idnits 2.17.1 draft-nottingham-safe-hint-09.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (April 17, 2019) is 1836 days in the past. Is this intentional? Checking references for intended status: Informational ---------------------------------------------------------------------------- ** Obsolete normative reference: RFC 7234 (Obsoleted by RFC 9111) Summary: 1 error (**), 0 flaws (~~), 1 warning (==), 1 comment (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 Network Working Group M. Nottingham 3 Internet-Draft April 17, 2019 4 Intended status: Informational 5 Expires: October 19, 2019 7 The "safe" HTTP Preference 8 draft-nottingham-safe-hint-09 10 Abstract 12 This specification defines a "safe" preference for HTTP requests that 13 expresses a desire to avoid objectionable content, according to the 14 definition of that term by the origin server. 16 Support for this preference by clients and servers is optional. 18 Status of This Memo 20 This Internet-Draft is submitted in full conformance with the 21 provisions of BCP 78 and BCP 79. 23 Internet-Drafts are working documents of the Internet Engineering 24 Task Force (IETF). Note that other groups may also distribute 25 working documents as Internet-Drafts. The list of current Internet- 26 Drafts is at https://datatracker.ietf.org/drafts/current/. 28 Internet-Drafts are draft documents valid for a maximum of six months 29 and may be updated, replaced, or obsoleted by other documents at any 30 time. It is inappropriate to use Internet-Drafts as reference 31 material or to cite them other than as "work in progress." 33 This Internet-Draft will expire on October 19, 2019. 35 Copyright Notice 37 Copyright (c) 2019 IETF Trust and the persons identified as the 38 document authors. All rights reserved. 40 This document is subject to BCP 78 and the IETF Trust's Legal 41 Provisions Relating to IETF Documents 42 (https://trustee.ietf.org/license-info) in effect on the date of 43 publication of this document. Please review these documents 44 carefully, as they describe your rights and restrictions with respect 45 to this document. Code Components extracted from this document must 46 include Simplified BSD License text as described in Section 4.e of 47 the Trust Legal Provisions and are provided without warranty as 48 described in the Simplified BSD License. 50 Table of Contents 52 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 53 2. The "safe" Preference . . . . . . . . . . . . . . . . . . . . 3 54 3. Implementation Status . . . . . . . . . . . . . . . . . . . . 4 55 4. Security Considerations . . . . . . . . . . . . . . . . . . . 5 56 5. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 6 57 6. References . . . . . . . . . . . . . . . . . . . . . . . . . 6 58 Appendix A. Acknowledgements . . . . . . . . . . . . . . . . . . 7 59 Appendix B. Sending "safe" from Web Browsers . . . . . . . . . . 7 60 Appendix C. Supporting "safe" on Web Sites . . . . . . . . . . . 7 61 Author's Address . . . . . . . . . . . . . . . . . . . . . . . . 8 63 1. Introduction 65 Many Web sites have a "safe" mode, to assist those who don't want to 66 be exposed (or have their children exposed) to content to which they 67 might object. 69 However, that goal is often difficult to achieve, because of the need 70 to go to every Web site that might be used, navigate to the 71 appropriate page (possibly creating an account along the way) to get 72 a cookie [RFC6265] set in the browser, for each browser on every 73 device used. 75 A more manageable approach is for the browser to proactively indicate 76 a preference for safe content. A user agent that supports doing so 77 (whether it be an individual browser, or through an Operating System 78 HTTP library) need only be configured once to assure that the 79 preference is advertised to a set of sites, or even all sites. 81 This specification defines how to declare this desire in requests as 82 a HTTP Preference [RFC7240]. 84 Note that this specification does not define what content might be 85 considered objectionable, and so the concept of "safe" is also not 86 precisely defined. Rather, the term is interpreted by the server and 87 within the scope of each Web site that chooses to act upon this 88 information. 90 That said, the intent of "safe" is to allow end users (or those 91 acting on their behalf) to express a desire to avoid content that is 92 considered objectionable within the cultural context of that site; 93 usually (but not always) content that is unsuitable for minors. The 94 "safe" preference is not intended to be used for other purposes. 96 Furthermore, sending "safe" does not guarantee that the Web site will 97 use it, nor that it will apply a concept of "objectionable" that is 98 consistent with the requester's views. As such, its effect can be 99 described as "best effort," and not to be relied upon. In other 100 words, sending the preference is no more reliable than going to each 101 Web site and manually selecting a "safe" mode, but it is considerably 102 easier. 104 It is also important to note that the "safe" preference is not a 105 reliable indicator that the end user is a child; other users might 106 have a desire for unobjectionable content, and some children might 107 browse without the preference being set. 109 Note also that the cultural context applies to the hosting location 110 of a site, the content provider, and the source of the content. It 111 cannot be guaranteed that a user-agent and origin server will have 112 the same view of the concept of what is objectionable. 114 Simply put, it is a statement by (or on behalf of) the end user to 115 the effect "If your site has a 'safe' setting, this user is hereby 116 opting into that, according to your definition of the term." 118 1.1. Notational Conventions 120 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 121 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 122 "OPTIONAL" in this document are to be interpreted as described in BCP 123 14 [RFC2119] [RFC8174] when, and only when, they appear in all 124 capitals, as shown here. 126 2. The "safe" Preference 128 When present in a request, the "safe" preference indicates that the 129 user prefers that the origin server to not respond with content which 130 is designated as objectionable, according to the origin server's 131 definition of the concept. 133 For example, a request that includes the "safe" preference: 135 GET /foo.html HTTP/1.1 136 Host: www.example.org 137 User-Agent: ExampleBrowser/1.0 138 Prefer: safe 140 Typically, user agents that emit the "safe" preference will include 141 it in all requests with the "https" URI scheme, although some might 142 expose finer-grained controls over when it is sent; this ensures that 143 the preference is available to the applicable resources. User agents 144 MUST NOT emit the "safe" preference on requests with the "http" URI 145 scheme (see Section 4). See Appendix B for more information about 146 configuring the set of resources "safe" is sent to. 148 Safe MAY be implemented in common HTTP libraries (e.g., an operating 149 system might choose to insert the preference in requests based upon 150 system-wide configuration). 152 Origin servers that utilize the "safe" preference ought to document 153 that they do so, along with the criteria that they use to denote 154 objectionable content. If a server has more fine-grained degrees of 155 "safety", it SHOULD select a reasonable default to use, and document 156 that; it MAY use additional mechanisms (e.g., cookies [RFC6265]) to 157 fine-tune. 159 A response corresponding to the request above might have headers that 160 look like this: 162 HTTP/1.1 200 OK 163 Transfer-Encoding: chunked 164 Content-Type: text/html 165 Preference-Applied: safe 166 Server: ExampleServer/2.0 167 Vary: Prefer 169 Here, the Preference-Applied response header ([RFC7240]) indicates 170 that the site has applied the preference. Servers are not required 171 to send Preference-Applied (even when they have applied the 172 preference), but are encouraged to where possible. 174 Note that the Vary response header needs to be sent if the response 175 is cacheable and might change depending on the value of the "Prefer" 176 header. This is not only true for those responses that are "safe", 177 but also the default "unsafe" response. 179 See [RFC7234] Section 4.1 for more information the interaction 180 between Vary and Web caching. 182 See Appendix C for additional advice specific to Web servers wishing 183 to use "safe". 185 3. Implementation Status 187 _Note to RFC Editor: Please remove this section before publication._ 189 This section records the status of known implementations of the 190 protocol defined by this specification at the time of posting of this 191 Internet-Draft. Please note that the listing of any individual 192 implementation here does not imply endorsement by the IETF. 194 Furthermore, no effort has been spent to verify the information 195 presented here that was supplied by IETF contributors. This is not 196 intended as, and must not be construed to be, a catalog of available 197 implementations or their features. Readers are advised to note that 198 other implementations may exist. 200 o Microsoft Internet Explorer - see https://support.microsoft.com/ 201 en-hk/help/2980016/ 203 o Microsoft Bing - see https://developer.microsoft.com/en-us/ 204 microsoft-edge/testdrive/demos/familysearch/ 206 o Mozilla Firefox - see https://support.mozilla.org/en-US/kb/block- 207 and-unblock-websites-parental-controls-firef 209 o Cisco - see http://blogs.cisco.com/security/filtering-explicit- 210 content 212 4. Security Considerations 214 The "safe" preference is not a secure mechanism; it can be inserted 215 or removed by intermediaries with access to the request stream (e.g. 216 for "http" URLs). Therefore, it is prohibited from being included in 217 requests with the "http" scheme. 219 Its presence reveals limited information about the user, which may be 220 of small assistance in "fingerprinting" the user by sites. 221 Therefore, user agents SHOULD NOT include it in requests when the 222 user has expressed a desire to avoid such attacks (e.g., some forms 223 of "private mode" browsing). 225 By its nature, including "safe" in requests does not assure that all 226 content will actually be safe; it is only when servers elect to honor 227 it that content might be "safe". 229 Even then, a malicious server might adapt content so that it is even 230 less "safe" (by some definition of the word). As such, this 231 mechanism on its own is not enough to assure that only "safe" content 232 is seen; those who wish to ensure that will need to combine its use 233 with other techniques (e.g., content filtering). 235 Furthermore, the server and user may have differing ideas regarding 236 the semantics of "safe." As such, the "safety" of the user's 237 experience when browsing from site to site as well as over time might 238 (and probably will) change. 240 5. IANA Considerations 242 This specification registers the following entry in the "HTTP 243 Preferences" registry [RFC7240]: 245 o Preference: safe 247 o Value: (no value) 249 o Description: Indicates that "safe" / "unobjectionable" content is 250 preferred. 252 o Reference: (this document) 254 o Notes: 256 6. References 258 6.1. Normative References 260 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 261 Requirement Levels", BCP 14, RFC 2119, 262 DOI 10.17487/RFC2119, March 1997, 263 . 265 [RFC7234] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 266 Ed., "Hypertext Transfer Protocol (HTTP/1.1): Caching", 267 RFC 7234, DOI 10.17487/RFC7234, June 2014, 268 . 270 [RFC7240] Snell, J., "Prefer Header for HTTP", RFC 7240, 271 DOI 10.17487/RFC7240, June 2014, 272 . 274 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 275 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 276 May 2017, . 278 6.2. Informative References 280 [RFC6265] Barth, A., "HTTP State Management Mechanism", RFC 6265, 281 DOI 10.17487/RFC6265, April 2011, 282 . 284 Appendix A. Acknowledgements 286 Thanks to Alissa Cooper, Ilya Grigorik, Emma Llanso, Jeff Hughes, 287 Lorrie Cranor, Doug Turner and Dave Crocker for their comments. 289 Appendix B. Sending "safe" from Web Browsers 291 As discussed in Section 2, there are many possible ways for the 292 "safe" preference to be generated. One possibility is for a Web 293 browser to allow its users to configure the preference to be sent. 295 When doing so, it is important not to misrepresent the preference as 296 binding to Web sites. For example, an appropriate setting might be a 297 checkbox with wording such as: 299 [] Request "safe" content from Web sites 301 ... along with further information available upon request. 303 Browsers might also allow the "safe" preference to be "locked" - that 304 is, prevent modification without administrative access, or a 305 passcode. 307 Note that this specification does not require browsers to send "safe" 308 on all requests, although that is one possible implementation; e.g., 309 alternate implementation strategies include blacklists and 310 whitelists. 312 Appendix C. Supporting "safe" on Web Sites 314 Web sites that allow configuration of a "safe" mode (for example, 315 using a cookie) can add support for the "safe" preference 316 incrementally; since the preference will not be supported by all 317 clients immediately, it is necessary to have another way to configure 318 it. 320 When honoring the safe preference, it is important that it not be 321 possible to disable it through the Web site's interface, since "safe" 322 may be configured and locked down by the browser or computer's 323 administrator (e.g., a parent). If the site has such a means of 324 configuration (e.g., stored user preferences) and the safe preference 325 is received in a request, the "safer" interpretation ought to be 326 used. 328 The appropriate level of "safety" is a site-specific decision. When 329 selecting it, sites ought to bear in mind that disabling the 330 preference might be considerably more onerous than through other 331 means, especially if the preference is generated based upon Operating 332 System configuration. 334 Sites might offer different levels of "safeness" through Web 335 configuration, they will need to either inform their users of what 336 level the "safe" hint corresponds to, or provide them with some means 337 of adjusting it. 339 If the user expresses a wish to disable "safe" mode, the site can 340 remind them that the safe preference is being sent, and ask them to 341 consult their administrator (since "safe" might be set by a locked- 342 down Operating System configuration). 344 As explained in Section 2, responses that change based upon the 345 presence of the "safe" preference need to either carry the "Vary: 346 Prefer" response header field, or be uncacheable by shared caches 347 (e.g., with a "Cache-Control: private" response header field). This 348 is to avoid an unsafe cached response being served to a client that 349 prefers safe content (or vice versa). 351 Author's Address 353 Mark Nottingham 355 EMail: mnot@mnot.net 356 URI: https://www.mnot.net/