idnits 2.17.1 draft-vandesompel-citeas-03.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- ** There are 4 instances of too long lines in the document, the longest one being 13 characters in excess of 72. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (June 25, 2018) is 2132 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 5988 (Obsoleted by RFC 8288) Summary: 2 errors (**), 0 flaws (~~), 1 warning (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group H. Van de Sompel 3 Internet-Draft Los Alamos National Laboratory 4 Intended status: Informational M. Nelson 5 Expires: December 27, 2018 Old Dominion University 6 G. Bilder 7 Crossref 8 J. Kunze 9 California Digital Library 10 S. Warner 11 Cornell University 12 June 25, 2018 14 cite-as: A Link Relation to Convey a Preferred URI for Referencing 15 draft-vandesompel-citeas-03 17 Abstract 19 A web resource is routinely referenced by means of the URI with which 20 it is directly accessed. But cases exist where referencing a 21 resource by means of a URI, different than that access URI, is 22 preferred. This specification defines a link relation type that can 23 be used to convey such a preference. 25 Note to Readers 27 Please discuss this draft on the ART mailing list 28 (). 30 Status of This Memo 32 This Internet-Draft is submitted in full conformance with the 33 provisions of BCP 78 and BCP 79. 35 Internet-Drafts are working documents of the Internet Engineering 36 Task Force (IETF). Note that other groups may also distribute 37 working documents as Internet-Drafts. The list of current Internet- 38 Drafts is at https://datatracker.ietf.org/drafts/current/. 40 Internet-Drafts are draft documents valid for a maximum of six months 41 and may be updated, replaced, or obsoleted by other documents at any 42 time. It is inappropriate to use Internet-Drafts as reference 43 material or to cite them other than as "work in progress." 45 This Internet-Draft will expire on December 27, 2018. 47 Copyright Notice 49 Copyright (c) 2018 IETF Trust and the persons identified as the 50 document authors. All rights reserved. 52 This document is subject to BCP 78 and the IETF Trust's Legal 53 Provisions Relating to IETF Documents 54 (https://trustee.ietf.org/license-info) in effect on the date of 55 publication of this document. Please review these documents 56 carefully, as they describe your rights and restrictions with respect 57 to this document. Code Components extracted from this document must 58 include Simplified BSD License text as described in Section 4.e of 59 the Trust Legal Provisions and are provided without warranty as 60 described in the Simplified BSD License. 62 Table of Contents 64 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 65 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 3 66 3. Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . 3 67 3.1. Persistent Identifiers . . . . . . . . . . . . . . . . . 3 68 3.2. Version Identifiers . . . . . . . . . . . . . . . . . . . 4 69 3.3. Preferred Social Identifier . . . . . . . . . . . . . . . 5 70 3.4. Multi-Resource Publications . . . . . . . . . . . . . . . 5 71 4. The "cite-as" Relation Type for Expressing a Preferred URI 72 for the Purpose of Referencing . . . . . . . . . . . . . . . 6 73 5. Distinction with Other Relation Types . . . . . . . . . . . . 7 74 5.1. bookmark . . . . . . . . . . . . . . . . . . . . . . . . 8 75 5.2. canonical . . . . . . . . . . . . . . . . . . . . . . . . 8 76 6. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . 10 77 6.1. Persistent HTTP URI . . . . . . . . . . . . . . . . . . . 10 78 6.2. Version URIs . . . . . . . . . . . . . . . . . . . . . . 11 79 6.3. Preferred Profile URI . . . . . . . . . . . . . . . . . . 12 80 6.4. Multi-Resource Publication . . . . . . . . . . . . . . . 13 81 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 13 82 7.1. Link Relation Type: cite-as . . . . . . . . . . . . . . . 13 83 8. Security Considerations . . . . . . . . . . . . . . . . . . . 14 84 9. References . . . . . . . . . . . . . . . . . . . . . . . . . 14 85 9.1. Normative References . . . . . . . . . . . . . . . . . . 14 86 9.2. Informative References . . . . . . . . . . . . . . . . . 15 87 Appendix A. Acknowledgements . . . . . . . . . . . . . . . . . . 16 88 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 16 90 1. Introduction 92 A web resource is routinely referenced (e.g. linked, bookmarked) by 93 means of the URI with which it is directly accessed. But cases exist 94 where referencing a resource by means of a different URI is 95 preferred, for example because the latter URI is intended to be more 96 persistent over time. Currently, there is no link relation type to 97 convey such alternative referencing preference; this specification 98 addresses this deficit by introducing a link relation type intended 99 for that purpose. 101 2. Terminology 103 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 104 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 105 document are to be interpreted as described in RFC 2119 [RFC2119]. 107 This specification uses the terms "link context" and "link target" as 108 defined in [RFC8288]. These terms respectively correspond with 109 "Context IRI" and "Target IRI" as used in [RFC5988]. Although 110 defined as IRIs, in common scenarios they are also URIs. 112 Additionally, this specification uses the following terms: 114 o "access URI": A URI at which a user agent accesses a web resource. 116 o "reference URI": A URI, other than the access URI, that should 117 preferentially be used for referencing. 119 By interacting with the access URI, the user agent may discover typed 120 links. For such links, the access URI is the link context. 122 3. Scenarios 124 3.1. Persistent Identifiers 126 Despite sound advice regarding the design of Cool URIs [CoolURIs], 127 link rot ("HTTP 404 Not Found") is a common phenomena when following 128 links on the web. Certain communities of practice have introduced 129 solutions to combat this problem that typically consist of: 131 o Accepting the reality that the web location of a resource - the 132 access URI - may change over time. 134 o Minting an additional URI for the resource - the reference URI - 135 that is specifically intended to remain persistent over time. 137 o Redirecting (typically "HTTP 301 Moved Permanently", "HTTP 302 138 Found", or "HTTP 303 See Other") from the reference URI to the 139 access URI. 141 o As a community, committing to adjust that redirection whenever the 142 access URI changes over time. 144 This approach is, for example, used by: 146 o Scholarly publishers that use DOIs [DOIs] to identify articles and 147 DOI URLs [DOI-URLs] as a means to keep cross-publisher article-to- 148 article links operational, even when the journals in which the 149 articles are published change hands from one publisher to another, 150 for example, as a result of an acquisition. 152 o Authors of controlled vocabularies that use PURLs [PURLs] for 153 vocabulary terms to ensure that the term URIs remain stable even 154 if management of the vocabulary is transfered to a new custodian. 156 o A variety of organizations, including libraries, archives, and 157 museums that assign ARK URLs [draft-kunze-ark-18] to information 158 objects in order to support long-term access. 160 In order for the investments in infrastructure involved in these 161 approaches to pay off, and hence for links to effectively remain 162 operational as intended, it is crucial that a resource be referenced 163 by means of its reference URI. However, the access URI is where a 164 user agent actually accesses the resource (e.g., it is the URI in the 165 browser's address bar). As such, there is a considerable risk that 166 the access URI instead of the reference URI is used for referencing 167 [PIDs-must-be-used]. 169 The link relation type defined in this specification allows to convey 170 to user agents that the reference URI is the preferred URI for 171 referencing. 173 3.2. Version Identifiers 175 Resource versioning systems often use a naming approach whereby: 177 o The most recent version of a resource is at any time available at 178 the same, generic URI. 180 o Each version of the resource - including the most recent one - has 181 a distinct version URI. 183 For example, Wikipedia uses generic URIs of the form 184 http://en.wikipedia.org/wiki/John_Doe and version URIs of the form 185 https://en.wikipedia.org/w/index.php?title=John_Doe&oldid=776253882. 187 While the current version of a resource is accessed at the generic 188 URI, some versioning systems adhere to a policy that favors linking 189 and referencing a specific version URI. To express this using the 190 terminology of Section 2, these policies intend that the generic URI 191 is the access URI, and that the version URI is the reference URI. 193 These policies are informed by the understanding that the content at 194 the generic URI is likely to evolve over time, and that accurate 195 links or references should lead to the content as it was at the time 196 of referencing. To that end, Wikipedia's "Permanent link" and "Cite 197 this page" functionalities promote the version URI, not the generic 198 URI. 200 The link relation type defined in this specification allows to convey 201 to user agents that the version URI is preferred over the generic URI 202 for referencing. 204 3.3. Preferred Social Identifier 206 A web user commonly has multiple profiles on the web, for example, 207 one per social network, a personal homepage, a professional homepage, 208 a FOAF profile [FOAF], etc. Each of these profiles is accessible at 209 a distinct URI. But the user may have a preference for one of those 210 profiles, for example, because it is most complete, kept up-to-date, 211 or expected to be long-lived. As an example, the first author of 212 this document has, among others, the following profile URIs: 214 o https://hvdsomp.info 216 o http://public.lanl.gov/herbertv/ 218 o https://www.linkedin.com/in/herbertvandesompel/ 220 o https://orcid.org/0000-0002-0715-6126 222 Of these, from the perspective of the person described by these 223 profiles, the first URI may be the preferred profile URI for the 224 purpose of referencing because the domain is not under the 225 custodianship of a third party. When an agent accesses another 226 profile URI, such as http://public.lanl.gov/herbertv/, this 227 preference for referencing by means of the first URI could be 228 expressed. 230 The link relation type defined in this specification allows to convey 231 to user agents that a profile URI - the reference URI - other than 232 the one the agent is accessing - the access URI - is preferred for 233 referencing. 235 3.4. Multi-Resource Publications 237 When publishing on the web, it is not uncommon to make distinct 238 components of a publication available as different web resources, 239 each with their own URI. For example: 241 o Contemporary scholarly publications routinely consists of a 242 traditional article as well as additional materials that are 243 considered an integral part of the publication such as 244 supplementary information, high-resolution images, a video 245 recording of an experiment. 247 o Scientific or governmental open data sets frequently consist of 248 multiple files. 250 o Online books typically consist of multiple chapters. 252 While each of these components are accessible at their distinct URI - 253 the access URI - they often also share a URI assigned to the 254 intellectual publication of which they are components - the reference 255 URI. 257 The link relation type defined in this specification allows to convey 258 to user agents that, for the purpose of referencing, the reference 259 URI of the intellectual publication is preferred over an access URI 260 of a component of the publication. 262 4. The "cite-as" Relation Type for Expressing a Preferred URI for the 263 Purpose of Referencing 265 A link with the "cite-as" relation type indicates that, for 266 referencing the link context, use of the URI of the link target is 267 preferred over use of the URI of the link context. It allows the 268 resource identified by the access URI (link context) to unambiguously 269 link to its corresponding reference URI (link target), thereby 270 expressing that the link target is preferred over the link context 271 for the purpose of permanent citation. 273 The link target of a "cite-as" link SHOULD support protocol-based 274 access as a means to ensure that applications that store them can 275 effectively re-use them for access. 277 The link target of a "cite-as" link SHOULD provide the ability for a 278 user agent to follow its nose back to the context of the link, e.g. 279 by following redirects and/or links. This helps a user agent to 280 establish trust in the target URI. 282 Because a link with the "cite-as" relation type expresses a preferred 283 URI for the purpose of referencing, the access URI SHOULD only 284 provide one link with that relation type. If more than one "cite-as" 285 link is provided, the user agent may decide to select one (e.g. an 286 HTTP URI over a mailto URI), for example, based on the purpose that 287 the reference URI will serve. 289 Providing a link with the "cite-as" relation type does not prevent 290 using the access URI for the purpose of referencing if such 291 specificity is needed for the application at hand. For example, in 292 the case of scenario Section 3.4 the access URI is likely required 293 for the purpose of annotating a specific component of an intellectual 294 publication. Yet, the annotation application may also want to 295 appropriately include the reference URI in the annotation. 297 Applications can leverage the information provided by a "cite-as" 298 link in a variety of ways, for example: 300 o Bookmarking tools and citation managers can take this preference 301 into account when recording a URI. 303 o Webometrics applications that trace URIs can trace both the access 304 URI and the reference URI. 306 o Discovery tools can support look-up by means of both the access 307 and the reference URI. This includes web archives that typically 308 make archived versions of web resources discoverable by means of 309 the original access URI of the archived resource; they can 310 additionally make these archived resources discoverable by means 311 of the associated reference URI. 313 5. Distinction with Other Relation Types 315 Some existing IANA-registered relationships intuitively resemble the 316 relationship that "cite-as" is intended to convey. But a closer 317 inspection of these candidates provided in the blog posts 318 [identifier-blog], [canonical-blog], and [bookmark-blog] shows that 319 they are not appropriate for various reasons and that a new relation 320 type is required. The remainder of this section provides a summary 321 of the detailed explanations provided in the referenced blog posts. 323 It can readily be seen that the following relation types are not fit 324 for purpose: 326 o "alternate" [RFC4287]: The link target provides an alternate 327 version of the content at the link context. These are typically 328 variants according to dimensions that are subject to content 329 negotiation, for example the same content with varying Content- 330 Type (e.g., application/pdf vs. text/html) and/or Content-Language 331 (e.g., en vs. fr). The representations provided by the context 332 URIs and target URIs in the scenarios of Section 3.1 through 333 Section 3.4 are not variants in the sense intended by [RFC4287], 334 and, as such, the use of "alternate" is not appropriate. 336 o "duplicate" [RFC6249]: The link target is a resource whose 337 available representations are byte-for-byte identical with the 338 corresponding representations of the link context, for example, an 339 identical file on a mirror site. In none of the above scenarios 340 do the link context and the link target provide identical content. 341 As such, the use of "duplicate" is not appropriate. 343 o "related" [RFC4287]: The link target is a resource that is related 344 to the link context. While "related" could be used in all of the 345 above scenarios, its semantics are too vague to convey the 346 specific semantics intended by "cite-as". 348 Two existing IANA-registered relationships deserve closer attention 349 and are discussed in the remainder of this section. 351 5.1. bookmark 353 "bookmark" [W3C.REC-html5-20151028]: The link target provides a URI 354 for the purpose of bookmarking the link context. 356 The intent of "bookmark" is closest to that of "cite-as" in that the 357 link target is intended to be a permalink for the link context, for 358 bookmarking purposes. The relation type dates back to the earliest 359 days of news syndication, before blogs and news feeds had permalinks 360 to identify individual resources that were aggregated into a single 361 page. As such, its intent is to provide permalinks for different 362 sections of an HTML document. It was originally used with HTML 363 elements such as
,

,

, etc. and, more recently, HTML5 364 revised it to be exclusively used with the
element. 365 Moreover, it is explictly excluded from use in the element in 366 HTML , and, as a consequence, in the HTTP Link header that is 367 semantically equivalent. For these technical and semantic reasons, 368 the use of "bookmark" to convey the relationship intented by "cite- 369 as" is not appropriate. 371 A more detailed justification regarding the inappropriatenss of 372 "bookmark", including a thorough overview of its turbulent history, 373 is provided in [bookmark-blog]. 375 5.2. canonical 377 "canonical" [RFC6596]: The meaning of "canonical" is commonly 378 misunderstood on the basis of its brief definition as being "the 379 preferred version of a resource." The description in the abstract of 380 [RFC6596] is more helpful and states that "canonical" is intended to 381 link to a resource that is preferred over resources with duplicative 382 content. A more detailed reading of [RFC6596] clarifies that the 383 intended meaning is preferred for the purpose of content indexing. A 384 typical use case is linking from each page in a multi-page magazine 385 article to a single page version of the article provided for indexing 386 by search engines: the former pages provide content that is 387 duplicative to the superset content that is available at the latter 388 page. 390 The semantics intended by "canonical" as preferred for the purpose of 391 content indexing differ from the semantics intended by "cite-as" as 392 preferred for the purpose of referencing. A further exploration of 393 the various scenarios shows that the use of "canonical" is not 394 appropriate to convey the semantics intended by "cite-as": 396 o Scenario of Section 3.1: The reference URI that is intended to be 397 persistent over time does not serve content that needs to be 398 indexed, it merely redirects to the access URI. Since the meaning 399 intended by "canonical" is "preferred for the purpose of content 400 indexing", it is not appropriate to point at the reference URI 401 (persistent identifier) using the "canonical" relation type. 402 Moreover, Section 6.1 shows that scholarly publishers that assign 403 persistent identifiers, already use the "canonical" relation type 404 for search engine optimization, and how that use contrasts with 405 the intended use of "cite-as". 407 o Scenario of Section 3.2: In most common cases, custodians of 408 resource versioning systems want search engines to index the most 409 recent version of a page and hence would use a "canonical" link to 410 point from version URIs of a resource to the associated generic 411 URI. Wikipedia effectively does this. However, for some resource 412 versioning systems, including Wikipedia, for the purpose of 413 referencing, version URIs are preferred. As such, a "cite-as" 414 link would point from the generic URI to the most recent version 415 URI. That is, in the opposite direction of the "canonical" link. 417 o Scenario of Section 3.3: The content at the link target and the 418 link context are different profiles for a same person. Each 419 profile, not just a preferred one, should be indexed. But a 420 single one could be preferred for referencing. 422 o Scenario of Section 3.4: The content at the link target, if any, 423 would typically be a landing page that includes descriptive 424 metadata pertaining to the multi-resource publication and links to 425 its component resources. Each component resource provides content 426 that is different, not duplicative, to the landing page. 428 A more detailed justification regarding the inappropriatenss of 429 "canonical", including examples, is provided in [canonical-blog]. 431 6. Examples 433 Sections Section 6.1 through Section 6.4 show examples of the use of 434 links with the "cite-as" relation type. They illustrate how the 435 typed links can be used in a response header and/or response body. 437 6.1. Persistent HTTP URI 439 PLOS ONE is one of many scholarly publishers that assigns DOIs to the 440 articles it publishes. For example, https://doi.org/10.1371/ 441 journal.pone.0171057 is the persistent identifier for such an 442 article. Via the DOI resolver, this persistent identifier redirects 443 to http://journals.plos.org/plosone/doi?id=10.1371/ 444 journal.pone.0171057 in the plos.org domain. This URI itself 445 redirects to http://journals.plos.org/plosone/article?id=10.1371/ 446 journal.pone.0171057, which delivers the actual article in HTML. 448 The HTML article contains a element with the "canonical" 449 relation type pointing at itself, 450 http://journals.plos.org/plosone/article?id=10.1371/ 451 journal.pone.0167475. As per Section 5.2, this indicates that the 452 article content at that URI should be indexed by search engines. 454 PLOS ONE can additionally provide a link with the "cite-as" relation 455 type pointing at the persistent identifier to indicate it is the 456 preferred URI for permanent citation of the article. Figure 1 shows 457 the addition of the "cite-as" link both in the HTTP header and the 458 HTML that results from an HTTP GET on the article URI 459 http://journals.plos.org/plosone/article?id=10.1371/ 460 journal.pone.0167475. 462 HTTP/1.1 200 OK 463 Link: ; rel="cite-as" 464 Content-Type: text/html;charset=utf-8 466 467 468 ... 469 470 472 ... 473 474 475 ... 476 477 479 Figure 1: Response to HTTP GET on the URI of a scholarly article 481 6.2. Version URIs 483 The preprint server arXiv.org has a versioning approach like the one 484 described in Section 3.2: 486 o The most recent version of a preprint is at any time available at 487 the same, generic URI. Consider the preprint with generic URI 488 https://arxiv.org/abs/1711.03787. 490 o Each version of the preprint - including the most recent one - has 491 a distinct version URI. The considered preprint has two versions 492 with respective version URIs https://arxiv.org/abs/1711.03787v1 493 (published 10 November 2017) and https://arxiv.org/ 494 abs/1711.03787v2 (published 24 January 2018). 496 A reader who accessed https://arxiv.org/abs/1711.03787 between 10 497 November 2017 and 23 January 2018, obtained the first version of the 498 preprint. Starting 24 January 2018, the second version was served at 499 that URI. In order to support accurate referencing, arXiv.org could 500 implement the "cite-as" link to point from the generic URI to the 501 most recent version URI. In doing so, assuming the existence of 502 reference manager tools that consume "cite-as" links: 504 o The reader who accesses https://arxiv.org/abs/1711.03787 between 505 10 November 2017 and 23 January 2018 would reference 506 https://arxiv.org/abs/1711.03787v1. 508 o The reader who accesses https://arxiv.org/abs/1711.03787 starting 509 24 January 2018 would reference https://arxiv.org/ 510 abs/1711.03787v2. 512 Figure 2 shows the header that arXiv.org would have returned in the 513 first case, in response to a HTTP HEAD on the generic URI 514 https://arxiv.org/abs/1711.03787. 516 HTTP/1.1 200 OK 517 Date: Sun, 24 Dec 2017 16:12:43 GMT 518 Content-Type: text/html; charset=utf-8 519 Link: ; rel="cite-as" 520 Vary: Accept-Encoding,User-Agent 522 Figure 2: Response to HTTP HEAD on the generic URI of the landing 523 page of an arXiv.org preprint 525 6.3. Preferred Profile URI 527 If the access URI is the home page of John Doe, John can add a link 528 with the "cite-as" relation type to it, as a means to convey that he 529 would preferably be referenced by means of the URI of his FOAF 530 profile. Figure 3 shows the response to an HTTP GET on the URI of 531 John's home page. 533 HTTP/1.1 200 OK 534 Content-Type: text/html;charset=utf-8 536 537 538 ... 539 541 ... 542 543 544 ... 545 546 548 Figure 3: Response to HTTP GET on the URI of John Doe's home page 550 6.4. Multi-Resource Publication 552 The Dryad Digital Repository at datadryad.org specializes in hosting 553 and preserving scientific datasets. Each dataset typically consists 554 of multiple resources. For example, the dataset "Data from: Climate, 555 demography, and lek stability in an Amazonian bird" consists of an 556 Excel spreadsheet, a csv file, and a zip file. Each of these 557 resources have different content and are accessible at their 558 respective URIs. In addition, the dataset has a landing page at 559 https://datadryad.org/resource/doi:10.5061/dryad.5d23f. 561 Each of these resources should be permanently cited by means of the 562 persistent identifier that was assigned to the entire dataset as an 563 intellectual publication, i.e. https://doi.org/10.5061/dryad.5d23f. 564 To that end, the Dryad Digital Repository can add "cite-as" links 565 pointing from the URIs of each of these resources to 566 https://doi.org/10.5061/dryad.5d23f. This is shown in Figure 4 for 567 the csv file that is a component resource of the dataset, through use 568 of the HTTP Link header. 570 HTTP/1.1 200 OK 571 Date: Tue, 12 Jun 2018 19:19:22 GMT 572 Last-Modified: Wed, 17 Feb 2016 18:37:02 GMT 573 Content-Type: text/csv;charset=ISO-8859-1 574 Content-Length: 25414 575 Link: ; rel="cite-as" 577 DATE,Year,PLOT/TRAIL,LOCATION,SPECIES CODE,BAND NUM,COLOR,SEX,AGE,TAIL,WING, 578 TARSUS,NARES,DEPTH,WIDTH,WEIGHT 579 6/26/02,2002,DANTA,325,PIPFIL,969,B/O,M,AHY,80,63,16,7.3,3.9,4.1,14.4 580 ... 581 2/3/13,2013,LAGO,,PIPFIL,BR-5095,O/YPI,M,SCB,78,65.5,14.2,7.5,3.8,3.7,14.3 583 Figure 4: Response to HTTP GET on the URI of a csv file that is a 584 component of a scientfic dataset 586 7. IANA Considerations 588 7.1. Link Relation Type: cite-as 590 The link relation type below has been registered by IANA per 591 Section 2.1.1 of [RFC8288]: 593 Relation Name: cite-as 594 Description: A link with the "cite-as" relation type indicates 595 that the link target is preferred over the link context for the 596 purpose of permanent citation. 598 Reference: [[ This document ]] 600 8. Security Considerations 602 In cases where there is no way for the agent to automatically verify 603 the correctness of the reference URI (cf. Section 4), out-of-band 604 mechanisms might be required to establish trust. 606 If a trusted site is compromised, the "cite-as" link relation could 607 be used with malicious intent to supply misleading URIs for 608 referencing. Use of these links might direct user agents to an 609 attacker's site, break the referencing record they are intended to 610 support, or corrupt algorithmic interpretation of referencing data. 612 9. References 614 9.1. Normative References 616 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 617 Requirement Levels", BCP 14, RFC 2119, 618 DOI 10.17487/RFC2119, March 1997, 619 . 621 [RFC4287] Nottingham, M., Ed. and R. Sayre, Ed., "The Atom 622 Syndication Format", RFC 4287, DOI 10.17487/RFC4287, 623 December 2005, . 625 [RFC5988] Nottingham, M., "Web Linking", RFC 5988, 626 DOI 10.17487/RFC5988, October 2010, 627 . 629 [RFC6249] Bryan, A., McNab, N., Tsujikawa, T., Poeml, P., and H. 630 Nordstrom, "Metalink/HTTP: Mirrors and Hashes", RFC 6249, 631 DOI 10.17487/RFC6249, June 2011, 632 . 634 [RFC6596] Ohye, M. and J. Kupke, "The Canonical Link Relation", 635 RFC 6596, DOI 10.17487/RFC6596, April 2012, 636 . 638 [RFC8288] Nottingham, M., "Web Linking", RFC 8288, 639 DOI 10.17487/RFC8288, October 2017, 640 . 642 [W3C.REC-html5-20151028] 643 Hickson, I., Berjon, R., Faulkner, S., Leithead, T., Doyle 644 Navara, E., O'Connor, E., and S. Pfeiffer, "HTML5", World 645 Wide Web Consortium Recommendation REC-HTML5-20141028, 646 October 2014, 647 . 649 9.2. Informative References 651 [bookmark-blog] 652 Nelson, M. and H. Van de Sompel, "rel=bookmark also does 653 not mean what you think it means", August 2017, 654 . 657 [canonical-blog] 658 Nelson, M. and H. Van de Sompel, "rel=canonical does not 659 mean what you think it means", August 2017, . 663 [CoolURIs] 664 Berners-Lee, T., "Cool URIs don't change", World Wide Web 665 Consortium Style, 1998, 666 . 668 [DOI-URLs] 669 Hendricks, G., "Display guidelines for Crossref DOIs", 670 June 2017, 671 . 673 [DOIs] "Information and documentation - Digital object identifier 674 system", ISO 26324:2012(en), 2012, 675 . 678 [draft-kunze-ark-18] 679 Kunze, J. and R. Rodgers, "The ARK Identifier Scheme", 680 Internet Draft draft-kunze-ark-18, April 2013, 681 . 683 [FOAF] Brickley, D. and L. Miller, "FOAF Vocabulary Specification 684 0.99", January 2014, . 686 [identifier-blog] 687 Nelson, M. and H. Van de Sompel, "Linking to Persistent 688 Identifiers with rel=identifier", July 2016, . 692 [PIDs-must-be-used] 693 Van de Sompel, H., Klein, M., and S. Jones, "Persistent 694 URIs Must Be Used To Be Persistent", February 2016, 695 . 697 [PURLs] "Persistent uniform resource locator", April 2017, 698 . 701 Appendix A. Acknowledgements 703 Thanks for comments and suggestions provided by Martin Klein, Harihar 704 Shankar, Peter Williams, John Howard, Mark Nottingham, Graham Klyne. 706 Authors' Addresses 708 Herbert Van de Sompel 709 Los Alamos National Laboratory 711 Email: herbertv@lanl.gov 712 URI: https://orcid.org/0000-0002-0715-6126 714 Michael Nelson 715 Old Dominion University 717 Email: mln@cs.odu.edu 718 URI: http://www.cs.odu.edu/~mln/ 720 Geoffrey Bilder 721 Crossref 723 Email: gbilder@crossref.org 724 URI: https://www.crossref.org/authors/geoffrey-bilder/ 726 John Kunze 727 California Digital Library 729 Email: jak@ucop.edu 730 URI: https://orcid.org/0000-0001-7604-8041 731 Simeon Warner 732 Cornell University 734 Email: simeon.warner@cornell.edu 735 URI: https://orcid.org/0000-0002-7970-7855