Preserves: an Expressive Data Language

Tony Garnock-Jones tonyg@leastfixedpoint.com
June 2025. Version 0.996.3.

Preserves is a data model, with associated serialization formats.

It supports records with user-defined labels, embedded references, and the usual suite of atomic and compound data types, including binary data as a distinct type from text strings. Its annotations allow separation of data from metadata such as comments, trace information, and provenance information.

Preserves departs from many other data languages in defining how to compare two values. Comparison is based on the data model, not on syntax or on data structures of any particular implementation language.

This document defines the core semantics and data model of Preserves and presents a handful of examples. Two other core documents define

a human-readable text syntax, and
a machine-oriented binary syntax

for the Preserves data model.

Values

Preserves values are given meaning independent of their syntax. We will write “Value” when we mean the set of all Preserves values or an element of that set.

Values fall into two broad categories: atomic and compound data. Every Value is finite and non-cyclic. Embedded values, called Embeddeds, are a third, special-case category.

                      Value = Atom
                            | Compound
                            | Embedded

                       Atom = Boolean
                            | Double
                            | SignedInteger
                            | String
                            | ByteString
                            | Symbol

                   Compound = Record
                            | Sequence
                            | Set
                            | Dictionary

Total order. As we go, we will incrementally specify a total order over Values. Two values of the same kind are compared using kind-specific rules. The ordering among values of different kinds is essentially arbitrary, but having a total order is convenient for many tasks, so we define it as follows:

        (Values)        Atom < Compound < Embedded

        (Compounds)     Record < Sequence < Set < Dictionary

        (Atoms)         Boolean < Double < SignedInteger
                          < String < ByteString < Symbol

Equivalence. Two Values are equal if neither is less than the other according to the total order.

Signed integers.

A SignedInteger is an arbitrarily-large signed integer. SignedIntegers are compared as mathematical integers.

Unicode strings.

A String is a sequence of Unicode scalar values.¹ Strings are compared lexicographically, scalar value by scalar value.²

Binary data.

A ByteString is a sequence of octets. ByteStrings are compared lexicographically.

Symbols.

Programming languages like Lisp and Prolog frequently use string-like values called symbols.³ Here, a Symbol is, like a String, a sequence of Unicode scalar values representing an identifier of some kind. Symbols are also compared lexicographically by scalar value.

Booleans.

There are two Booleans, “false” and “true”. The “false” value is less-than the “true” value.

IEEE floating-point values.

Doubles are double-precision IEEE 754 floating-point values.⁴ Doubles and SignedIntegers are disjoint; by the rules above, every Double is less than every SignedInteger. Two Doubles are to be ordered by the totalOrder predicate defined in section 5.10 of IEEE Std 754-2008.

Records.

A Record is a labelled tuple of Values, the record’s fields. A label can be any Value, but is usually a Symbol.⁵ ⁶ Records are ordered first by label, then lexicographically⁷ by field sequence.

Sequences.

A Sequence is a sequence of Values. Sequences are compared lexicographically.⁷

Sets.

A Set is an unordered finite set of Values. It contains no duplicate values, following the equivalence relation induced by the total order on Values. Two Sets are compared by sorting their elements ascending using the total order and comparing the resulting Sequences.⁷

Dictionaries.

A Dictionary is an unordered finite collection of pairs of Values. Each pair comprises a key and a value. Keys in a Dictionary are pairwise distinct. Instances of Dictionary are compared by lexicographic⁷ comparison of the sequences resulting from ordering each Dictionary’s pairs in ascending order by key.

Embeddeds.

An Embedded allows inclusion of domain-specific, potentially stateful or located data into a Value.⁸ Embeddeds may be used to denote stateful objects, network services, object capabilities, file descriptors, Unix processes, or other possibly-stateful things. Because each Embedded is a domain-specific datum, comparison of two Embeddeds is done according to domain-specific rules.

Motivating Examples. In a Java or Python implementation, an Embedded may denote a reference to a Java or Python object; comparison would be done via the language’s own rules for equivalence and ordering. In a Unix application, an Embedded may denote an open file descriptor or a process ID. In an HTTP-based application, each Embedded might be a URL, compared according to RFC 6943. When a Value is serialized for storage or transfer, Embeddeds will usually be represented as ordinary Values, in which case the ordinary rules for comparing Values will apply.

Appendix. Examples

The definitions above are independent of any particular concrete syntax. The examples of Values that follow are written using the Preserves text syntax, and the example encoded byte sequences use the Preserves binary encoding.

Ordering.

The total ordering specified above means that the following statements are true:

"bzz" < "c" < "caa" < #:"a"
#t < 3.0 < 3 < "3" < '3' < [] < #:#t
[#f] < [foo], because Boolean appears before Symbol in the kind ordering
[x] < [x y], because there is no element remaining to compare against y
[a b] < [x], because a is smaller than x
[x y] < [x z], because y is ordered before z

Simple examples.

Value	Encoded byte sequence
`<capture <discard>>`	B4 B3 07 ‘c’ ‘a’ ‘p’ ‘t’ ‘u’ ‘r’ ‘e’ B4 B3 07 ‘d’ ‘i’ ‘s’ ‘c’ ‘a’ ‘r’ ‘d’ 84 84
`[1 2 3 4]`	B5 B0 01 01 B0 01 02 B0 01 03 B0 01 04 84
`[-2 -1 0 1]`	B5 B0 01 FE B0 01 FF B0 00 B0 01 01 84
`"hello"`	B1 05 ‘h’ ‘e’ ‘l’ ‘l’ ‘o’
`"z水𝄞"`	B1 08 ‘z’ E6 B0 B4 F0 9D 84 9E
`"z水\uD834\uDD1E"`	B1 08 ‘z’ E6 B0 B4 F0 9D 84 9E
`["a" b #"c" [] #{} #t #f]`	B5 B1 01 ‘a’ B3 01 ‘b’ B2 01 ‘c’ B5 84 B6 84 81 80 84
`-257`	B0 02 FE FF
`-1`	B0 01 FF
`0`	B0 00
`1`	B0 01 01
`255`	B0 02 00 FF
`1.0`	87 08 3F F0 00 00 00 00 00 00
`-1.202e300`	87 08 FE 3C B7 B7 59 BF 04 26
`#xd"fff0000000000000"`, negative `Double` infinity	87 08 FF F0 00 00 00 00 00 00

The next example uses a non-Symbol label for a record.⁹ The Record

<[titled person 2 thing 1] 101 "Blackwell" <date 1821 2 3> "Dr">

encodes to

B4                                # Record
  B5                                # Sequence
    B3 06 74 69 74 6C 65 64           # Symbol, "titled"
    B3 06 70 65 72 73 6F 6E           # Symbol, "person"
    B0 01 02                          # SignedInteger, "2"
    B3 05 74 68 69 6E 67              # Symbol, "thing"
    B0 01 01                          # SignedInteger, "1"
  84                                # End (sequence)
  B0 01 65                          # SignedInteger, "101"
  B1 09 42 6C 61 63 6B 77 65 6C 6C  # String, "Blackwell"
  B4                                # Record
    B3 04 64 61 74 65                 # Symbol, "date"
    B0 02 07 1D                       # SignedInteger, "1821"
    B0 01 02                          # SignedInteger, "2"
    B0 01 03                          # SignedInteger, "3"
  84                                # End (record)
  B1 02 44 72                       # String, "Dr"
84                                # End (record)

JSON examples.

Preserves text syntax is a superset of JSON,¹⁰ so the examples from RFC 8259 read as valid Preserves.

The JSON literals true, false and null all read as Symbols, and JSON numbers read (unambiguously) either as SignedIntegers or as Doubles.¹¹

The first RFC 8259 example:

{
  "Image": {
      "Width":  800,
      "Height": 600,
      "Title":  "View from 15th Floor",
      "Thumbnail": {
          "Url":    "http://www.example.com/image/481989943",
          "Height": 125,
          "Width":  100
      },
      "Animated" : false,
      "IDs": [116, 943, 234, 38793]
    }
}

when read using the Preserves text syntax encodes via the binary syntax as follows:

B7
  B1 05 "Image"
  B7
    B1 03 "IDs"      B5
                       B0 01 74
                       B0 02 03 AF
                       B0 02 00 EA
                       B0 03 00 97 89
                     84
    B1 05 "Title"    B1 14 "View from 15th Floor"
    B1 05 "Width"    B0 02 03 20
    B1 06 "Height"   B0 02 02 58
    B1 08 "Animated" B3 05 "false"
    B1 09 "Thumbnail"
      B7
        B1 03 "Url"    B1 26 "http://www.example.com/image/481989943"
        B1 05 "Width"  B0 01 64
        B1 06 "Height" B0 01 7D
      84
  84
84

The second RFC 8259 example:

[
  {
     "precision": "zip",
     "Latitude":  37.7668,
     "Longitude": -122.3959,
     "Address":   "",
     "City":      "SAN FRANCISCO",
     "State":     "CA",
     "Zip":       "94107",
     "Country":   "US"
  },
  {
     "precision": "zip",
     "Latitude":  37.371991,
     "Longitude": -122.026020,
     "Address":   "",
     "City":      "SUNNYVALE",
     "State":     "CA",
     "Zip":       "94085",
     "Country":   "US"
  }
]

encodes to binary as follows:

B5
  B7
    B1 03 "Zip"        B1 05 "94107"
    B1 04 "City"       B1 0D "SAN FRANCISCO"
    B1 05 "State"      B1 02 "CA"
    B1 07 "Address"    B1 00
    B1 07 "Country"    B1 02 "US"
    B1 08 "Latitude"   87 08 40 42 E2 26 80 9D 49 52
    B1 09 "Longitude"  87 08 C0 5E 99 56 6C F4 1F 21
    B1 09 "precision"  B1 03 "zip"
  84
  B7
    B1 03 "Zip"        B1 05 "94085"
    B1 04 "City"       B1 09 "SUNNYVALE"
    B1 05 "State"      B1 02 "CA"
    B1 07 "Address"    B1 00
    B1 07 "Country"    B1 02 "US"
    B1 08 "Latitude"   87 08 40 42 AF 9D 66 AD B4 03
    B1 09 "Longitude"  87 08 C0 5E 81 AA 4F CA 42 AF
    B1 09 "precision"  B1 03 "zip"
  84
84

Appendix. Merging Values

The merge of two Values is a combination of the two values that includes all information from each that is missing from the other. If the values are incompatible, they have no merge.

the merge of two Atoms has no value if they are not equal; otherwise, it has value equal to (an arbitrary) one of the atoms.
the merge of two Embeddeds depends on the interpretation of the embedded values, and so is implementation-defined.
the merge of two Compounds is:
- if both are Sequences, let n be the minimum of the lengths of the two sequences. If every merge of corresponding positions up to n in the sequences is defined, the result is defined, with elements merged up to position n and simply copied from the longer sequence from position n onward; otherwise, it is undefined.
- if both are Records, the Record with the merge of the two input records’ labels as its label and the merge of the inputs’ field sequences as its fields;
- if both are Dictionarys, if every merge of values associated with keys common to both inputs is defined, the result is defined, with merged values at common keys and simply copied from either side for keys unique to that side; otherwise, it is undefined.
Otherwise, the merge is undefined.

Examples.

merge [1, [2], 3] [1, [2, 99], 3, 4, 5] = [1, [2, 99], 3, 4, 5]
merge [1, 2, 3] [1, 5, 3] = ⊥
merge #{a, b, c} #{a, b, c} = ⊥
merge {a: 1, b: [2]} {b: [2, 99] c: 3} = {a: 1, b: [2, 99], c: 3}
merge {a: 1, b: [2]} {a: 5, b: [2]} = ⊥

Notes

All Unicode scalar values are permitted, including NUL (scalar value zero). Because scalar values are defined as code points excluding surrogate code points (D800₁₆–DFFF₁₆), surrogates are not permitted in Preserves Unicode data. ↩
Happily, the design of UTF-8 is such that this gives the same result as a lexicographic byte-by-byte comparison of the UTF-8 encoding of a string! ↩
Even Java has quasi-symbols in the form of its “interned strings”. A Java Preserves implementation might intern Preserves Symbols while leaving Preserves Strings uninterned. ↩
Every value inhabiting a smaller IEEE 754 type (e.g. single- or half-precision) can be injected into and projected from double-precision losslessly and in an order-preserving way. ↩
The Racket programming language defines “prefab” structure types, which map well to our Records. Racket supports record extensibility by encoding record supertypes into record labels as specially-formatted lists. ↩
It is occasionally (but seldom) necessary to interpret such Symbol labels as IRIs. Where a label can be read as a relative IRI, it is notionally interpreted with respect to the IRI urn:uuid:6bf094a6-20f1-4887-ada7-46834a9b5b34; where a label can be read as an absolute IRI, it stands for that IRI; and otherwise, it cannot be read as an IRI at all, and so the label simply stands for itself—for its own Value. ↩
When comparing sequences of values for the total order, lexicographical ordering is used. Elements are drawn pairwise from the two sequences to be compared. If one is smaller than the other according to the total order, the sequence it was drawn from is the smaller of the sequences. If the end of one sequence is reached, while the other sequence has elements remaining, the shorter sequence is considered smaller. Otherwise, all the elements compared equal and neither was longer than the other, so they compare equal. For example,
- [#f] is ordered before [foo] because Boolean appears before Symbol in the kind ordering;
- [x] before [x y] because there is no element remaining to compare against y;
- [a b] before [x] because a is smaller than x; and
- [x y] before [x z] because y is ordered before z according to the ordering rules for Symbol.
↩ ↩² ↩³ ↩⁴
Rationale. Why include Embeddeds as a special class, distinct from, say, a specially-labeled Record? First, a Record can only hold other Values: in order to embed values such as live pointers to Java objects, some means of “escaping” from the Value data type must be provided. Second, Embeddeds are meant to be able to denote stateful entities, for which comparison by address is appropriate; however, we do not wish to place restrictions on the nature of these entities: if we had used Records instead of distinct Embeddeds, users would have to invent an encoding of domain data into Records that reflected domain ordering into Value ordering. This is often difficult and may not always be possible. Finally, because Embeddeds are intended to be able to represent network and memory locations, they must be able to be rewritten at network and process boundaries. Having a distinct class allows generic Embedded rewriting without the quotation-related complications of encoding references as, say, Records. ↩
It happens to line up with Racket’s representation of a record label for an inheritance hierarchy where titled extends person extends thing:
```
(struct date (year month day) #:prefab)
(struct thing (id) #:prefab)
(struct person thing (name date-of-birth) #:prefab)
(struct titled person (title) #:prefab)
```
For more detail on Racket’s representations of record labels, see the Racket documentation for make-prefab-struct. ↩
There is one caveat to be aware of. Section 8.2 of RFC 8259 explicitly permits unpaired surrogate code points in JSON texts without specifying an interpretation for them. Preserves mandates UTF-8 in its binary syntax, forbids unpaired surrogates in its text syntax, and disallows surrogate code points in Strings and Symbols, meaning that any valid JSON text including an unpaired surrogate will not be parseable using the Preserves text syntax rules. ↩

The following schema definitions match exactly the JSON subset of a Preserves input:

version 1 .
JSON = @string string / @integer int / @double double / @boolean JSONBoolean / @null =null
     / @array [JSON ...] / @object { string: JSON ...:... } .
JSONBoolean = =true / =false .

↩