Simplified-Bookmarks-Spec

Overview

The contents of this repository define a specification describing the format of bookmark data shared between clients and server in the Library Simplified ecosystem. The intention is to declare a common format for bookmarks that clients on different platforms (Web, iOS, Android) can use to synchronize reading positions. The specification is described as executable Literate Haskell and can be executed and inspected directly using ghci.

$ ghci -W -Wall -Werror -pgmL markdown-unlit Bookmarks.lhs

Typographic Conventions

Within this document, commands given at the GHCI prompt are prefixed with *Bookmarks> to indicate that the commands are being executed within the Bookmarks module.

The main specification definitions are given in the Bookmarks module:

{-# LANGUAGE Haskell2010, ExplicitForAll #-}
module Bookmarks where

import qualified Data.Map as DM

The specification makes references to RFC 3986 URI values. Within this specification, URIs are treated as opaque strings.

type URI = String

Terminology

• User: A human (typically a library patron) using one or more of the Library Simplified applications.

• Client: An application running on a user's device. This can refer to native applications such as SimplyE, or the web-based interface.

• Bookmark: A stored position within a publication that can be used to navigate to that position at a later date.

Web Annotations

The base format for bookmark data is the W3C Web Annotations format. The bookmark data described in this specification is expressed in terms of an annotation with a set of strictly-defined required and optional fields.

Compatibility

Historically, the Library Simplified applications have not had a consistent standard with regard to how bookmarks are serialized. Applications MAY accept bookmarks in older formats, but MUST serialize all new bookmarks using the format described here. This allows for a degree of migration compatibility; over time, all bookmarks in circulation will effectively be converted to the new format.

Locators

A Locator uniquely identifies a position within a book. There are specific types of locators tailored to specific reading contexts and book formats, because each of those contexts typically has a different means to specify locations within books.

A Locator is one of the following:

• LocatorLegacyCFI
• LocatorHrefProgression
• LocatorPage
• LocatorAudioBookTime

data Locator
  = L_CFI             LocatorLegacyCFI
  | L_HrefProgression LocatorHrefProgression
  | L_Page            LocatorPage
  | L_AudioBookTime   LocatorAudioBookTime
  deriving (Eq, Ord, Show)

Chapter Progression

A progression value is a real number in the range [0, 1] where 0 is the beginning of a chapter, and 1 is the end of the chapter.

data Progression
  = Progression Double
  deriving (Eq, Ord, Show)

progression :: Double -> Progression
progression x =
  if (x >= 0.0 && x <= 1.0)
  then Progression x
  else error "Progression must be in the range [0,1]"

LocatorLegacyCFI

A LocatorLegacyCFI value consists of a set of properties used to express content fragment identifiers, such as those frequently consumed by the Readium 1 reader. There is very little consistency in the values consumed by Library Simplified applications between platforms, hence the legacy status of this locator type and the optional fields. Applications are encouraged to attempt to write a non-Nothing value to at least one of the fields.

The lcIdRef property refers to the id value of the spine item of the target EPUB. This, in practice, is the idRef value returned by Readium 1.

The lcContentCFI property refers to the content fragment identifier used to point to a specific element within the specified spine item.

data LocatorLegacyCFI = LocatorLegacyCFI {
  lcIdRef              :: Maybe String,
  lcContentCFI         :: Maybe String,
  lcChapterProgression :: Maybe Progression
} deriving (Eq, Ord, Show)

LocatorHrefProgression

A LocatorHrefProgression consists of a URI that uniquely identifies a chapter within a publication, and a progression value.

LocatorHrefProgression values are used to describe the positions of books being consumed in the Readium 2 reader and are expected to be the preferred form for sharing book locations for the forseeable future.

data LocatorHrefProgression = LocatorHrefProgression {
  hpChapterHref        :: URI,
  hpChapterProgression :: Progression
} deriving (Eq, Ord, Show)

LocatorPage

A LocatorPage consists of a single integer value that uniquely identifies a page within an integer page-based publication such as PDF.

A Page number must be non-negative.

data Page
  = Page Integer
  deriving (Eq, Ord, Show)

page :: Integer -> Page
page x =
  if (x >= 0)
  then Page x
  else error "Page must be in non-negative"

data LocatorPage = LocatorPage {
  ipPage :: Page
} deriving (Eq, Ord, Show)

LocatorAudioBookTime

A LocatorAudioBookTime consists of a part and chapter number, and a time in milliseconds. This is expected to uniquely identify a position within an audio book.

Part and Chapter numbers must be non-negative, as must TimeMilliseconds values.

data Part
  = Part Integer
  deriving (Eq, Ord, Show)

data Chapter
  = Chapter Integer
  deriving (Eq, Ord, Show)

data Title
  = Title String
  deriving (Eq, Ord, Show)

data AudiobookID
  = AudiobookID String
  deriving (Eq, Ord, Show)

data Duration
  = Duration Integer
  deriving (Eq, Ord, Show)

data TimeMilliseconds
  = TimeMilliseconds Integer
  deriving (Eq, Ord, Show)

part :: Integer -> Part
part x =
  if (x >= 0)
  then Part x
  else error "Part must be in non-negative"

chapter :: Integer -> Chapter
chapter x =
  if (x >= 0)
  then Chapter x
  else error "Chapter must be in non-negative"

duration :: Integer -> Duration
duration x =
  if (x >= 0)
  then Duration x
  else error "Duration must be in non-negative"

time :: Integer -> TimeMilliseconds
time x =
  if (x >= 0)
  then TimeMilliseconds x
  else error "TimeMilliseconds must be in non-negative"

data LocatorAudioBookTime = LocatorAudioBookTime {
  abtPart    :: Part,
  abtChapter    :: Chapter,
  abtTitle    :: Title,
  abtAudiobookID  :: AudiobookID,
  abtDuration      :: Duration,
  abtTime          :: TimeMilliseconds
} deriving (Eq, Ord, Show)

Interpretation

Audiobook players differ in their support for part values. Some manifests will not contain part numbers, whilst other manifests are provided to players that actually require them in order to work at all. Manifests that represent Findaway audiobooks, for example, include both findaway:part and findaway:sequence values in each entry of the manifest's readingOrder, and the Findaway player cannot work without access to these values. Other manifest formats do not include part and chapter numbers at all, and simply assume that players will walk through the list of chapters in manifest declaration order. This raises the question of how the abtPart and abtChapter fields in LocatorAudioBookTime values should be interpreted when loaded into an arbitrary audiobook player.

For Findaway audiobooks, the abtPart and abtChapter fields for a serialized locator should be equal to the findaway:part and findaway:sequence fields, respectively, of the readingOrder manifest element that was active when the locator was serialized.

For all other audiobooks, the abtPart field should be 0, and the abtChapter field should be equal to the index of the readingOrder manifest element that was active when the locator was serialized.

When loading a locator value L in a Findaway player, search for a readingOrder element that contains a findaway:part and findaway:sequence value equal to the L.abtPart and L.abtChapter fields, respectively.

LocatorAudioBookTime L;

for (element in readingOrder) {
  if (element.part == L.abtPart && element.chapter == L.abtChapter) {
    openForReading (element);
    return;
  }
}

throw ErrorNoSuchChapter();

When loading a locator value L in any other player, use readingOrder[L.abtChapter].

LocatorAudioBookTime L;

if (L.abtChapter < readingOrder.size) {
  openForReading (readingOrder [L.abtChapter]);
  return;
}

throw ErrorNoSuchChapter();

Serialization

Locators MUST be serialized using the following JSON schema:

{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "urn:org.librarysimplified.bookmarks:locator:1.0",
  "title": "Simplified Bookmark Locator",
  "description": "A bookmark locator",
  "type": "object",
  "oneOf": [
    {
      "type": "object",
      "properties": {
        "@type": {
          "description": "The type of locator",
          "type": "string",
          "pattern": "LocatorHrefProgression"
        },
        "href": {
          "description": "The unique identifier for a chapter (hpChapterHref)",
          "type": "string"
        },
        "progressWithinChapter": {
          "description": "The progress within a chapter (hpChapterProgression)",
          "type": "number",
          "minimum": 0.0,
          "maximum": 1.0
        }
      },
      "required": [
        "@type",
        "href",
        "progressWithinChapter"
      ]
    },

    {
      "type": "object",
      "properties": {
        "@type": {
          "description": "The type of locator",
          "type": "string",
          "pattern": "LocatorLegacyCFI"
        },
        "idref": {
          "description": "The unique identifier for a chapter (lcIdRef)",
          "type": "string"
        },
        "contentCFI": {
          "description": "The content fragment identifier (lcContentCFI)",
          "type": "string"
        },
        "progressWithinChapter": {
          "description": "The progress within a chapter (lcChapterProgression)",
          "type": "number",
          "minimum": 0.0,
          "maximum": 1.0
        }
      },
      "required": [
        "@type"
      ]
    },

    {
      "type": "object",
      "properties": {
        "@type": {
          "description": "The type of locator",
          "type": "string",
          "pattern": "LocatorPage"
        },
        "page": {
          "description": "The integer page number (ipPage)",
          "type": "number",
          "minimum": 0
        }
      },
      "required": [
        "@type",
        "page"
      ]
    },

    {
      "type": "object",
      "properties": {
        "@type": {
          "description": "The type of locator",
          "type": "string",
          "pattern": "LocatorAudioBookTime"
        },
        "part": {
          "description": "The part number (abtPart)",
          "type": "number",
          "minimum": 0
        },
        "chapter": {
          "description": "The chapter number (abtChapter)",
          "type": "number",
          "minimum": 0
        },
        "title": {
          "description": "The title (abtTitle)",
          "type": "string"
        },
        "audiobookID": {
          "description": "The audiobook ID (abtAudiobookID)",
          "type": "string"
        },
        "duration": {
          "description": "The duration (abtDuration)",
          "type": "number",
          "minimum": 0
        }
        "time": {
          "description": "The time (abtTime)",
          "type": "number",
          "minimum": 0
        }
      },
      "required": [
        "@type",
        "part",
        "chapter",
    "title",
    "audiobookID",
    "duration",
        "time"
      ]
    }
  ]
}

A LocatorHrefProgression value MUST be serialized using the schema with @type = LocatorHrefProgression.

A LocatorLegacyCFI value MUST be serialized using the schema with @type = LocatorLegacyCFI.

A LocatorPage value MUST be serialized using the schema with @type = LocatorPage.

A LocatorAudioBookTime value MUST be serialized using the schema with @type = LocatorAudioBookTime.

When encountering a locator without a @type property, applications SHOULD assume that the format is LocatorLegacyCFI and parse it accordingly.

Examples

An example of a valid, serialized locator is given in valid-locator-0.json:

{
  "@type": "LocatorHrefProgression",
  "href": "/xyz.html",
  "progressWithinChapter": 0.666
}

An example of a valid, serialized locator is given in valid-locator-1.json:

{
  "@type": "LocatorLegacyCFI",
  "idref": "xyz-html",
  "contentCFI": "/4/2/2/2",
  "progressWithinChapter": 0.25
}

An example of a valid, serialized locator is given in valid-locator-2.json:

{
  "@type": "LocatorPage",
  "page": 23
}

An example of a valid, serialized locator is given in valid-locator-3.json:

{
  "@type": "LocatorAudioBookTime",
  "part": 3,
  "chapter": 32,
  "title": "Chapter title",
  "audiobookID": "urn:uuid:b309844e-7d4e-403e-945b-fbc78acd5e03",
  "duration": 190000,
  "time": 78000
}

Bookmarks

A Bookmark is a Web Annotation with the following data:

• A body containing optional metadata such as the reader's current progress through the entire publication.
• A motivation indicating the type of bookmark.
• A target that uniquely identifies the publication, and includes a selector that includes a serialized Locator.
• An optional id value that uniquely identifies the bookmark. This is typically assigned by the server, and a server publishing bookmarks to a client MUST include this value in each bookmark.

data Bookmark = Bookmark {
  bookmarkId         :: Maybe URI,
  bookmarkTarget     :: BookmarkTarget,
  bookmarkMotivation :: Motivation,
  bookmarkBody       :: BookmarkBody
} deriving (Eq, Show)

Bodies

A body contains metadata that applications MAY use to derive extra data for display in the application. Currently, bodies are defined as simple maps of strings to strings with a couple of extra mandatory fields.

data BookmarkBody = BookmarkBody {
  bodyDeviceId :: String,
  bodyTime     :: String,
  bodyOthers   :: DM.Map String String
} deriving (Eq, Show)

The bodyTime field MUST contain an RFC 3339 timestamp indicating the creation time of the bookmark. The timestamp MUST be in the UTC time zone.

The bodyDeviceId field denotes the unique identifier of the device that created the bookmark. This is typically a UUID value expressed as a URN, such as:

urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c

Clients that do not have access to an identifier in this form SHOULD use a string value of null. Note that this does mean serializing the literal value null as a quoted string:

{
  ...
  "http://librarysimplified.org/terms/device" = "null",
  ...
}

Targets

A target uniquely identifies a publication, and uses a Locator to uniquely identify a position within that publication. The value of the targetSource field is typically taken from metadata included in the publication, or from the OPDS feed that originally delivered the publication.

data BookmarkTarget = BookmarkTarget {
  targetLocator :: Locator,
  targetSource  :: String
} deriving (Eq, Show)

Motivations

A motivation is value that simply indicates whether a bookmark was created explicitly by the user, or created implicitly by the application each time the user navigates to a new page. Explicitly created bookmarks are denoted by the bookmarking motivation, whilst implicitly created bookmarks are denoted by the idling motivation. In practice, there is exactly one idling bookmark in the user's set of bookmarks at any given time, and the reading application effectively replaces the current idling bookmark each time the user turns a page in a given publication.

data Motivation
  = Bookmarking
  | Idling
  deriving (Eq, Ord, Show)

JSON Serialization

Bookmarks MUST be serialized as Web Annotation values according to the following rules:

• Body values MUST be serialized as string-typed properties with string-typed values in the annotation's body property, with the following extra constraints:

  • The bodyDeviceId field MUST be serialized as string-typed property with the name http://librarysimplified.org/terms/device.
  • The bodyTime field MUST be serialized as string-typed property with the name http://librarysimplified.org/terms/time.

• Motivation values MUST be serialized as one of the two possible string values according to the motivationJSON function:

motivationJSON :: Motivation -> String
motivationJSON Bookmarking = "http://www.w3.org/ns/oa#bookmarking"
motivationJSON Idling      = "http://librarysimplified.org/terms/annotation/idling"

• Target values MUST be serialized with:
  • A selector property containing an object with:
    • A type property equal to "oa:FragmentSelector".
    • A value property containing a Locator serialized as a string value.
  • A source property with a string value that uniquely identifies the publication.

If present, the bookmark's id field MUST be serialized as an id property with a string value equal to the id field.

The bookmark SHOULD be serialized with a type property set to the string value "Annotation", and a @context property set to the string "http://www.w3.org/ns/anno.jsonld".

An example of a valid bookmark is given in valid-bookmark-0.json:

{
  "@context": "http://www.w3.org/ns/anno.jsonld",
  "type": "Annotation",
  "id": "urn:uuid:715885bc-23d3-4d7d-bd87-f5e7a042c4ba",

  "body": {
    "http://librarysimplified.org/terms/time": "2021-03-12T16:32:49Z",
    "http://librarysimplified.org/terms/device": "urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c"
  },

  "motivation": "http://librarysimplified.org/terms/annotation/idling",

  "target": {
    "selector": {
      "type": "oa:FragmentSelector",
      "value": "{\n  \"@type\": \"LocatorHrefProgression\",\n  \"href\": \"/xyz.html\",\n  \"progressWithinChapter\": 0.666\n}\n"
    },
    "source": "urn:uuid:1daa8de6-94e8-4711-b7d1-e43b572aa6e0"
  }
}

Test Cases

This specification includes a number of test cases. Applications MUST include unit tests that give the results specified below for each test case, and MUST succeed or fail for the reasons specified. For tests cases that must succeed, their required interpretation is listed below.

File	Type	Result	Reason
invalid-bookmark-0.json	bookmark	❌ failure	Missing a body
invalid-bookmark-1.json	bookmark	❌ failure	Missing a motivation
invalid-bookmark-2.json	bookmark	❌ failure	Missing a target
invalid-bookmark-3.json	bookmark	❌ failure	Target selector has an invalid type
invalid-bookmark-4.json	bookmark	❌ failure	Target selector has an invalid value
invalid-bookmark-5.json	bookmark	❌ failure	Body lacks device ID property
invalid-bookmark-6.json	bookmark	❌ failure	Body lacks time property
invalid-bookmark-7.json	bookmark	❌ failure	Target selector lacks page
invalid-locator-1.json	locator	❌ failure	Missing href property
invalid-locator-2.json	locator	❌ failure	Missing progressWithinChapter property
invalid-locator-3.json	locator	❌ failure	Chapter progression is negative
invalid-locator-4.json	locator	❌ failure	Chapter progression is greater than 1.0
invalid-locator-5.json	locator	❌ failure	Chapter number is negative
invalid-locator-6.json	locator	❌ failure	Page number is negative
valid-bookmark-0.json	bookmark	✅ success	Valid bookmark
valid-bookmark-1.json	bookmark	✅ success	Valid bookmark
valid-bookmark-2.json	bookmark	✅ success	Valid bookmark
valid-bookmark-3.json	bookmark	✅ success	Valid bookmark
valid-bookmark-4.json	bookmark	✅ success	Valid bookmark
valid-bookmark-5.json	bookmark	✅ success	Valid bookmark
valid-locator-0.json	locator	✅ success	Valid locator
valid-locator-1.json	locator	✅ success	Valid locator
valid-locator-2.json	locator	✅ success	Valid locator
valid-locator-3.json	locator	✅ success	Valid locator

valid-bookmark-0.json

validBookmark0 :: Bookmark
validBookmark0 = Bookmark {
  bookmarkId   = Just "urn:uuid:715885bc-23d3-4d7d-bd87-f5e7a042c4ba",
  bookmarkBody = BookmarkBody {
    bodyDeviceId = "urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c",
    bodyTime     = "2021-03-12T16:32:49Z",
    bodyOthers   = DM.empty
  },
  bookmarkMotivation = Idling,
  bookmarkTarget = BookmarkTarget {
    targetLocator = L_HrefProgression $ LocatorHrefProgression {
      hpChapterHref        = "/xyz.html",
      hpChapterProgression = progression 0.666
    },
    targetSource = "urn:uuid:1daa8de6-94e8-4711-b7d1-e43b572aa6e0"
  }
}

valid-bookmark-1.json

validBookmark1 :: Bookmark
validBookmark1 = Bookmark {
  bookmarkId   = Nothing,
  bookmarkBody = BookmarkBody {
    bodyDeviceId = "urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c",
    bodyTime     = "2021-03-12T16:32:49Z",
    bodyOthers   = DM.empty
  },
  bookmarkMotivation = Idling,
  bookmarkTarget = BookmarkTarget {
    targetLocator = L_HrefProgression $ LocatorHrefProgression {
      hpChapterHref        = "/xyz.html",
      hpChapterProgression = progression 0.666
    },
    targetSource = "urn:uuid:1daa8de6-94e8-4711-b7d1-e43b572aa6e0"
  }
}

valid-bookmark-2.json

validBookmark2 :: Bookmark
validBookmark2 = Bookmark {
  bookmarkId   = Nothing,
  bookmarkBody = BookmarkBody {
    bodyDeviceId = "urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c",
    bodyTime     = "2021-03-12T16:32:49Z",
    bodyOthers   = DM.empty
  },
  bookmarkMotivation = Bookmarking,
  bookmarkTarget = BookmarkTarget {
    targetLocator = L_HrefProgression $ LocatorHrefProgression {
      hpChapterHref        = "/xyz.html",
      hpChapterProgression = progression 0.666
    },
    targetSource = "urn:uuid:1daa8de6-94e8-4711-b7d1-e43b572aa6e0"
  }
}

valid-bookmark-3.json

validBookmark3 :: Bookmark
validBookmark3 = Bookmark {
  bookmarkId   = Just "urn:uuid:715885bc-23d3-4d7d-bd87-f5e7a042c4ba",
  bookmarkBody = BookmarkBody {
    bodyDeviceId = "urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c",
    bodyTime     = "2021-03-12T16:32:49Z",
    bodyOthers   = DM.empty
  },
  bookmarkMotivation = Bookmarking,
  bookmarkTarget = BookmarkTarget {
    targetLocator = L_HrefProgression $ LocatorHrefProgression {
      hpChapterHref        = "/xyz.html",
      hpChapterProgression = progression 0.666
    },
    targetSource = "urn:uuid:1daa8de6-94e8-4711-b7d1-e43b572aa6e0"
  }
}

valid-bookmark-4.json

validBookmark4 :: Bookmark
validBookmark4 = Bookmark {
  bookmarkId   = Just "urn:uuid:715885bc-23d3-4d7d-bd87-f5e7a042c4ba",
  bookmarkBody = BookmarkBody {
  bodyChapter   = "Chapter title",
    bodyDeviceId = "urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c",
    bodyTime     = "2022-06-27T12:47:49Z"
  },
  bookmarkMotivation = Bookmarking,
  bookmarkTarget = BookmarkTarget {
    targetLocator = L_ABT $ LocatorAudioBookTime {
      hpTitle             = "Chapter title",
      hpAudiobookID       = "urn:uuid:b309844e-7d4e-403e-945b-fbc78acd5e03",
      hpChapter             = chapter 32,
      hpDuration         = duration 190000,
      hpTime             = time 78000,
      hpPart             = part 3
    },
    targetSource = "urn:uuid:1daa8de6-94e8-4711-b7d1-e43b572aa6e0"
  }
}

valid-bookmark-5.json

validBookmark5 :: Bookmark
validBookmark5 = Bookmark {
  bookmarkId   = Just "urn:uuid:715885bc-23d3-4d7d-bd87-f5e7a042c4ba",
  bookmarkBody = BookmarkBody {
    bodyDeviceId = "urn:uuid:c83db5b1-9130-4b86-93ea-634b00235c7c",
    bodyTime     = "2022-08-05T16:32:49Z"
  },
  bookmarkMotivation = Bookmarking,
  bookmarkTarget = BookmarkTarget {
    targetLocator = L_P $ LocatorPage {
      hpPage             = page 2
    },
    targetSource = "urn:uuid:1daa8de6-94e8-4711-b7d1-e43b572aa6e0"
  }
}

valid-locator-0.json

validLocator0 :: Locator
validLocator0 = L_HrefProgression $ LocatorHrefProgression {
  hpChapterHref        = "/xyz.html",
  hpChapterProgression = progression 0.666
}

valid-locator-1.json

validLocator1 :: Locator
validLocator1 = L_CFI $ LocatorLegacyCFI {
  lcIdRef              = Just "xyz-html",
  lcContentCFI         = Just "/4/2/2/2",
  lcChapterProgression = Just $ progression 0.25
}

valid-locator-2.json

validLocator2 :: Locator
validLocator2 = L_P $ LocatorPage {
  lcPage         = Just $ page 2
}

valid-locator-3.json

validLocator3 :: Locator
validLocator3 = L_ABT$ LocatorAudioBookTime {
  lcPart               = Just $ part 3,
  lcChapter        = Just $ chapter 32,
  lcTitle         = Just "Chapter title",
  lcAudiobookID       = Just "urn:uuid:b309844e-7d4e-403e-945b-fbc78acd5e03",
  lcDuration       = Just $ duration 190000,
  lcTime         = Just $ time 78000
}