the face of uncertain funding. 857 One option is to form or join a consortium of like-minded 858 organizations with the purpose of providing mutual preservation 859 support. The first goal of such a consortium would be to perpetually 860 rent a hostname on which to establish a web server that simply 861 redirects incoming member organization requests to the appropriate 862 member server; using ARKs, for example, a 150-member consortium could 863 run a very small server (24x7) that contained nothing more than 150 864 rewrite rules in its configuration file. Even more helpful would be 865 additional consortial support for a member organization that was 866 unable to continue providing services and needed to find a successor 867 archival organization. This would be a low-cost, low-tech way to 868 publish ARKs (or URLs) under highly persistent hostnames. 870 There are no obvious reasons why the organizations registering DNS 871 names, URN Namespaces, and DOI publisher IDs should have among them 872 one that is intrinsically more fallible than the next. Moreover, it 873 is a misconception that the demise of DNS and of HTTP need adversely 874 affect the persistence of URLs. At such a time, certainly URLs from 875 the present day might not then be actionable by our present-day 876 mechanisms, but resolution systems for future non-actionable URLs are 877 no harder to imagine than resolution systems for present-day non- 878 actionable URNs and DOIs. There is no more stable a namespace than 879 one that is dead and frozen, and that would then characterize the 880 space of names bearing the "http://" prefix. It is useful to 881 remember that just because hostnames have been carelessly chosen in 882 their brief history does not mean that they are unsuitable in NMAHs 883 (and URLs) intended for use in situations demanding the highest level 884 of persistence available in the Internet environment. A well-planned 885 name assignment strategy is everything. 887 3.5. Assigners of ARKs 889 A Name Assigning Authority (NAA) is an organization that creates (or 890 delegates creation of) long-term associations between identifiers and 891 information objects. Examples of NAAs include national libraries, 892 national archives, and publishers. An NAA may arrange with an 893 external organization for identifier assignment. The US Library of 894 Congress, for example, allows OCLC (the Online Computer Library 895 Center, a major world cataloger of books) to create associations 896 between Library of Congress call numbers (LCCNs) and the books that 897 OCLC processes. A cataloging record is generated that testifies to 898 each association, and the identifier is included by the publisher, 899 for example, in the front matter of a book. 901 An NAA does not so much create an identifier as create an 902 association. The NAA first draws an unused identifier string from 903 its namespace, which is the set of all identifiers under its control. 904 It then records the assignment of the identifier to an information 905 object having sundry witnessed characteristics, such as a particular 906 author and modification date. A namespace is usually reserved for an 907 NAA by agreement with recognized community organizations (such as 908 IANA and ISO) that all names containing a particular string be under 909 its control. In the ARK an NAA is represented by the Name Assigning 910 Authority Number (NAAN). 912 The ARK namespace reserved for an NAA is the set of names bearing its 913 particular NAAN. For example, all strings beginning with 914 "ark:/12025/" are under control of the NAA registered under 12025, 915 which might be the National Library of Finland. Because each NAA has 916 a different NAAN, names from one namespace cannot conflict with those 917 from another. Each NAA is free to assign names from its namespace 918 (or delegate assignment) according to its own policies. These 919 policies must be documented in a manner similar to the declarations 920 required for URN Namespace registration [URNNID]. 922 To register for a NAAN, please read about the mapping authority 923 discovery file in the next section and send email to ark@cdlib.org. 925 3.6. NAAN Namespace Management 927 Every NAA must have a namespace management strategy. A time-honored 928 technique is to hierarchically partition a namespace into 929 subnamespaces using prefixes that guarantee non-collision of names in 930 different partition. This practice is strongly encouraged for all 931 NAAs, especially when subnamespace management will be delegated to 932 other departments, units, or projects within an organization. For 933 example, with a NAAN that is assigned to a university and managed by 934 its main library, care should be taken to reserve semantically opaque 935 prefixes that will set aside large parts of the unused namespace for 936 future assignments. Prefix-based partition management is an 937 important responsibility of the NAA. 939 This sort of delegation by prefix is well-used in the formation of 940 DNS names and ISBN identifiers. An important difference is that in 941 the former, the hierarchy is deliberately exposed and in the latter 942 it is hidden. Rather than using lexical boundary markers such as the 943 period (`.') found in domain names, the ISBN uses a publisher prefix 944 but doesn't disclose where the prefix ends and the publisher's 945 assigned name begins. This practice of non-disclosure, borrowed from 946 the ISBN and ISSN schemes, is encouraged in assigning ARKs, because 947 it reduces the visibility of an assertion that is probably not 948 important now and may become a vulnerability later. 950 Reasonable prefixes for assigned names usually consist of consonants 951 and digits and are 1-5 characters in length. For example, the 952 constant prefix "x9t" might be delegated to a book digitization 953 project that creates identifiers such as 955 http://444.berkeley.edu/ark:/28722/x9t38rk45c 957 If longevity is the goal, it is important to keep the prefixes free 958 of recognizable semantics; for example, using an acronym representing 959 a project or a department is discouraged. At the same time, you may 960 wish to set aside a subnamespace for testing purposes under a prefix 961 such as "fk..." that can serve as a visual clue and reminder to 962 maintenance staff that this "fake" identifier was never published. 964 There are other measures one can take to avoid user confusion, 965 transcription errors, and the appearance of accidental semantics when 966 creating identifiers. If you are generating identifiers 967 automatically, pure numeric identifiers are likeley to be 968 semantically opaque enough, but it's probably useful to avoid leading 969 zeroes because some users mistakenly treat them as optional, thinking 970 (arithmetically) that they don't contribute to the "value" of the 971 identifier. 973 If you need lots of identifiers and you don't want them to get too 974 long, you can mix digits with consonants (but avoid vowels since they 975 might accidentally spell words) to get more identifiers without 976 increasing the string length. In this case you may not want more 977 than a two letters in a row because it reduces the chance of 978 generating acronyms. Generator tools such as [NOID] provide support 979 for these sorts of identifiers, and can also add a computed check 980 character as a guarantee against the most common transcription 981 errors. 983 3.7. Sub-Object Naming 985 As mentioned previously, semantically opaque identifiers are very 986 useful for long-term naming of abstract objects, however, it may be 987 appropriate to extend these names with less opaque extensions that 988 reference contemporary service entry points (sub-objects) in support 989 of the object. Sub-object extensions beginning with a digit or 990 underscore (`_') are reserved for the possibilty of developing a 991 future registry of canonical service points (e.g., numeric references 992 to versions, formats, languages, etc). 994 4. Finding a Name Mapping Authority 996 In order to derive an actionable identifier (these days, a URL) from 997 an ARK, a hostport (hostname or hostname plus port combination) for a 998 working Name Mapping Authority (NMA) must be found. An NMA is a 999 service that is able to respond to the three basic ARK service 1000 requests. Relying on registration and client-side discovery, NMAs 1001 make known which NAAs' identifiers they are willing to service. 1003 Upon encountering an ARK, a user (or client software) looks inside it 1004 for the optional NMAH part (the hostport of the NMA's ARK service). 1005 If it contains an NMAH that is working, this NMAH discovery step may 1006 be skipped; the NMAH effectively uses the beginning of an ARK to 1007 cache the results of a prior mapping authority discovery process. If 1008 a new NMAH needs to found, the client looks inside the ARK again for 1009 the NAAN (Name Assigning Authority Number). Querying a global 1010 database, it then uses the NAAN to look up all current NMAHs that 1011 service ARKs issued by the identified NAA. The global database is 1012 key, and two specific methods for querying it are given in this 1013 section. 1015 In the interests of long-term persistence, however, ARK mechanisms 1016 are first defined in high-level, protocol-independent terms so that 1017 mechanisms may evolve and be replaced over time without compromising 1018 fundamental service objectives. Either or both specific methods 1019 given here may eventually be supplanted by better methods since, by 1020 design, the ARK scheme does not depend on a particular method, but 1021 only on having some method to locate an active NMAH. 1023 At the time of issuance, at least one NMAH for an ARK should be 1024 prepared to service it. That NMA may or may not be administered by 1025 the Name Assigning Authority (NAA) that created it. Consider the 1026 following hypothetical example of providing long-term access to a 1027 cancer research journal. The publisher wishes to turn a profit and 1028 the National Library of Medicine wishes to preserve the scholarly 1029 record. An agreement might be struck whereby the publisher would act 1030 as the NAA and the national library would archive the journal issue 1031 when it appears, but without providing direct access for the first 1032 six months. During the first six months of peak commercial 1033 viability, the publisher would retain exclusive delivery rights and 1034 would charge access fees. Again, by agreement, both the library and 1035 the publisher would act as NMAs, but during that initial period the 1036 library would redirect requests for issues less than six months old 1037 to the publisher. At the end of the waiting period, the library 1038 would then begin servicing requests for issues older than six months 1039 by tapping directly into its own archives. Meanwhile, the publisher 1040 might routinely redirect incoming requests for older issues to the 1041 library. Long-term access is thereby preserved, and so is the 1042 commercial incentive to publish content. 1044 Although it will be common for an NAA also to run an NMA service, it 1045 is never a requirement. Over time NAAs and NMAs will come and go. 1046 One NMA will succeed another, and there might be many NMAs serving 1047 the same ARKs simultaneously (e.g., as mirrors or as competitors). 1048 There might also be asymmetric but coordinated NMAs as in the 1049 library-publisher example above. 1051 4.1. Looking Up NMAHs in a Globally Accessible File 1053 This subsection describes a way to look up NMAHs using a simple name 1054 authority table represented as a plain text file. For efficient 1055 access the file may be stored in a local filesystem, but it needs to 1056 be reloaded periodically to incorporate updates. It is not expected 1057 that the size of the file or frequency of update should impose an 1058 undue maintenance or searching burden any time soon, for even 1059 primitive linear search of a file with ten-thousand NAAs is a 1060 subsecond operation on modern server machines. The proposed file 1061 strategy is similar to the /etc/hosts file strategy that supported 1062 Internet host address lookup for a period of years before the advent 1063 of DNS. 1065 The name authority table file is updated on an ongoing basis and is 1066 available for copying over the internet from the California Digital 1067 Library at http://www.cdlib.org/inside/diglib/ark/natab and from a 1068 number of mirror sites. The file contains comment lines (lines that 1069 begin with `#') explaining the format and giving the file's 1070 modification time, reloading address, and NAA registration 1071 instructions. There is even a Perl script that processes the file 1072 embedded in the file's comments. As of February 2006, currently 1073 registered Name Assigning Authorities are: 1075 12025 National Library of Medicine 1076 12026 Library of Congress 1077 12027 National Agriculture Library 1078 13030 California Digital Library 1079 13038 World Intellectual Property Organization 1080 20775 University of California San Diego 1081 29114 University of California San Francisco 1082 28722 University of California Berkeley 1083 21198 University of California Los Angeles 1084 15230 Rutgers University 1085 13960 Internet Archive 1086 64269 Digital Curation Centre 1087 62624 New York University 1088 67531 University of North Texas 1089 27927 Ithaka Electronic-Archiving Initiative 1090 12148 Bibliothque nationale de France / National Library of France 1091 88435 Princeton University 1092 78428 University of Washington 1093 89901 Archives of Region of Vstra Gtaland and City of Gothenburg, Sweden 1094 80444 Northwest Digital Archives 1095 25593 Emory University 1097 A snapshot of the name authority table file appears in an appendix. 1099 4.2. Looking up NMAHs Distributed via DNS 1101 This subsection introduces a method for looking up NMAHs that is 1102 based on the method for discovering URN resolvers described in 1103 [NAPTR]. It relies on querying the DNS system already installed in 1104 the background infrastructure of most networked computers. A query 1105 is submitted to DNS asking for a list of resolvers that match a given 1106 NAAN. DNS distributes the query to the particular DNS servers that 1107 can best provide the answer, unless the answer can be found more 1108 quickly in a local DNS cache as a side-effect of a recent query. 1109 Responses come back inside Name Authority Pointer (NAPTR) records. 1110 The normal result is one or more candidate NMAHs. 1112 In its full generality the [NAPTR] algorithm ambitiously accommodates 1113 a complex set of preferences, orderings, protocols, mapping services, 1114 regular expression rewriting rules, and DNS record types. This 1115 subsection proposes a drastic simplification of it for the special 1116 case of ARK mapping authority discovery. The simplified algorithm is 1117 called Maptr. It uses only one DNS record type (NAPTR) and restricts 1118 most of its field values to constants. The following hypothetical 1119 excerpt from a DNS data file for the NAAN known as 12026 shows three 1120 example NAPTR records ready to use with the Maptr algorithm. 1122 12026.ark.arpa. 1123 ;; US Library of Congress 1124 ;; order pref flags service regexp replacement 1125 IN NAPTR 0 0 "h" "ark" "USLC" lhc.nlm.nih.gov:8080 1126 IN NAPTR 0 0 "h" "ark" "USLC" foobar.zaf.org 1127 IN NAPTR 0 0 "h" "ark" "USLC" sneezy.dopey.com 1129 All the fields are held constant for Maptr except for the "flags", 1130 "regexp", and "replacement" fields. The "service" field contains the 1131 constant value "ark" so that NAPTR records participating in the Maptr 1132 algorithm will not be confused with other NAPTR records. The "order" 1133 and "pref" fields are held to 0 (zero) and otherwise ignored for now; 1134 the algorithm may evolve to use these fields for ranking decisions 1135 when usage patterns and local administrative needs are better 1136 understood. 1138 When a Maptr query returns a record with a flags field of "h" (for 1139 hostport, a Maptr extension to the NAPTR flags), the replacement 1140 field contains the NMAH (hostport) of an ARK service provider. When 1141 a query returns a record with a flags field of "" (the empty string), 1142 the client needs to submit a new query containing the domain name 1143 found in the replacement field. This second sort of record exploits 1144 the distributed nature of DNS by redirecting the query to another 1145 domain name. It looks like this. 1147 12345.ark.arpa. 1148 ;; Digital Library Consortium 1149 ;; order pref flags service regexp replacement 1150 IN NAPTR 0 0 "" "ark" "" dlc.spct.org. 1152 Here is the Maptr algorithm for ARK mapping authority discovery. In 1153 it replace with the NAAN from the ARK for which an NMAH is 1154 sought. 1156 (1) Initialize the DNS query: type=NAPTR, 1157 query=.ark.arpa. 1159 (2) Submit the query to DNS and retrieve (NAPTR) records, 1160 discarding any record that does not have "ark" for the service 1161 field. 1163 (3) All remaining records with a flags fields of "h" contain 1164 candidate NMAHs in their replacement fields. Set them aside, if 1165 any. 1167 (4) Any record with an empty flags field ("") has a replacement 1168 field containing a new domain name to which a subsequent query 1169 should be redirected. For each such record, set 1170 query= then go to step (2). When all such records 1171 have been recursively exhausted, go to step (5). 1173 (5) All redirected queries have been resolved and a set of 1174 candidate NMAHs has been accumulated from steps (3). If there 1175 are zero NMAHs, exit - no mapping authority was found. If there 1176 is one or more NMAH, choose one using any criteria you wish, 1177 then exit. 1179 A Perl script that implements this algorithm is included here. 1181 #!/depot/bin/perl 1183 use Net::DNS; # include simple DNS package 1184 my $qtype = "NAPTR"; # initialize query type 1185 my $naa = shift; # get NAAN script argument 1186 my $mad = new Net::DNS::Resolver; # mapping authority discovery 1188 &maptr("$naa.ark.arpa"); # call maptr - that's it 1190 sub maptr { # recursive maptr algorithm 1191 my $dname = shift; # domain name as argument 1192 my ($rr, $order, $pref, $flags, $service, $regexp, 1193 $replacement); 1194 my $query = $mad->query($dname, $qtype); 1195 return # non-productive query 1196 if (! $query || ! $query->answer); 1197 foreach $rr ($query->answer) { 1198 next # skip records of wrong type 1199 if ($rr->type ne $qtype); 1200 ($order, $pref, $flags, $service, $regexp, 1201 $replacement) = split(/\s/, $rr->rdatastr); 1202 if ($flags eq "") { 1203 &maptr($replacement); # recurse 1204 } elsif ($flags eq "h") { 1205 print "$replacement\n"; # candidate NMAH 1206 } 1207 } 1208 } 1210 The global database thus distributed via DNS and the Maptr algorithm 1211 can easily be seen to mirror the contents of the Name Authority Table 1212 file described in the previous section. 1214 5. Generic ARK Service Definition 1216 An ARK request's output is delivered information; examples include 1217 the object itself, a policy declaration (e.g., a promise of support), 1218 a descriptive metadata record, or an error message. The experience 1219 of object delivery is expected to be an evolving mix of information 1220 that reflects changing service expectations and technology 1221 requirements; contemporary examples include such things as an object 1222 summary and component links formatted for human consumption. ARK 1223 services must be couched in high-level, protocol-independent terms if 1224 persistence is to outlive today's networking infrastructural 1225 assumptions. The high-level ARK service definitions listed below are 1226 followed in the next section by a concrete method (one of many 1227 possible methods) for delivering these services with today's 1228 technology. 1230 5.1. Generic ARK Access Service (access, location) 1232 Returns (a copy of) the object or a redirect to the same, although a 1233 sensible object proxy may be substituted. Examples of sensible 1234 substitutes include, 1236 - a table of contents instead of a large complex document, 1237 - a home page instead of an entire web site hierarchy, 1238 - a rights clearance challenge before accessing protected data, 1239 - directions for access to an offline object (e.g., a book), 1240 - a description of an intangible object (a disease, an event), or 1241 - an applet acting as "player" for a large multimedia object. 1243 May also return a discriminated list of alternate object locators. 1244 If access is denied, returns an explanation of the object's current 1245 (perhaps permanent) inaccessibility. 1247 5.2. Generic Policy Service (permanence, naming, etc.) 1249 Returns declarations of policy and support commitments for given 1250 ARKs. Declarations are returned in either a structured metadata 1251 format or a human readable text format; sometimes one format may 1252 serve both purposes. Policy subareas may be addressed in separate 1253 requests, but the following areas should should be covered: object 1254 permanence, object naming, object fragment addressing, and 1255 operational service support. 1257 The permanence declaration for an object is a rating defined with 1258 respect to an identified permanence provider (guarantor), which will 1259 be the NMA. It may include the following aspects. 1261 (a) "object availability" - whether and how access to the object 1262 is supported (e.g., online 24x7, or offline only), 1264 (b) "identifier validity" - under what conditions the identifier 1265 will be or has been re-assigned, 1267 (c) "content invariance" - under what conditions the content of 1268 the object is subject to change, and 1270 (d) "change history" - access to corrections, migrations, and 1271 revisions, whether through links to the changed objects 1272 themselves or through a document summarizing the change history 1274 One approach to a permanence rating framework, conceived 1275 independently from ARKs, is given in [NLMPerm]. Under ongoing 1276 development and limited deployment at the US National Library of 1277 Medicine, it identifies the following "permanence levels": 1279 Not Guaranteed: No commitment has been made to retain this 1280 resource. It could become unavailable at any time. Its 1281 identifier could be changed. 1283 Permanent: Dynamic Content: A commitment has been made to keep 1284 this resource permanently available. Its identifier will always 1285 provide access to the resource. Its content could be revised or 1286 replaced. 1288 Permanent: Stable Content: A commitment has been made to keep 1289 this resource permanently available. Its identifier will always 1290 provide access to the resource. Its content is subject only to 1291 minor corrections or additions. 1293 Permanent: Unchanging Content: A commitment has been made to 1294 keep this resource permanently available. Its identifier will 1295 always provide access to the resource. Its content will not 1296 change. 1298 Naming policy for an object includes an historical description of the 1299 NAA's (and its successor NAA's) policies regarding differentiation of 1300 objects. Since it the NMA who responds to requests for policy 1301 statements, it is useful for the NMA to be able to produce or 1302 summarize these historical NAA documents. Naming policy may include 1303 the following aspects. 1305 (i) "similarity" - (or "unity") the limit, defined by the NAA, 1306 to the level of dissimilarity beyond which two similar objects 1307 warrant separate identifiers but before which they share one 1308 single identifier, and 1310 (ii) "granularity" - the limit, defined by the NAA, to the level 1311 of object subdivision beyond which sub-objects do not warrant 1312 separately assigned identifiers but before which sub-objects are 1313 assigned separate identifiers. 1315 Subnaming policy for an object describes the qualifiers that the NMA, 1316 in fulfilling its ongoing and evolving service obligations, allows as 1317 extensions to an NAA-assigned ARK. To the conceptual object that the 1318 NAA named with an ARK, the NMA may add component access points and 1319 derivatives (e.g., format migrations in aid of preservation) in order 1320 to provide both basic and value-added services. 1322 Addressing policy for an object includes a description of how, during 1323 access, object components (e.g., paragraphs, sections) or views 1324 (e.g., image conversions) may or may not be "addressed", in other 1325 words, how the NMA permits arguments or parameters to modify the 1326 object delivered as the result of an ARK request. If supported, 1327 these sorts of operations would provide things like byte-ranged 1328 fragment delivery and open-ended format conversions, or any set of 1329 possible transformations that would be too numerous to list or to 1330 identify with separately assigned ARKs. 1332 Operational service support policy includes a description of general 1333 operational aspects of the NMA service, such as after-hours staffing 1334 and trouble reporting procedures. 1336 5.3. Generic Description Service 1338 Returns a description of the object. Descriptions are returned in 1339 either a structured metadata format or a human readable text format; 1340 sometimes one format may serve both purposes. A description must at 1341 a minimum answer the who, what, when, and where questions concerning 1342 an expression of the object. Standalone descriptions should be 1343 accompanied by the modification date and source of the description 1344 itself. May also return discriminated lists of ARKs that are related 1345 to the given ARK. 1347 6. Overview of The HTTP URL Mapping Protocol (THUMP) 1349 The HTTP URL Mapping Protocol (THUMP) is a way of taking a key (a 1350 kind of identifier) and asking such questions as, what information 1351 does this identify and how permanent is it? [THUMP] is in fact one 1352 specific method under development for delivering ARK services. The 1353 protocol runs over HTTP to exploit the web browser's current pre- 1354 eminence as user interface to the Internet. THUMP is designed so 1355 that a person can enter ARK requests directly into the location field 1356 of current browser interfaces. Because it runs over HTTP, THUMP can 1357 be simulated and tested within keyboard-based [TELNET] sessions. 1359 The asker (a person or client program) starts with an identifier, 1360 such as an ARK or a URL. The identifier reveals to the asker (or 1361 allows the asker to infer) the Internet host name and port number of 1362 a server system that responds to questions. Here, this is just the 1363 NMAH that is obtained by inspection and possibly lookup based on the 1364 ARK's NAAN. The asker then sets up an HTTP session with the server 1365 system, sends a question via a THUMP request (contained within an 1366 HTTP request), receives an answer via a THUMP response (contained 1367 within an HTTP response), and closes the session. That concludes the 1368 connected portion of the protocol. 1370 A THUMP request is a string of characters beginning with a `?' 1371 (question mark) that is appended to the identifier string. The 1372 resulting string is sent as an argument to HTTP's GET command. 1373 Request strings too long for GET may be sent using HTTP's POST 1374 command. The three most common requests correspond to three 1375 degenerate special cases that keep the user's learning and typing 1376 burden low. First, a simple key with no request at all is the same 1377 as an ordinary access request. Thus a plain ARK entered into a 1378 browser's location field behaves much like a plain URL, and returns 1379 access to the primary identified object, for instance, an HTML 1380 document. 1382 The second special case is a minimal ARK description request string 1383 consisting of just "?". For example, entering the string, 1385 ark.nlm.nih.gov/12025/psbbantu? 1387 into the browser's location field directly precipitates a request for 1388 a metadata record describing the object identified by 1389 ark:/12025/psbbantu. The browser, unaware of THUMP, prepares and 1390 sends an HTTP GET request in the same manner as for a URL. THUMP is 1391 designed so that the response (indicated by the returned HTTP content 1392 type) is normally displayed, whether the output is structured for 1393 machine processing (text/plain) or formatted for human consumption 1394 (text/html). 1396 In the following example THUMP session, each line has been annotated 1397 to include a line number and whether it was the client or server that 1398 sent it. Without going into much depth, the session has four pieces 1399 separated from each other by blank lines: the client's piece (lines 1400 1-3), the server's HTTP/THUMP response headers (4-7), and the body of 1401 the server's response (8-17). The first and last lines (1 and 17) 1402 correspond to the client's steps to start the TCP session and the 1403 server's steps to end it, respectively. 1405 1 C: [opens session] 1406 C: GET http://ark.nlm.nih.gov/ark:/12025/psbbantu? HTTP/1.1 1407 C: 1408 S: HTTP/1.1 200 OK 1409 5 S: Content-Type: text/plain 1410 S: THUMP-Status: 0.1 200 OK 1411 S: 1412 S: |set: NLM | 12025/psbbantu? | 20030731 1413 S: | http://ark.nlm.nih.gov/ark:/12025/psbbantu? 1414 10 S: here: 1 | 1 | 1 1415 S: 1416 S: erc: 1417 S: who: Lederberg, Joshua 1418 S: what: Studies of Human Families for Genetic Linkage 1419 15 S: when: 1974 1420 S: where: http://profiles.nlm.nih.gov/BB/A/N/T/U/_/bbantu.pdf 1421 S: [closes session] 1423 The first two server response lines (4-5) above are typical of HTTP. 1424 The next line (6) is peculiar to THUMP, and indicates the THUMP 1425 version and a normal return status. The balance of the response 1426 consists of a record set header (lines 8-10) and a single metadata 1427 record (12-16) that comprises the ARK description service response. 1428 The record set header identifies (8-9) who created the set, what its 1429 title is, when it was created, and where an automated process can 1430 access the set; it ends in a line (10) whose respective sub-elements 1431 indicate that here in this communication the recipient can expect to 1432 find 1 record, starting at the record numbered 1, from a set 1433 consisting of a total of 1 record (i.e., here is the entire set, 1434 consisting of exactly one record). 1436 The returned record (12-16) is in the format of an Electronic 1437 Resource Citation [ERC], which is discussed in more detail in the 1438 next section. For now, note that it contains four elements that 1439 answer the top priority questions regarding an expression of the 1440 object: who played a major role in expressing it, what the 1441 expression was called, when is was created, and where the expression 1442 may be found. This quartet of elements comes up again and again in 1443 ERCs. 1445 The third degenerate special case of an ARK request (and no other 1446 cases will be described in this document) is the string "??", 1447 corresponding to a minimal permanence policy request. It can be seen 1448 in use appended to an ARK (on line 2) in the example session that 1449 follows. 1451 1 C: [opens session] 1452 C: GET http://ark.nlm.nih.gov/ark:/12025/psbbantu?? HTTP/1.1 1453 C: 1454 S: HTTP/1.1 200 OK 1455 5 S: Content-Type: text/plain 1456 S: THUMP-Status: 0.1 200 OK 1457 S: 1458 S: |set: NLM | 12025/psbbantu?? | 20030731 1459 S: | http://ark.nlm.nih.gov/ark:/12025/psbbantu?? 1460 10 S: here: 1 | 1 | 1 1461 S: 1462 S: erc: 1463 S: who: Lederberg, Joshua 1464 S: what: Studies of Human Families for Genetic Linkage 1465 15 S: when: 1974 1466 S: where: http://profiles.nlm.nih.gov/BB/A/N/T/U/_/bbantu.pdf 1467 S: erc-support: 1468 S: who: USNLM 1469 S: what: Permanent, Unchanging Content 1470 20 S: when: 20010421 1471 S: where: http://ark.nlm.nih.gov/yy22948 1472 S: [closes session] 1474 Again, a single metadata record (lines 12-21) is returned, but it 1475 consists of two segments. The first segment (12-16) gives the same 1476 basic citation information as in the previous example. It is 1477 returned in order to establish context for the persistence 1478 declaration in the second segment (17-21). 1480 Each segment in an ERC tells a different story relating to the 1481 object, so although the same four questions (elements) appear in 1482 each, the answers depend on the segment's story type. While the 1483 first segment tells the story of an expression of the object, the 1484 second segment tells the story of the support commitment made to it: 1486 who made the commitment, what the nature of the commitment was, when 1487 it was made, and where a fuller explanation of the commitment may be 1488 found. 1490 7. Overview of Electronic Resource Citations (ERCs) 1492 An Electronic Resource Citation (or ERC, pronounced e-r-c) [ERC] is a 1493 simple, compact, and printable record designed to hold data 1494 associated with an information resource. By design, the ERC is a 1495 metadata format that balances the needs for expressive power, very 1496 simple machine processing, and direct human manipulation. 1498 A founding principle of the ERC is that direct human contact with 1499 metadata will be a necessary and sufficient condition for the near 1500 term rapid development of metadata standards, systems, and services. 1501 Thus the machine-processable ERC format must only minimally strain 1502 people's ability to read, understand, change, and transmit ERCs 1503 without their relying on intermediation with specialized software 1504 tools. The basic ERC needs to be succinct, transparent, and 1505 trivially parseable by software. 1507 In the current Internet, it is natural seriously to consider using 1508 XML as an exchange format because of predictions that it will obviate 1509 many ad hoc formats and programs, and unify much of the world's 1510 information under one reliable data structuring discipline that is 1511 easy to generate, verify, parse, and render. It appears, however, 1512 that XML is still only catching on after years of standards work and 1513 implementation experience. The reasons for it are unclear, but for 1514 now very simple XML interpretation is still out of reach. Another 1515 important caution is that XML structures are hard on the eyeballs, 1516 taking up an amount of display (and page) space that significantly 1517 exceeds that of traditional formats. Until these conflicts with ERC 1518 principle are resolved, XML is not a first choice for representing 1519 ERCs. Borrowing instead from the data structuring format that 1520 underlies the successful spread of email and web services, the first 1521 ERC format uses [ANVL], which is based on email and HTTP headers 1522 [RFC822]. There is a naturalness to ANVL's label-colon-value format 1523 (seen in the previous section) that barely needs explanation to a 1524 person beginning to enter ERC metadata. 1526 Besides simplicity of ERC system implementation and data entry 1527 mechanics, ERC semantics (what the record and its constituent parts 1528 mean) must also be easy to explain. ERC semantics are based on a 1529 reformulation and extension of the Dublin Core [DCORE] hypothesis, 1530 which suggests that the fifteen Dublin Core metadata elements have a 1531 key role to play in cross-domain resource description. The ERC 1532 design recognizes that the Dublin Core's primary contribution is the 1533 international, interdisciplinary consensus that identified fifteen 1534 semantic buckets (element categories), regardless of how they are 1535 labeled. The ERC then adds a definition for a record and some 1536 minimal compliance rules. In pursuing the limits of simplicity, the 1537 ERC design combines and relabels some Dublin Core buckets to isolate 1538 a tiny kernel (subset) of four elements for basic cross-domain 1539 resource description. 1541 For the cross-domain kernel, the ERC uses the four basic elements - 1542 who, what, when, and where - to pretend that every object in the 1543 universe can have a uniform minimal description. Each has a name or 1544 other identifier, a location, some responsible person or party, and a 1545 date. It doesn't matter what type of object it is, or whether one 1546 plans to read it, interact with it, smoke it, wear it, or navigate 1547 it. Of course, this approach is flawed because uniformity of 1548 description for some object types requires more semantic contortion 1549 and sacrifice than for others. That is why at the beginning of this 1550 document, the ARK was said to be suited to objects that accommodate 1551 reasonably regular electronic description. 1553 While insisting on uniformity at the most basic level provides 1554 powerful cross-domain leverage, the semantic sacrifice is great for 1555 many applications. So the ERC also permits a semantically rich and 1556 nuanced description to co-exist in a record along with a basic 1557 description. In that way both sophisticated and naive recipients of 1558 the record can extract the level of meaning from it that best suits 1559 their needs and abilities. Key to unlocking the richer description 1560 is a controlled vocabulary of ERC record types (not explained in this 1561 document) that permit knowledgeable recipients to apply defined sets 1562 of additional assumptions to the record. 1564 7.1. ERC Syntax 1566 An ERC record is a sequence of metadata elements ending in a blank 1567 line. An element consists of a label, a colon, and an optional 1568 value. Here is an example of a record with five elements. 1570 erc: 1571 who: Gibbon, Edward 1572 what: The Decline and Fall of the Roman Empire 1573 when: 1781 1574 where: http://www.ccel.org/g/gibbon/decline/ 1576 A long value may be folded (continued) onto the next line by 1577 inserting a newline and indenting the next line. A value can be thus 1578 folded across multiple lines. Here are two example elements, each 1579 folded across four lines. 1581 who/created: University of California, San Francisco, AIDS 1582 Program at San Francisco General Hospital | University 1583 of California, San Francisco, Center for AIDS Prevention 1584 Studies 1585 what/Topic: 1586 Heart Attack | Heart Failure 1587 | Heart 1588 Diseases 1590 An element value folded across several lines is treated as if the 1591 lines were joined together on one long line. For example, the second 1592 element from the previous example is considered equivalent to 1594 what/Topic: Heart Attack | Heart Failure | Heart Diseases 1596 An element value may contain multiple values, each one separated from 1597 the next by a `|' (pipe) character. The element from the previous 1598 example contains three values. 1600 For annotation purposes, any line beginning with a `#' (hash) 1601 character is treated as if it were not present; this is a "comment" 1602 line (a feature not available in email or HTTP headers). For 1603 example, the following element is spread across four lines and 1604 contains two values: 1606 what/Topic: 1607 Heart Attack 1608 # | Heart Failure -- hold off until next review cycle 1609 | Heart Diseases 1611 7.2. ERC Stories 1613 An ERC record is organized into one or more distinct segments, where 1614 where each segment tells a story about a different aspect of the 1615 information resource. A segment boundary occurs whenever a segment 1616 label (an element beginning with "erc") is encountered. The basic 1617 label "erc:" introduces the story of an object's expression (e.g., 1618 its publication, installation, or performance). The label "erc- 1619 about:" introduces the story of an object's content (what it is 1620 about) and "erc-support:" introduces the story of a support 1621 commitment made to it. A story segment that concerns the ERC itself 1622 is introduced by the label "erc-from:". It is an important segment 1623 that tells the story of the ERC's provenance. Elements beginning 1624 with "erc" are reserved for segment labels and their associated story 1625 types. From an earlier example, here is an ERC with two segments. 1627 erc: 1628 who: Lederberg, Joshua 1629 what: Studies of Human Families for Genetic Linkage 1630 when: 1974 1631 where: http://profiles.nlm.nih.gov/BB/A/N/T/U/_/bbantu.pdf 1632 erc-support: 1633 who: NIH/NLM/LHNCBC 1634 what: Permanent, Unchanging Content 1635 # Note to ops staff: date needs verification. 1636 when: 2001 04 21 1637 where: http://ark.nlm.nih.gov/yy22948 1639 Segment stories are told according to journalistic tradition. While 1640 any number of pertinent elements may appear in a segment, priority is 1641 placed on answering the questions who, what, when, and where at the 1642 beginning of each segment so that readers can make the most important 1643 selection or rejection decisions as soon as possible. To make things 1644 simple, the listed ordering of the questions is maintained in each 1645 segment (as it happens most people who have been exposed to this 1646 story telling technique are already familiar with the above 1647 ordering). 1649 The four questions are answered by using corresponding element 1650 labels. The four element labels can be re-used in each story 1651 segment, but their meaning changes depending on the segment (the 1652 story type) in which they appear. In the example above, "who" is 1653 first used to name a document's author and subsequently used to name 1654 the permanence guarantor (provider). Similarly, "when" first lists 1655 the date of object creation and in the next segment lists the date of 1656 a commitment decision. Four labels appearing across three segments 1657 effectively map to twelve semantically distinct elements. Distinct 1658 element meanings are mapped to Dublin Core elements in a later 1659 section. 1661 7.3. The ERC Anchoring Story 1663 Each ERC contains an anchoring story. It is usually the first 1664 segment labeled "erc:" and it concerns an "anchoring" expression of 1665 the object. An "anchoring" expression is the one that a provider 1666 deemed the most suitable basic referent given the audience and 1667 application for which it produced the ERC. If it sounds like the 1668 provider has great latitude in choosing its anchoring expression, it 1669 is because it does. A typical anchoring story in an ERC for a born- 1670 digital document would be the story of the document's release on a 1671 web site; such a document would then be the anchoring expression. 1673 An anchoring story need not be the central descriptive goal of an ERC 1674 record. For example, a museum provider may create an ERC for a 1675 digitized photograph of a painting but choose to anchor it in the 1676 story of the original painting instead of the story of the electronic 1677 likeness; although the ERC may through other segments prove to be 1678 centrally concerned with describing the electronic likeness, the 1679 provider may have chosen this particular anchoring story in order to 1680 make the ERC visible in a way that is most natural to patrons (who 1681 would find the Mona Lisa under da Vinci sooner than they would find 1682 it under the name of the person who snapped the photograph or scanned 1683 the image). In another example, a provider that creates an ERC for a 1684 dramatic play as an abstract work has the task of describing a piece 1685 of intangible intellectual property. To anchor this abstract object 1686 in the concrete world, if only through a derivative expression, it 1687 makes sense for the provider to choose a suitable printed edition of 1688 the play as the anchoring object expression (to describe in the 1689 anchoring story) of the ERC. 1691 The anchoring story has special rules designed to keep ERC processing 1692 simple and predictable. Each of the four basic elements (who, what, 1693 when, and where) must be present, unless a best effort to supply it 1694 fails. In the event of failure, the element still appears but a 1695 special value (described later) is used to explain the missing value. 1696 While the requirement that each of the four elements be present only 1697 applies to the anchoring story segment, as usual these elements 1698 appear at the beginning of the segment and may only be used in the 1699 prescribed order. A minimal ERC would normally consist of just an 1700 anchoring story and the element quartet, as illustrated in the next 1701 example. 1703 erc: 1704 who: National Research Council 1705 what: The Digital Dilemma 1706 when: 2000 1707 where: http://books.nap.edu/html/digital%5Fdilemma 1709 A minimal ERC can be abbreviated so that it resembles a traditional 1710 compact bibliographic citation that is nonetheless completely machine 1711 processable. The required elements and ordering makes it possible to 1712 eliminate the element labels, as shown here. 1714 erc: National Research Council | The Digital Dilemma | 2000 1715 | http://books.nap.edu/html/digital%5Fdilemma 1717 7.4. ERC Elements 1719 As mentioned, the four basic ERC elements (who, what, when, and 1720 where) take on different specific meanings depending on the story 1721 segment in which they are used. By appearing in each segment, albeit 1722 in different guises, the four elements serve as a valuable mnemonic 1723 device - a kind of checklist - for constructing minimal story 1724 segments from scratch. Again, it is only in the anchoring segment 1725 that all four elements are mandatory. 1727 Here are some mappings between ERC elements and Dublin Core [DCORE] 1728 elements. 1730 Segment ERC Element Equivalent Dublin Core Element 1731 --------- ----------- ------------------------------ 1732 erc who Creator/Contributor/Publisher 1733 erc what Title 1734 erc when Date 1735 erc where Identifier 1736 erc-about who 1737 erc-about what Subject 1738 erc-about when Coverage (temporal) 1739 erc-about where Coverage (spatial) 1741 The basic element labels may also be qualified to add nuances to the 1742 semantic categories that they identify. Elements are qualified by 1743 appending a `/' (slash) and a qualifier term. Often qualifier terms 1744 appear as the past tense form of a verb because it makes re-using 1745 qualifiers among elements easier. 1747 who/published: ... 1748 when/published: ... 1749 where/published: ... 1751 Using past tense verbs for qualifiers also reminds providers and 1752 recipients that element values contain transient assertions that may 1753 have been true once, but that tend to become less true over time. 1754 Recipients that don't understand the meaning of a qualifier can fall 1755 back onto the semantic category (bucket) designated by the 1756 unqualified element label. Inevitably recipients (people and 1757 software) will have diverse abilities in understanding elements and 1758 qualifiers. 1760 Any number of other elements and qualifiers may be used in 1761 conjunction with the quartet of basic segment questions. The only 1762 semantic requirement is that they pertain to the segment's story. 1763 Also, it is only the four basic elements that change meaning 1764 depending on their segment context. All other elements have meaning 1765 independent of the segment in which they appear. If an element label 1766 stripped of its qualifier is still not recognized by the recipient, a 1767 second fall back position is to ignore it and rely on the four basic 1768 elements. 1770 Elements may be either Canonical, Provisional, or Local. Canonical 1771 elements are officially recognized via a registry as part of the 1772 metadata vernacular. All elements, qualifiers, and segment labels 1773 used in this document up until now belong to that vernacular. 1774 Provisional elements are also officially recognized via the registry, 1775 but have only been proposed for inclusion in the vernacular. To be 1776 promoted to the vernacular, a provisional element passes through a 1777 vetting process during which its documentation must be in order and 1778 its community acceptance demonstrated. Local elements are any 1779 elements not officially recognized in the registry. The registry 1780 [DERC] is a work in progress. 1782 Local elements can be immediately distinguishable from Canonical or 1783 Provisional elements because all terms that begin with an upper case 1784 letter are reserved for spontaneous local use. No term beginning 1785 with an upper case letter will ever be assigned Canonical or 1786 Provisional status, so it should be safe to use such terms for local 1787 purposes. Any recipient of external ERCs containing such terms will 1788 understand them to be part of the originating provider's local 1789 metadata dialect. Here's an example ERC with three segments, one 1790 local element, and two local qualifiers. The segment boundaries have 1791 been emphasized by comment lines (which, as before, are ignored by 1792 processors). 1794 erc: 1795 who: Bullock, TH | Achimowicz, JZ | Duckrow, RB 1796 | Spencer, SS | Iragui-Madoz, VJ 1797 what: Bicoherence of intracranial EEG in sleep, 1798 wakefulness and seizures 1799 when: 1997 12 00 1800 where: http://cogprints.soton.ac.uk/%{ 1801 documents/disk0/00/00/01/22/index.html %} 1802 in: EEG Clin Neurophysiol | 1997 12 00 | v103, i6, p661-678 1803 IDcode: cog00000122 1804 # ---- new segment ---- 1805 erc-about: 1806 what/Subcategory: Bispectrum | Nonlinearity | Epilepsy 1807 | Cooperativity | Subdural | Hippocampus | Higher moment 1808 # ---- new segment ---- 1809 erc-from: 1810 who: NIH/NLM/NCBI 1811 what: pm9546494 1812 when/Reviewed: 1998 04 18 021600 1813 where: http://ark.nlm.nih.gov/12025/pm9546494? 1815 The local element "IDcode" immediately precedes the "erc-about" 1816 segment, which itself contains an element with the local qualifier 1817 "Subcategory". The second to last element also carries the local 1818 qualifier "Reviewed". Finally, what might be a provisional element 1819 "in" appears near the end of the first segment. It might have been 1820 proposed as a way to complete a citation for an object originally 1821 appearing inside another object (such as an article appearing in a 1822 journal or an encyclopedia). 1824 7.5. ERC Element Values 1826 ERC element values tend to be straightforward strings. If the 1827 provider intends something special for an element, it will so 1828 indicate with markers at the beginning of its value string. The 1829 markers are designed to be uncommon enough that they would not likely 1830 occur in normal data except by deliberate intent. Markers can only 1831 occur near the beginning of a string, and once any octet of non- 1832 marker data has been encountered, no further marker processing is 1833 done for the element value. In the absence of markers the string is 1834 considered pure data; this has been the case with all the examples 1835 seen thus far. The fullest form of an element value with all three 1836 optional markers in place looks like this. 1838 VALUE = [markup_flags] (:ccode) , DATA 1840 In processing, the first non-whitespace character of an ERC element 1841 value is examined. An initial `[' is reserved to introduce a 1842 bracketed set of markup flags (not described in this document) that 1843 ends with `]'. If ERC data is machine-generated, each value string 1844 may be preceded by "[]" to prevent any of its data from being 1845 mistaken for markup flags. Once past the optional markup, the 1846 remaining value may optionally begin with a controlled code. A 1847 controlled code always has the form "(:ccode)", for example, 1849 who: (:unkn) Anonymous 1850 what: (:791) Bee Stings 1852 Any string after such a code is taken to be an uncontrolled (e.g., 1853 natural language) equivalent. The code "unkn" indicates a 1854 conventional explanation for a missing value (stating that the value 1855 is unknown). The remainder of the string makes an equivalent 1856 statement in a form that the provider deemed most suitable to its 1857 (probably human) audience. The code "791" could be a fixed numeric 1858 topic identifier within an unspecified topic vocabulary. Any code 1859 may be ignored by those that do not understand it. 1861 There are several codes to explain different ways in which a required 1862 element's value may go missing. 1864 (:unac) temporarily inaccessible 1865 (:unal) unallowed, suppressed intentionally 1866 (:unap) not applicable, makes no sense 1867 (:unas) value unassigned (e.g., Untitled) 1868 (:unav) value unavailable indefinitely 1869 (:unkn) unknown (e.g., Anonymous, Inconnue) 1870 (:etal) too numerous to list (I). 1871 (:none) never had a value, never will 1872 (:null) explicitly empty 1873 (:tba) to be assigned or announced later 1875 Once past an optional controlled code, the remaining string value is 1876 subjected to one final test. If the first next non-whitespace 1877 character is a `,' (comma), it indicates that the string value is 1878 "sort-friendly". This means that the value is (a) laid out with an 1879 inverted word order useful for sorting items having comparably laid 1880 out element values (items might be the containing ERC records) and 1881 (b) that the value may contain other commas that indicate inversion 1882 points should it become necessary to recover the value in natural 1883 word order. Typically, this feature is used to express Western-style 1884 personal names in family-name-given-name order. It can also be used 1885 wherever natural word order might make sorting tricky, such as when 1886 data contains titles or corporate names. Here are some example 1887 elements. 1889 who: , van Gogh, Vincent 1890 who:,Howell, III, PhD, 1922-1987, Thurston 1891 who:, Acme Rocket Factory, Inc., The 1892 who:, Mao Tse Tung 1893 who:, McCartney, Paul, Sir, 1894 what:, Health and Human Services, United States Government 1895 Department of, The, 1897 There are rules to use in recovering a copy of the value in natural 1898 word order, if desired. The above example strings have the following 1899 natural word order values, respectively. 1901 Vincent van Gogh 1902 Thurston Howell, III, PhD, 1922-1987 1903 The Acme Rocket Factory, Inc. 1904 Mao Tse Tung 1905 Sir Paul McCartney 1906 The United States Government Department of Health and Human Services 1908 7.6. ERC Element Encoding and Dates 1910 Some characters that need to appear in ERC element values might 1911 conflict with special characters used for structuring ERCs, so there 1912 needs to be a way to include them as literal characters that are 1913 protected from special interpretation. This is accomplished through 1914 an encoding mechanism that resembles the %-encoding familiar to [URI] 1915 handlers. 1917 The ERC encoding mechanism also uses `%', but instead of taking two 1918 following hexadecimal digits, it takes one non-alphanumeric character 1919 or two alphabetic characters that cannot be mistaken for hex digits. 1920 It is designed not to be confused with normal web-style %-encoding. 1921 In particular it can be decoded without risking unintended decoding 1922 of normal %-encoded data (which would introduce errors). Here are 1923 the one-character (non-alphanumeric) ERC encoding extensions. 1925 ERC Purpose 1926 --- ------------------------------------------------ 1927 %! decodes to the element separator `|' 1928 %% decodes to a percent sign `%' 1929 %. decodes to a comma `,' 1930 %_ a non-character used as syntax shim 1931 %{ a non-character that begins an expansion block 1932 %} a non-character that ends an expansion block 1934 One particularly useful construct in ERC element values is the pair 1935 of special encoding markers ("%{" and "%}") that indicates a 1936 "expansion" block. Whatever string of characters they enclose will 1937 be treated as if none of the contained whitespace (SPACEs, TABs, 1938 Newlines) were present. This comes in handy for writing long, multi- 1939 part URLs in a readable way. For example, the value in 1941 where: http://foo.bar.org/node%{ 1942 ? db = foo 1943 & start = 1 1944 & end = 5 1945 & buf = 2 1946 & query = foo + bar + zaf 1947 %} 1949 is decoded into an equivalent element, but with a correct and intact 1950 URL: 1952 where: 1953 http://foo.bar.org/node?db=foo&start=1&end=5&buf=2&query=foo+bar+zaf 1955 In a parting word about ERC element values, a commonly recurring 1956 value type is a date, possibly followed by a time. ERC dates use the 1957 [TEMPER] format, taking on one of the following forms: 1959 1999 (four digit year) 1960 2000 12 29 (year, month, day) 1961 2000 12 29 235955 (year, month, day, hour, minute, second) 1963 In dates, all internal whitespace is squeezed out to achieve a 1964 normalized form suitable for lexical comparison and sorting. This 1965 means that the following dates 1967 2000 12 29 235955 (recommended for readability) 1968 2000 12 29 23 59 55 1969 20001229 23 59 55 1970 20001229235955 (normalized date and time) 1972 are all equivalent. The first form is recommended for readability. 1973 The last form (shortest and easiest to compute with) is the 1974 normalized form. Hyphens and commas are reserved to create date 1975 ranges and lists, for example, 1977 1996-2000 (a range of four years) 1978 1952, 1957, 1969 (a list of three years) 1979 1952, 1958-1967, 1985 (a mixed list of dates and ranges) 1980 20001229-20001231 (a range of three days) 1982 7.7. ERC Stub Records and Internal Support 1984 The ERC design introduces the concept of a "stub" record, which is an 1985 incomplete ERC record intended to be supplemented with additional 1986 elements before being released as a standalone ERC record. A stub 1987 ERC record has no minimum required elements. It is just a group of 1988 elements that does not begin with "erc:" but otherwise conforms to 1989 the ERC record syntax. 1991 ERC stubs may be useful in supporting internal procedures using the 1992 ERC syntax. Often they rely on the convenience and accuracy of 1993 automatically supplied elements, even the basic ones. To be ready 1994 for external use, however, an ERC stub must be transformed into a 1995 complete ERC record having the usual required elements. An ERC stub 1996 record can be convenient for metadata embedded in a document, where 1997 elements such as location, modification date, and size - which one 1998 would not omit from an externalized record - are omitted simply 1999 because they are much better supplied by a computation. A separate 2000 local administrative procedure, not defined for ERC's in general, 2001 would effect the promotion of stubs into complete records. 2003 While the ERC is a general-purpose container for exchange of resource 2004 descriptions, it does not dictate how records must be internally 2005 stored, laid out, or assembled by data providers or recipients. 2006 Arbitrary internal descriptive frameworks can support ERCs simply by 2007 mapping (e.g., on demand) local records to the ERC container format 2008 and making them available for export. Therefore, to support ERCs 2009 there is no need for a data provider to convert internal data to be 2010 stored in an ERC format. On the other hand, any provider (such as 2011 one just getting started in the business of resource description) may 2012 choose to store and manipulate local data natively in the ERC format. 2014 8. Advice to Web Clients 2016 This section offers some advice to web client software developers. 2017 It is hard to write about because it tries to anticipate a series of 2018 events that might lead to native web browser support for ARKs. 2020 ARKs are envisaged to appear wherever durable object references are 2021 planned. Library cataloging records, literature citations, and 2022 bibliographies are important examples. In many of these places URLs 2023 (Uniform Resource Locators) currently stand in, and URNs, DOIs, and 2024 PURLs have been proposed as alternatives. 2026 The strings representing ARKs are also envisaged to appear in some of 2027 the places where URLs currently appear: in hypertext links (where 2028 they are not normally shown to users) and in rendered text (displayed 2029 or printed). Internet search engines, for example, tend to include 2030 both actionable and manifest links when listing each item found. A 2031 normal HTML link for which the URL is not displayed looks like this. 2033 Click Here 2035 The same link with an ARK instead of a URL: 2037 Click Here 2039 Web browsers would in general require a small modification to 2040 recognize and convert this ARK, via mapping authority discovery, to 2041 the URL form. 2043 Click Here 2045 A browser that knows how to make that conversion could also 2046 automatically detect and replace a non-working NMAH. 2048 An NAA will typically make known the associations it creates by 2049 publishing them in catalogs, actively advertizing them, or simply 2050 leaving them on web sites for visitors (e.g., users, indexing 2051 spiders) to stumble across in browsing. 2053 9. Security Considerations 2055 The ARK naming scheme poses no direct risk to computers and networks. 2056 Implementors of ARK services need to be aware of security issues when 2057 querying networks and filesystems for Name Mapping Authority 2058 services, and the concomitant risks from spoofing and obtaining 2059 incorrect information. These risks are no greater for ARK mapping 2060 authority discovery than for other kinds of service discovery. For 2061 example, recipients of ARKs with a specified hostport (NMAH) should 2062 treat it like a URL and be aware that the identified ARK service may 2063 no longer be operational. 2065 Apart from mapping authority discovery, ARK clients and servers 2066 subject themselves to all the risks that accompany normal operation 2067 of the protocols underlying mapping services (e.g., HTTP, Z39.50). 2068 As specializations of such protocols, an ARK service may limit 2069 exposure to the usual risks. Indeed, ARK services may enhance a kind 2070 of security by helping users identify long-term reliable references 2071 to information objects. 2073 10. Authors' Addresses 2075 John A. Kunze 2076 California Digital Library 2077 University of California, Office of the President 2078 415 20th St, 4th Floor 2079 Oakland, CA 94612-3550, USA 2081 Fax: +1 510-893-5212 2082 EMail: jak@ucop.edu 2084 R. P. C. Rodgers 2085 US National Library of Medicine 2086 8600 Rockville Pike, Bldg. 38A 2087 Bethesda, MD 20894, USA 2089 Fax: +1 301-496-0673 2090 EMail: rodgers@nlm.nih.gov 2092 11. References 2094 [ANVL] J. Kunze, B. Kahle, et al, "A Name-Value Language", work 2095 in progress, 2096 http://www.cdlib.org/inside/diglib/ark/anvlspec.pdf 2098 [ARK] J. Kunze, "Towards Electronic Persistence Using ARK 2099 Identifiers", Proceedings of the 3rd ECDL Workshop on Web 2100 Archives, August 2003, (PDF) 2101 http://bibnum.bnf.fr/ecdl/2003/proceedings.php?f=kunze 2103 [DCORE] Dublin Core Metadata Initiative, "Dublin Core Metadata 2104 Element Set, Version 1.1: Reference Description", July 2105 1999, http://dublincore.org/documents/dces/. 2107 [DERC] J. Kunze, "Dictionary of the ERC", work in progress within 2108 the Dublin Core Metadata Initiative's Kernel Working 2109 Group, http://dublincore.org/groups/kernel/ 2111 [DNS] P.V. Mockapetris, "Domain Names - Concepts and 2112 Facilities", RFC 1034, November 1987. 2114 [DOI] International DOI Foundation, "The Digital Object 2115 Identifier (DOI) System", February 2001, 2116 http://dx.doi.org/10.1000/203. 2118 [ERC] J. Kunze, "A Metadata Kernel for Electronic Permanence", 2119 Journal of Digital Information, Vol 2, Issue 2, January 2120 2002, ISSN 1368-7506, (PDF) 2121 http://jodi.ecs.soton.ac.uk/Articles/v02/i02/Kunze/ 2123 [Handle] L. Lannom, "Handle System Overview", ICSTI Forum, No. 30, 2124 April 1999, http://www.icsti.org/forum/30/#lannom 2126 [HTTP] R. Fielding, et al, "Hypertext Transfer Protocol -- 2127 HTTP/1.1", RFC 2616, June 1999. 2129 [MD5] R. Rivest, "The MD5 Message-Digest Algorithm", RFC 1321, 2130 April 1992. 2132 [NAPTR] M. Mealling, Daniel, R., "The Naming Authority Pointer 2133 (NAPTR) DNS Resource Record", RFC 2915, September 2000. 2135 [NLMPerm] M. Byrnes, "Defining NLM's Commitment to the Permanence of 2136 Electronic Information", ARL 212:8-9, October 2000, 2137 http://www.arl.org/newsltr/212/nlm.html 2139 [NOID] J. Kunze, "Nice Opaque Identifiers", February 2005, 2140 http://www.cdlib.org/inside/diglib/ark/noid.pdf 2142 [PURL] K. Shafer, et al, "Introduction to Persistent Uniform 2143 Resource Locators", 1996, 2144 http://purl.oclc.org/OCLC/PURL/INET96 2146 [RFC822] D. Crocker, "Standard for the format of ARPA Internet text 2147 messages", RFC 822, August 1982. 2149 [TELNET] J. Postel, J.K. Reynolds, "Telnet Protocol Specification", 2150 RFC 854, May 1983. 2152 [TEMPER] J. Kunze, "Temporal Enumerated Ranges", work in progress, 2153 http://www.cdlib.org/inside/diglib/ark/temperspec.pdf 2155 [THUMP] K. Gamiel, J. Kunze, N. Nassar, "The HTTP URL Mapping 2156 Protocol", work in progress. 2158 [URI] T. Berners-Lee, et al, "Uniform Resource Identifiers 2159 (URI): Generic Syntax", RFC 2396, August 1998. 2161 [URNBIB] C. Lynch, et al, "Using Existing Bibliographic Identifiers 2162 as Uniform Resource Names", RFC 2288, February 1998. 2164 [URNSYN] R. Moats, "URN Syntax", RFC 2141, May 1997. 2166 [URNNID] L. Daigle, et al, "URN Namespace Definition Mechanisms", 2167 RFC 2611, June 1999. 2169 12. Appendix: ARK Implementations 2171 Currently, the primary implementation activity is at the California 2172 Digital Library (CDL), 2174 http://ark.cdlib.org/ 2176 housed at the University of California Office of the President, where 2177 over 200,000 ARKs have been assigned to objects that the CDL owns or 2178 controls. Some experimentation in ARKs is taking place at JSTOR, the 2179 Digital Curation Centre, WIPO and at the University of California's 2180 San Diego, San Francisco, and Berkeley campuses. 2182 The US National Library of Medicine (NLM) also has an experimental, 2183 prototype ARK service under development. It is being made available 2184 for purposes of demonstrating various aspects of the ARK system, but 2185 is subject to temporary or permanent withdrawal (without notice) 2186 depending upon the circumstances of the small research group 2187 responsible for making it available. It is described at: 2189 http://ark.nlm.nih.gov/ 2191 Comments and feedback may be addressed to rodgers@nlm.nih.gov. 2193 13. Appendix: Current ARK Name Authority Table 2195 This appendix contains a copy of the Name Authority Table (a file) at 2196 the time of writing. It may be loaded into a local filesystem (e.g., 2197 /etc/natab) for use in mapping NAAs (Name Assigning Authorities) to 2198 NMAHs (Name Mapping Authority Hostports). It contains Perl code that 2199 can be copied into a standalone script that processes the table (as a 2200 file). Because this is still a proposed file, none of the values in 2201 it are real. 2203 # 2204 # Name Assigning Authority / Name Mapping Authority Lookup Table 2205 # Last change: 2006.01.12 2206 # Reload from: http://ark.nlm.nih.gov/etc/natab 2207 # Mirrored at: http://www.cdlib.org/inside/diglib/ark/natab 2208 # To register: mailto:ark@cdlib.org?Subject=naareg 2209 # Process with: Perl script at end of this file (optional) 2210 # 2211 # Each NAA appears at the beginning of a line with the NAA Number 2212 # first, a colon, and an ARK or URL to a statement of naming policy 2213 # (see http://ark.cdlib.org for an example). 2214 # All the NMA hostports that service an NAA are listed, one per 2215 # line, indented, after the corresponding NAA line. 2216 # 2217 # National Library of Medicine 2218 12025: http://www.nlm.nih.gov/xxx/naapolicy.html 2219 ark.nlm.nih.gov USNLM 2220 foobar.zaf.org UCSF 2221 # 2222 # Library of Congress 2223 12026: http://www.loc.gov/xxx/naapolicy.html 2224 foobar.zaf.org USLC 2225 # 2226 # National Agriculture Library 2227 12027: http://www.nal.gov/xxx/naapolicy.html 2228 foobar.zaf.gov:80 USNAL 2229 # 2230 # California Digital Library 2231 13030: http://www.cdlib.org/inside/diglib/ark/ 2232 ark.cdlib.org CDL 2233 # 2234 # World Intellectual Property Organization 2235 13038: http://www.wipo.int/xxx/naapolicy.html 2236 www.wipo.int WIPO 2237 # 2238 # University of California San Diego 2239 20775: http://library.ucsd.edu/xxx/naapolicy.html 2240 ucsd.edu UCSD 2241 # 2242 # University of California San Francisco 2243 29114: http://library.ucsf.edu/xxx/naapolicy.html 2244 ucsf.edu UCSF 2245 # 2246 # University of California Berkeley 2247 28722: http://library.berkeley.edu/xxx/naapolicy.html 2248 berkeley.edu UCB 2249 # 2250 # University of California Los Angeles 2251 21198: http://library.ucla.edu/xxx/naapolicy.html 2252 ucla.edu UCLA 2253 # 2254 # Rutgers University 2255 15230: http://rci.rutgers.edu/xxx/naapolicy.html 2256 rutgers.edu RU 2257 # 2258 # Internet Archive 2259 13960: http://www.archive.org/xxx/naapolicy.html 2260 archive.org IA 2261 # 2262 # Digital Curation Centre 2263 64269: http://www.dcc.ac.uk/xxx/naapolicy.html 2264 dcc.ac.uk DCC 2265 # 2266 # New York University 2267 62624: http://library.nyu.edu/xxx/naapolicy.html 2268 nyu.edu NYU 2269 # 2270 # University of North Texas 2271 67531: http://www.library.unt.edu/xxx/naapolicy.html 2272 unt.edu UNT 2273 # 2274 # Ithaka Electronic-Archiving Initiative 2275 27927: http://www.ithaka.org/xxx/naapolicy.html 2276 ithaka.org ITHAKA 2277 # 2278 # Bibliothque nationale de France / National Library of France 2279 12148: http://www.bnf.fr/xxx/naapolicy.html 2280 bnf.fr BNF 2281 # 2282 # Princeton University 2283 88435: http://diglib.princeton.edu/xxx/naapolicy.html 2284 princeton.edu PU 2285 # 2286 # University of Washington 2287 78428: http://u.washington.edu/xxx/naapolicy.html 2288 u.washington.edu UW 2289 # 2290 # Archives of Region of Vstra Gtaland and City of Gothenburg, Sweden 2291 89901: http://www.arkivnamnden.org/xxx/naapolicy.html 2292 arkivnamnden.org AVGG 2293 # 2294 # Northwest Digital Archives 2295 80444: http://nwda.wsulibs.wsu.edu/xxx/naapolicy.html 2296 nwda.wsulibs.wsu.edu NWDA 2297 # 2298 # Emory University 2299 25593: http://id.library.emory.edu/xxx/naapolicy.html 2300 id.library.emory.edu EMORY 2301 # 2302 #--- end of data --- 2303 # The following Perl script takes an NAA as argument and outputs 2304 # the NMAs in this file listed under any matching NAA. 2306 # 2307 # my $naa = shift; 2308 # while (<>) { 2309 # next if (! /^$naa:/); 2310 # while (<>) { 2311 # last if (! /^[#\s]./); 2312 # print "$1\n" if (/^\s+(\S+)/); 2313 # } 2314 # } 2315 # 2316 # Create a g/t/nroff-safe version of this table with the UNIX command, 2317 # 2318 # expand natab | sed 's/\\/\\\e/g' > natab.roff 2319 # 2320 # end of file 2322 14. Copyright Notice 2324 Copyright (C) The Internet Society (2006). This document is subject 2325 to the rights, licenses and restrictions contained in BCP 78, and 2326 except as set forth therein, the authors retain all their rights. 2328 This document and the information contained herein are provided on an 2329 "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS 2330 OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET 2331 ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, 2332 INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE 2333 INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED 2334 WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. 2336 Expires 23 August 2006 2337 Table of Contents 2339 Status of this Document . . . . . . . . . . . . . . . . . . . . . . 1 2340 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2341 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2342 1.1. Reasons to Use ARKs . . . . . . . . . . . . . . . . . . . . . 4 2343 1.2. Three Requirements of ARKs . . . . . . . . . . . . . . . . . . 4 2344 1.3. Organizing Support for ARKs: Our Stuff vs. Their Stuff . . . 5 2345 1.4. Definition of Identifier . . . . . . . . . . . . . . . . . . . 7 2346 2. ARK Anatomy . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2347 2.1. The Name Mapping Authority Hostport (NMAH) . . . . . . . . . . 8 2348 2.2. The ARK Label Part - ark: . . . . . . . . . . . . . . . . . . 9 2349 2.3. The Name Assigning Authority Number (NAAN) . . . . . . . . . . 10 2350 2.4. The Name Part . . . . . . . . . . . . . . . . . . . . . . . . 10 2351 2.5. The Qualifier Part . . . . . . . . . . . . . . . . . . . . . . 11 2352 2.5.1. ARKs that Reveal Object Hierarchy . . . . . . . . . . . . . 12 2353 2.5.2. ARKs that Reveal Object Variants . . . . . . . . . . . . . . 13 2354 2.6. Character Repertoires . . . . . . . . . . . . . . . . . . . . 14 2355 2.7. Normalization and Lexical Equivalence . . . . . . . . . . . . 15 2356 3. Naming Considerations . . . . . . . . . . . . . . . . . . . . . 16 2357 3.1. ARKS Embedded in Language . . . . . . . . . . . . . . . . . . 16 2358 3.2. Objects Should Wear Their Identifiers . . . . . . . . . . . . 17 2359 3.3. Names are Political, not Technological . . . . . . . . . . . . 17 2360 3.4. Choosing a Hostname or NMA . . . . . . . . . . . . . . . . . . 17 2361 3.5. Assigners of ARKs . . . . . . . . . . . . . . . . . . . . . . 19 2362 3.6. NAAN Namespace Management . . . . . . . . . . . . . . . . . . 20 2363 3.7. Sub-Object Naming . . . . . . . . . . . . . . . . . . . . . . 21 2364 4. Finding a Name Mapping Authority . . . . . . . . . . . . . . . . 21 2365 4.1. Looking Up NMAHs in a Globally Accessible File . . . . . . . . 22 2366 4.2. Looking up NMAHs Distributed via DNS . . . . . . . . . . . . . 23 2367 5. Generic ARK Service Definition . . . . . . . . . . . . . . . . . 25 2368 5.1. Generic ARK Access Service (access, location) . . . . . . . . 26 2369 5.2. Generic Policy Service (permanence, naming, etc.) . . . . . . 26 2370 5.3. Generic Description Service . . . . . . . . . . . . . . . . . 28 2371 6. Overview of The HTTP URL Mapping Protocol (THUMP) . . . . . . . 28 2372 7. Overview of Electronic Resource Citations (ERCs) . . . . . . . . 31 2373 7.1. ERC Syntax . . . . . . . . . . . . . . . . . . . . . . . . . . 32 2374 7.2. ERC Stories . . . . . . . . . . . . . . . . . . . . . . . . . 33 2375 7.3. The ERC Anchoring Story . . . . . . . . . . . . . . . . . . . 34 2376 7.4. ERC Elements . . . . . . . . . . . . . . . . . . . . . . . . . 35 2377 7.5. ERC Element Values . . . . . . . . . . . . . . . . . . . . . . 37 2378 7.6. ERC Element Encoding and Dates . . . . . . . . . . . . . . . . 39 2379 7.7. ERC Stub Records and Internal Support . . . . . . . . . . . . 41 2380 8. Advice to Web Clients . . . . . . . . . . . . . . . . . . . . . 41 2381 9. Security Considerations . . . . . . . . . . . . . . . . . . . . 42 2382 10. Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . 42 2383 11. References . . . . . . . . . . . . . . . . . . . . . . . . . . 43 2384 12. Appendix: ARK Implementations . . . . . . . . . . . . . . . . 44 2385 13. Appendix: Current ARK Name Authority Table . . . . . . . . . . 45 2386 14. Copyright Notice . . . . . . . . . . . . . . . . . . . . . . . 48