[Docs] [txt|pdf] [Tracker] [Email] [Nits]

Versions: 00 01 02 03 04 05 06 07 08 09 10 11 RFC 3687

INTERNET-DRAFT                                                S. Legg
draft-legg-ldapext-component-matching-00.txt      Adacel Technologies
                                                     October 23, 2000


                 LDAP & X.500 Component Matching Rules

    Copyright (C) The Internet Society (2000). All Rights Reserved.

   Status of this Memo


   This document is an Internet-Draft and is in full conformance with
   all provisions of Section 10 of RFC2026.

   Internet-Drafts are working documents of the Internet Engineering
   Task Force (IETF), its areas, and its working groups.  Note that
   other groups may also distribute working documents as Internet-
   Drafts.

   Internet-Drafts are draft documents valid for a maximum of six months
   and may be updated, replaced, or obsoleted by other documents at any
   time.  It is inappropriate to use Internet-Drafts as reference
   material or to cite them other than as "work in progress".

   The list of current Internet-Drafts can be accessed at
   http://www.ietf.org/ietf/1id-abstracts.txt

   The list of Internet-Draft Shadow Directories can be accessed at
   http://www.ietf.org/shadow.html.

   Distribution of this document is unlimited.  Comments should be sent
   to the LDAPEXT working group mailing list <ietf-ldapext@netscape.com>
   or to the author.

   This Internet-Draft expires on 23 April 2000.


   1. Abstract

   The syntaxes of attributes in an LDAP or X.500 directory range from
   simple data types, such as text string, integer, or boolean, to
   complex structured data types, such as the syntaxes of the directory
   schema operational attributes.  The matching rules defined for the
   complex syntaxes, if any, usually only provide the most immediately
   useful matching capability.  This document defines generic matching
   rules that can match any user selected component parts in an
   attribute value of any arbitrarily complex attribute syntax.  Generic



Legg                     Expires 23 April 2000                  [Page 1]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   string encodings for attribute and assertion values of arbitrary
   syntax are also defined.

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and  "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119 [RFC2119].


   2. Table of Contents

   1. Abstract                                                         1
   2. Table of Contents                                                2
   3. Introduction                                                     3
   4. ComponentAssertion                                               5
   4.1 Component Reference                                             5
   4.1.1 Component Type Substitutions                                  7
   4.1.2 Referencing SET, SEQUENCE and CHOICE Components               8
   4.1.3 Referencing SET OF and SEQUENCE OF Components                 9
   4.1.4 Referencing Components of Parameterized Types                10
   4.1.5 Component Referencing Example                                10
   4.1.6 Referencing Components of Open Types                         11
   4.1.6.1 Open Type Referencing Example                              12
   4.2 Matching of Components                                         12
   4.2.1 Applicability of Existing Matching Rules                     14
   4.2.1.1 String Matching                                            14
   4.2.1.2 Telephone Number Matching                                  16
   4.2.1.3 Distinguished Name Matching                                16
   4.2.2 Additional Useful Matching Rules                             16
   4.2.2.1 The rdnMatch Matching Rule                                 16
   4.2.2.2 The enumeratedMatch Matching Rule                          17
   4.2.2.3 The presentMatch Matching Rule                             18
   4.2.3 Summary of Useful Matching Rules                             19
   5. ComponentFilter                                                 20
   6. The componentFilterMatch Matching Rule                          21
   7. Equality Matching of Complex Components                         23
   7.1 The allComponentsMatch Matching Rule                           24
   7.2 Deriving Component Equality Matching Rules                     26
   7.3 The directoryComponentsMatch Matching Rule                     27
   8. String Encodings for Values of Arbitrary ASN.1 Types            28
   9. Component Matching Examples                                     35
   10. Security Considerations                                        41
   11. Acknowledgements                                               41
   12. References                                                     42
   13. Intellectual Property Notice                                   43
   14. Copyright Notice                                               43
   15. Author's Address                                               44





Legg                     Expires 23 April 2000                  [Page 2]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   3. Introduction

   The structure or data type of data held in an attribute of an LDAP
   [RFC2251] or X.500 [X500] directory is described by the attribute's
   syntax.  Attribute syntaxes range from simple data types, such as
   text string, integer, or boolean, to complex data types, for example,
   the syntaxes of the directory schema operational attributes.

   In X.500, the attribute syntaxes are explicitly described by ASN.1
   [X680] type definitions.  ASN.1 type notation has a number of simple
   data types (e.g. PrintableString, INTEGER, BOOLEAN), and combining
   types (i.e. SET, SEQUENCE, SET OF, SEQUENCE OF, and CHOICE) for
   constructing arbitrarily complex data types from simpler component
   types.  In LDAP, the attributes syntaxes are usually described by
   ABNF [RFC2234] though there is an implied association between the
   LDAP attribute syntaxes and the X.500 ASN.1 types.  To a large
   extent, the data types of attribute values in either an LDAP or X.500
   directory are described by ASN.1 types.  This formal description can
   be exploited to identify component parts of an attribute value for a
   variety of purposes.  This document addresses attribute value
   matching.

   With any complex attribute syntax there is normally a requirement to
   partially match an attribute value of that syntax by matching only
   selected components of the value.  Typically, matching rules specific
   to the attribute syntax are defined to fill this need.  These highly
   specific matching rules usually only provide the most immediately
   useful matching capability.  Some complex attribute syntaxes don't
   even have an equality matching rule let alone any additional matching
   rules for partial matching.  This document defines a generic way of
   matching user selected components in an attribute value of any
   arbitrarily complex attribute syntax, where that syntax is described
   using ASN.1 type notation.

   Section 4 describes the ComponentAssertion, a testable assertion
   about the value of a component of an attribute value of any complex
   syntax.

   Section 5 introduces the ComponentFilter assertion, which is an
   expression of ComponentAssertions.  The ComponentFilter enables more
   powerful filter matching of components in an attribute value.

   Section 6 defines the componentFilterMatch matching rule, which
   enables a ComponentFilter to be evaluated against attribute values.

   Section 7 defines matching rules for component-wise equality matching
   of values of any syntax described by an ASN.1 type definition.




Legg                     Expires 23 April 2000                  [Page 3]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   The LDAP string encodings for attribute syntaxes defined in ABNF do
   not clearly or consistently delineate the component parts of an
   attribute value.  A regular and uniform string encoding for arbitrary
   component data types is needed for the assertion syntax of the
   componentFilterMatch matching rule.  Section 8 defines a human
   readable text encoding (based on ASN.1 value notation) for an ASN.1
   value of any ASN.1 type.  Though primarily intended for assertion
   syntaxes in component matching rules, these generic string encoding
   rules could also be used for new attribute syntaxes, or in other
   domains where human readable renderings of ASN.1 values would be
   useful.  Note that "ASN.1 value" does not mean a BER encoded value.
   The ASN.1 value is an abstract concept that is independent of any
   particular encoding.  BER is just one possible encoding of an ASN.1
   value.  The component matching rules operate at the abstract level
   without regard for the possible encodings of a value.

   Examples showing the usage of componentFilterMatch are in Section 9.

   For a new attribute syntax, the specifications in sections 4 to 8 of
   this document make it possible to fully and precisely define, the
   LDAP string encoding, the LDAP and X.500 binary encoding (and
   possibly other encodings in the future, e.g. XML via XER), a suitable
   equality matching rule, and a comprehensive collection of component
   matching capabilities, by simply writing down an ASN.1 type
   definition for the syntax.  These implicit definitions are also
   automatically extended if the ASN.1 type is later extended.  The
   algorithmic relationship between the ASN.1 type definition, the
   various encodings and the component matching behaviour makes
   directory server implementation support for the component matching
   rules amenable to automatic code generation from ASN.1 type
   definitions.

   Schema designers have the choice of storing related items of data as
   a single attribute value of a complex syntax in some entry, or as a
   subordinate entry where the related data items are stored as separate
   attribute values of simpler syntaxes.  The inability to search
   component parts of a complex syntax has been used as an argument for
   favouring the subordinate entries approach.  The component matching
   rules provide the analogous matching capability on an attribute value
   of a complex syntax that a search filter has on a subordinate entry.

   The corresponding ASN.1 type definitions for LDAP syntaxes are
   usually not reproduced or referenced alongside the formal definition
   of the LDAP syntax.  As an aid to those wishing to use the component
   matching rules, the corresponding ASN.1 type definitions for the
   syntaxes from [RFC2252] are provided in a companion document
   [SYNTAX].  Syntaxes defined with only a string encoding (i.e. without
   an explicit or implied corresponding ASN.1 type definition) cannot



Legg                     Expires 23 April 2000                  [Page 4]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   use the component matching capabilities described in this document
   unless and until a semantically equivalent ASN.1 type definition is
   defined for them.

   In the remainder of this document "type" will be taken to mean an
   ASN.1 type unless explicitly qualified as an attribute type, and
   "value" will be taken to mean an ASN.1 value unless explicitly
   qualified as an attribute value.


   4. ComponentAssertion

   A ComponentAssertion is an assertion about the presence, or values
   of, components within an ASN.1 value, i.e. an instance of an ASN.1
   type.  The ASN.1 value is typically an attribute value, where the
   ASN.1 type is the syntax of the attribute.  However a
   ComponentAssertion may also be applied to a component part of an
   attribute value.  The assertion evaluates to either TRUE, FALSE or
   undefined for each tested ASN.1 value.

   A ComponentAssertion is described by the following ASN.1 type
   (assumed to be defined with "EXPLICIT TAGS" in force):

   ComponentAssertion ::= SEQUENCE {
       component         [0] ComponentReference,
       useDefaultValues  [1] BOOLEAN DEFAULT TRUE,
       rule              [2] MATCHING-RULE.&id,
       value             [3] MATCHING-RULE.&AssertionType }

   ComponentReference ::= UTF8String

   MATCHING-RULE.&id equates to the OBJECT IDENTIFIER of a matching
   rule.  MATCHING-RULE.&AssertionType is an open type (formally known
   as the ANY type).

   The "component" field of a ComponentAssertion identifies which
   component part of a value of some ASN.1 type is to be tested, the
   "useDefaultValues" field indicates whether DEFAULT values are to be
   substituted for absent component values, the "rule" field indicates
   how that component is to be tested, and the "value" field is an
   asserted ASN.1 value against which the component is tested.  The
   ASN.1 type of the asserted value is determined by the chosen rule.

   The fields of a ComponentAssertion are described in detail in the
   following sections.


   4.1 Component Reference



Legg                     Expires 23 April 2000                  [Page 5]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   The component field in a ComponentAssertion is a UTF8 character
   string [RFC2279] whose textual content is a component reference,
   identifying a component part of some ASN.1 type or value.  A
   component reference conforms to the following ABNF [RFC2234], which
   extends the notation defined in Clause 14 of [X680]:

   component-reference = ComponentId *( "." ComponentId )
   ComponentId         = identifier /
                         from-beginning /
                         count /
                         from-end /       ; extends Clause 14
                         all

   identifier          = lowercase *alphanumeric *(hyphen 1*alphanumeric)
   alphanumeric        = uppercase / lowercase / decimal-digit
   uppercase           = %x41-5A  ; "A" to "Z"
   lowercase           = %x61-7A  ; "a" to "z"
   hyphen              = "-"

   from-beginning      = positive-number
   count               = "0"
   from-end            = "-" positive-number
   all                 = "*"

   positive-number     = non-zero-digit *decimal-digit

   decimal-digit       = %x30-39  ; "0" to "9"
   non-zero-digit      = %x31-39  ; "1" to "9"

   An <identifier> conforms to the definition of an identifier in ASN.1
   notation (Clause 11.3 of [X680]).  It begins with a lowercase letter
   and is followed by zero or more letters, digits, and hyphens.  A
   hyphen is not permitted to be the last character and a hyphen is not
   permitted to be followed by another hyphen.

   A component reference is a sequence of one or more ComponentIds where
   each successive ComponentId identifies an inner component at the next
   level of nesting of the ASN.1 combining types, i.e. SET, SEQUENCE,
   SET OF, SEQUENCE OF, and CHOICE.

   A component reference is always considered in the context of a
   particular complex ASN.1 type.  When applied to the ASN.1 type the
   component reference identifies a specific component type.  When
   applied to a value of the ASN.1 type a component reference identifies
   zero, one or more component values of that component type.  The
   component values are potentially in a DEFAULT value if
   useDefaultValues is TRUE.  The specific component type identified by
   the component reference determines what matching rules are capable of



Legg                     Expires 23 April 2000                  [Page 6]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   being used to match the component values.

   An empty string for a component reference, which would identify the
   whole ASN.1 value, is NOT supported since assertions about a whole
   value are already possible by the direct application of a matching
   rule to an attribute value.

   A valid component reference for a particular complex ASN.1 type is
   constructed by starting with the outermost combining type and
   repeatedly selecting one of the permissible forms of ComponentId to
   identify successively deeper nested components.  A component
   reference MAY identify a component with a complex ASN.1 type, i.e. it
   is NOT required that the component type identified by a component
   reference be a simple ASN.1 type.


   4.1.1 Component Type Substitutions

   ASN.1 type notation has a number of constructs for referencing other
   defined types, and constructs that are irrelevant for matching
   purposes. These constructs are not represented in a component
   reference in any way and substitutions of the component type are
   performed to eliminate them from further consideration.  These
   substitutions automatically occur prior to each ComponentId, whether
   constructing or interpreting a component reference, but do not occur
   after the last ComponentId, except as allowed by Section 4.2.

   If the ASN.1 type is an ASN.1 type reference then the component type
   is taken to be the actual definition on the right hand side of the
   type assignment for the referenced type.

   If the ASN.1 type is a tagged type then the component type is taken
   to be the type without the tag.

   If the ASN.1 type is a constrained type then the component type is
   taken to be the type without the constraint.

   If the ASN.1 type is an ObjectClassFieldType (Clause 14 of [X681])
   that denotes a specific ASN.1 type (e.g. MATCHING-RULE.&id denotes
   the OBJECT IDENTIFIER type) then the component type is taken to be
   the denoted type.  Section 4.1.6 describes the case where the
   ObjectClassFieldType denotes an open type.

   If the ASN.1 type is a selection type other than one used in the list
   of components for a SET or SEQUENCE type then the component type is
   taken to be the selected alternative type from the named CHOICE.

   If the ASN.1 type is a TypeFromObject (Clause 15 of [X681]) then the



Legg                     Expires 23 April 2000                  [Page 7]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   component type is taken to be the denoted type.

   If the ASN.1 type is a ValueSetFromObjects (Clause 15 of [X681]) then
   the component type is taken to be the governing type of the denoted
   values.


   4.1.2 Referencing SET, SEQUENCE and CHOICE Components

   If the ASN.1 type is a SET or SEQUENCE type then the <identifier>
   form of ComponentId MAY be used to identify the component type within
   that SET or SEQUENCE having that identifier.  If <identifier>
   references an OPTIONAL component type and that component is not
   present in a particular value then there are no corresponding
   component values.  If <identifier> references a DEFAULT component
   type and useDefaultValues is TRUE (the default setting for
   useDefaultValues) and that component is not present in a particular
   value then the component value is taken to be the default value.  If
   <identifier> references a DEFAULT component type and useDefaultValues
   is FALSE and that component is not present in a particular value then
   there are no corresponding component values.

   If the ASN.1 type is a CHOICE type then the <identifier> form of
   ComponentId MAY be used to identify the alternative type within that
   CHOICE having that identifier.  If <identifier> references an
   alternative other than the one used in a particular value then there
   are no corresponding component values.

   The COMPONENTS OF notation in Clause 24 of [X680] augments the
   defined list of components in a SET or SEQUENCE type by including all
   the components of another defined SET or SEQUENCE type respectively.
   These included components are referenced directly by identifier as
   though they were defined in-line in the type using the COMPONENTS OF
   notation.

   The SelectionType (Clause 29 of [X680]), when used in the list of
   components for a SET or SEQUENCE type, includes a single component
   from a defined CHOICE type.  This included component is referenced
   directly by identifier as though it was defined in-line in the SET or
   SEQUENCE type.

   The REAL type is treated as though it is the SEQUENCE type defined in
   Clause 20.5 of [X680].

   The EMBEDDED PDV type is treated as though it is the SEQUENCE type
   defined in Clause 32.5 of [X680].

   The EXTERNAL type is treated as though it is the SEQUENCE type



Legg                     Expires 23 April 2000                  [Page 8]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   defined in Clause 33.5 of [X680].

   The unrestricted CHARACTER STRING type is treated as though it is the
   SEQUENCE type defined in Clause 39.5 of [X680].

   The INSTANCE OF type is treated as though it is the SEQUENCE type
   defined in Annex C of [X681].

   The <identifier> form MUST NOT be used on any other ASN.1 type.


   4.1.3 Referencing SET OF and SEQUENCE OF Components

   If the ASN.1 type is a SET OF or SEQUENCE OF type then the <from-
   beginning>, <from-end>, <count> and <all> forms of ComponentId can be
   used.

   The <from-beginning> form of ComponentId MAY be used to identify one
   instance (i.e. value) of the component type of the SET OF or SEQUENCE
   OF type (e.g. if Foo ::= SET OF Bar, then Bar is the component type),
   where the instances are numbered from one upwards.  If <from-
   beginning> references a higher numbered instance than the last
   instance in a particular value of the SET OF or SEQUENCE OF type then
   there is no corresponding component value.

   The <from-end> form of ComponentId MAY be used to identify one
   instance of the component type of the SET OF or SEQUENCE OF type,
   where "-1" is the last instance, "-2" is the second last instance,
   and so on.  If <from-end> references a lower numbered instance than
   the first instance in a particular value of the SET OF or SEQUENCE OF
   type then there is no corresponding component value.

   The <count> form of ComponentId identifies a notional count of the
   number of instances of the component type in a value of the SET OF or
   SEQUENCE OF type.  This count is not explicitly represented but for
   matching purposes it has an assumed ASN.1 type of INTEGER.  A
   ComponentId of the <count> form MUST be the last ComponentId in a
   component reference.

   The <all> form of ComponentId MAY be used to simultaneously identify
   all instances of the component type of the SET OF or SEQUENCE OF
   type.  It is through the <all> form that a component reference can
   identify more than one component value.  However, if a particular
   value of the SET OF or SEQUENCE OF type is an empty list there are no
   corresponding component values.

   Where multiple component values are identified, the remaining
   ComponentIds in the component reference, if any, can identify zero,



Legg                     Expires 23 April 2000                  [Page 9]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   one or more subcomponent values for each of the higher level
   component values.

   The corresponding ASN.1 type for the <from-beginning>, <from-end>,
   and <all> forms of ComponentId is the component type of the SET OF or
   SEQUENCE OF type.

   The <from-beginning>, <count>, <from-end> and <all> forms MUST NOT be
   used on ASN.1 types other than SET OF or SEQUENCE OF.


   4.1.4 Referencing Components of Parameterized Types

   A component reference cannot be formed for a parameterized type
   unless the type has been used with actual parameters, in which case
   the type is treated as though the DummyReferences [X683] have been
   substituted with the actual parameters.


   4.1.5 Component Referencing Example

   Consider the following ASN.1 type definitions.

   ExampleType ::= SEQUENCE {
       part1       [0] INTEGER,
       part2       [1] ExampleSet,
       part3       [2] SET OF OBJECT IDENTIFIER,
       part4       [3] ExampleChoice }

   ExampleSet ::= SET {
       option      PrintableString,
       setting     BOOLEAN }

   ExampleChoice ::= CHOICE {
       eeny-meeny  BIT STRING,
       miney-mo    OCTET STRING }

   Following are component references constructed with respect to the
   type ExampleType.

   The component reference "part1" identifies a component of a value of
   ExampleType having the ASN.1 tagged type [0] INTEGER.

   The component reference "part2" identifies a component of a value of
   ExampleType having the ASN.1 type of [1] ExampleSet

   The component reference "part2.option" identifies a component of a
   value of ExampleType having the ASN.1 type of PrintableString.  A



Legg                     Expires 23 April 2000                 [Page 10]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   ComponentAssertion could also be applied to a value of ASN.1 type
   ExampleSet, in which case the component reference "option" would
   identify the same kind of information.

   The component reference "part3" identifies a component of a value of
   ExampleType having the ASN.1 type of [2] SET OF OBJECT IDENTIFIER.

   The component reference "part3.2" identifies the second instance of
   the part3 SET OF.  The instance has the ASN.1 type of OBJECT
   IDENTIFIER.

   The component reference "part3.0" identifies the count of the number
   of instances in the part3 SET OF.  The count has the corresponding
   ASN.1 type of INTEGER.

   The component reference "part3.*" identifies all the instances in the
   part3 SET OF.  Each instance has the ASN.1 type of OBJECT IDENTIFIER.

   The component reference "part4" identifies a component of a value of
   ExampleType having the ASN.1 type of [3] ExampleChoice.

   The component reference "part4.miney-mo" identifies a component of a
   value of ExampleType having the ASN.1 type of OCTET STRING.


   4.1.6 Referencing Components of Open Types

   If a sequence of ComponentIds identifies an ObjectClassFieldType that
   denotes an open type (e.g. MATCHING-RULE.&AssertionType denotes an
   open type) then the ASN.1 type of the component varies.  An open type
   is typically constrained by some other component in an outer
   enclosing type (e.g. in a ComponentAssertion, MATCHING-
   RULE.&AssertionType is constrained by the chosen matching rule) so
   the actual ASN.1 type of a value of the open type will generally be
   known.  The constraint will also limit the range of permissible
   types.

   When constructing a component reference, any one of the permitted
   types MAY be chosen as the implied component type and subcomponents
   of that type can be identified with further ComponentIds.  However
   note that the remainder of the component reference is potentially
   ambiguous in that it may also identify subcomponents of some other
   permitted type.  If a different type is used in a particular value of
   the open type then the component reference may still be compatible
   and yield component values.  If the component reference is not
   compatible there will be no corresponding component values.

   Users of ComponentAssertion SHOULD NOT rely on the component



Legg                     Expires 23 April 2000                 [Page 11]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   reference being compatible with only one of the permitted types in an
   open type.  Instead the ComponentAssertion should be combined with
   another ComponentAssertion, on the component that constrains the open
   type, to verify that the open type is of the expected type.  The
   method for combining ComponentAssertions is described in Section 5.


   4.1.6.1 Open Type Referencing Example

   The ASN.1 type AttributeTypeAndValue from [X501] describes a single
   attribute value of a nominated attribute type.

   AttributeTypeAndValue ::= SEQUENCE {
       type    ATTRIBUTE.&id ({SupportedAttributes}),
       value   ATTRIBUTE.&Type ({SupportedAttributes}{@type}) }

   ATTRIBUTE.&id denotes an OBJECT IDENTIFIER and
   ({SupportedAttributes}) constrains the OBJECT IDENTIFIER to be a
   supported attribute type (see [X682] for the semantics of ASN.1
   constraint notation).

   ATTRIBUTE.&Type denotes an open type and
   ({SupportedAttributes}{@type}) constrains it to be of the attribute
   syntax for the attribute type.

   The component reference "value" on AttributeTypeAndValue refers to an
   open type, in this case an attribute value.

   One of the X.500 standard attributes is facsimileTelephoneNumber,
   which is defined [X520] to have the following syntax.

   FacsimileTelephoneNumber ::= SEQUENCE {
       telephoneNumber  PrintableString (SIZE(1.. ub-telephone-number)),
       parameters       G3FacsimileNonBasicParameters OPTIONAL }

   If the attribute type is known to be facsimileTelephoneNumber, or any
   other attribute with the same attribute syntax, the component
   reference "value.telephoneNumber" on AttributeTypeAndValue is
   compatible, but it would also be compatible with any other attribute
   whose syntax is a SET, SEQUENCE or CHOICE with a component identified
   as "telephoneNumber".

   In practice there should be another ComponentAssertion on
   AttributeTypeAndValue with a component reference of "type" that tests
   whether the attribute type is indeed facsimileTelephoneNumber.


   4.2 Matching of Components



Legg                     Expires 23 April 2000                 [Page 12]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   The rule in a ComponentAssertion specifies how the zero, one or more
   component values identified by the component reference are tested by
   the assertion.  Attribute matching rules are used to specify the
   semantics of the test.

   Each matching rule has a notional set of attribute syntaxes
   (typically one), defined as ASN.1 types, to which it may be applied.
   When used in a ComponentAssertion these matching rules apply to the
   same ASN.1 types, only in this context the corresponding ASN.1 values
   are not complete attribute values.

   Note that the referenced component type may be a tagged and/or
   constrained version of the expected attribute syntax (e.g. [0]
   INTEGER, whereas integerMatch would expect simply INTEGER), or an
   open type.  Additional type substitutions of the kind described in
   Section 4.1.1 are performed as required to reduce the component type
   to the same type as the expected attribute syntax.  If an open type
   is encountered the actual ASN.1 type of the component value is
   substituted before continuing.

   If a matching rule applies to more than one attribute syntax (e.g.
   objectIdentifierFirstComponentMatch [X520]) then the minimum number
   of substitutions required to conform to any one of those syntaxes are
   performed.  If a matching rule can apply to any attribute syntax
   (e.g. the allComponentsMatch rule defined in Section 7.1) then the
   referenced component type is used as is, with no additional
   substitutions.

   The value in a ComponentAssertion will be of the assertion syntax
   (i.e. ASN.1 type) required by the chosen matching rule.  Note that
   the assertion syntax of a matching rule is not necessarily the same
   as the attribute syntax(es) to which the rule may be applied.

   Some matching rules do not have a fixed assertion syntax (e.g.
   allComponentsMatch).  For these rules the ASN.1 type of the
   referenced component, after any additional substitutions, is used in
   place of an attribute syntax to decide the required assertion syntax.

   The ComponentAssertion is undefined if:

   i) the matching rule in the ComponentAssertion is not known to the
   evaluating procedure,

   ii) if no part of the component reference identifies an open type and
   the matching rule is not applicable to the referenced component type,
   even with the additional substitutions,

   iii) the value in the ComponentAssertion does not conform to the



Legg                     Expires 23 April 2000                 [Page 13]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   assertion syntax defined for the matching rule,

   iv) an open type in the tested value cannot be decoded, or

   v) the implementation does not support the particular combination of
   component reference and matching rule.

   If the ComponentAssertion is not undefined then the
   ComponentAssertion evaluates to TRUE if there is at least one
   component value for which the matching rule applied to that component
   value returns TRUE, and evaluates to FALSE otherwise (which includes
   the case where there are no component values).


   4.2.1 Applicability of Existing Matching Rules

   4.2.1.1 String Matching

   ASN.1 has a number of built in restricted character string types with
   different character sets and/or different character encodings.  A
   directory user generally has little interest in the particular
   character set or encoding used to represent a character string
   component value, and some directory server implementations make no
   distinction between the different string types in their internal
   representation of values.  So rather than define string matching
   rules for each of the restricted character string types, the existing
   case ignore and case exact string matching rules are extended to
   apply to component values of any of the restricted character string
   types, in addition to component values of the DirectoryString type.
   This extension is only for the purposes of component matching
   described in this document.

   The relevant string matching rules are: caseIgnoreMatch,
   caseIgnoreOrderingMatch, caseIgnoreSubstringsMatch, caseExactMatch,
   caseExactOrderingMatch and caseExactSubstringsMatch.  The relevant
   restricted character string types are: NumericString,
   PrintableString, VisibleString, IA5String, UTF8String, BMPString,
   UniversalString, TeletexString, VideotexString, GraphicString and
   GeneralString.

   The assertion syntax of the string matching rules is still
   DirectoryString regardless of the string syntax of the component
   being matched.  Thus an implementation will be called upon to compare
   a DirectoryString value to a value of one of the restricted character
   string types.  As is the case when comparing two DirectoryStrings
   where the chosen alternatives are of different string types, the
   comparison proceeds so long as the corresponding characters are
   representable in both character sets.  Otherwise matching returns



Legg                     Expires 23 April 2000                 [Page 14]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   FALSE.

   It is not uncommon for ASN.1 specifications to define types that are
   a CHOICE between two or more alternative ASN.1 string types, where
   the particular alternative chosen carries no semantic significance
   (DirectoryString being a prime example).  Such types are defined to
   avoid having to use a complicated character encoding for all values
   when most values could use a simpler string type, or to deal with
   evolving requirements that compel the use of a broader character set
   while still maintaining backward compatibility.  It is convenient to
   also be able to use the existing character string matching rules on
   types that are a purely syntactic choice of string types.

   While there are certain ASN.1 constructs that betray the semantic
   significance of the alternatives within a CHOICE type, the absence of
   those constructs does not necessarily mean a CHOICE type is purely
   syntactic.  Therefore, it is necessary for specifications to declare
   the purely syntactic CHOICE types so that they may be matched with
   the existing string matching rules.  These CHOICE types will be
   referred to as ChoiceOfStrings types.  The string matching rules
   above are extended to also apply to any ChoiceOfStrings type.

   To be eligible to be declared a ChoiceOfStrings type an ASN.1 type
   MUST satisfy the following conditions.

   i) The type is a CHOICE type.

   ii) The component type of each alternative is one of the following
   ASN.1 restricted string types:  NumericString, PrintableString,
   TeletexString (T61String), VideotexString, IA5String, GraphicString,
   VisibleString (ISO646String), GeneralString, BMPString,
   UniversalString or UTF8String.

   iii) All the alternatives are of different restricted string types,
   i.e. no two alternatives have the same ASN.1 restricted string type.

   iv) Either none of the alternatives has a constraint, or all of the
   alternatives have exactly the same constraint.

   Tagging on the alternative types is ignored.

   Consider the ASN.1 parameterized type definition of DirectoryString.

   DirectoryString { INTEGER : maxSize } ::= CHOICE {
       teletexString     TeletexString (SIZE (1..maxSize)),
       printableString   PrintableString (SIZE (1..maxSize)),
       bmpString         BMPString (SIZE (1..maxSize)),
       universalString   UniversalString (SIZE (1..maxSize)),



Legg                     Expires 23 April 2000                 [Page 15]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


       uTF8String        UTF8String (SIZE (1..maxSize)) }

   Any use of the DirectoryString parameterized type with an actual
   parameter defines a ASN.1 type that satisfies the above conditions.
   Recognising that the alternative within a DirectoryString carries no
   semantic significance, this document declares (each and every use of)
   DirectoryString{} to be a ChoiceOfStrings type.

   Other specifications MAY declare other types satisfying the above
   conditions to be ChoiceOfStrings types.  The declaration SHOULD be
   made at the point where the ASN.1 type is defined, otherwise it
   SHOULD be made at the point where it is introduced as, or in, an
   attribute syntax.


   4.2.1.2 Telephone Number Matching

   Early editions of X.520 [X520] gave the syntax of the telephoneNumber
   attribute as a constrained PrintableString.  The fourth edition of
   X.520 equates the ASN.1 type name TelephoneNumber to the constrained
   PrintableString and uses TelephoneNumber as the attribute and
   assertion syntax.  For the purposes of component matching,
   telephoneNumberMatch and telephoneNumberSubstringsMatch are permitted
   to be applied to any PrintableString value, as well as to
   TelephoneNumber values.


   4.2.1.3 Distinguished Name Matching

   The DistinguishedName type is defined by assignment to be the same as
   the RDNSequence type, however RDNSequence is sometimes directly used
   in other type definitions.  For the purposes of component matching,
   distinguishedNameMatch is also permitted to be applied to values of
   the RDNSequence type.


   4.2.2 Additional Useful Matching Rules

   This section defines additional matching rules that may prove useful
   in ComponentAssertions.  These rules MAY also be used in
   extensibleMatch search filters.


   4.2.2.1 The rdnMatch Matching Rule

   The distinguishedNameMatch matching rule can match whole
   distinguished names but it is sometimes useful to be able to match
   specific RDNs in a DN without regard for the other RDNs in the DN.



Legg                     Expires 23 April 2000                 [Page 16]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   The rdnMatch matching rule allows component RDNs of a DN to be
   tested.

   The LDAP style definitions for rdnMatch and its assertion syntax are:

   ( <OID-rdnMatch> NAME 'rdnMatch' SYNTAX <OID-RDN> )

   ( <OID-RDN> DESC 'RDN' )

   [ Editor's Note: <OID-rdnMatch> and <OID-RDN> to be supplied. ]

   The LDAP string encoding for a value of the RDN syntax is given by
   <RelativeDistinguishedNameValue> in Section 8.

   The X.500 style definition for rdnMatch is:

   rdnMatch MATCHING-RULE ::= {
       SYNTAX  RelativeDistinguishedName
       ID      <OID-rdnMatch> }

   The rdnMatch rule evaluates to true if the component value and
   assertion value are the same RDN, using the same RDN comparison
   method as distinguishedNameMatch.

   When using rdnMatch to match components of DNs it is important to
   note that the LDAP string encoding of a DN [RFC2253] reverses the
   order of the RDNs.  So for the DN represented in LDAP as "cn=Steven
   Legg, o=Adacel, c=au", the RDN "cn=Steven Legg" corresponds to the
   component reference "3", or alternatively, "-1".


   4.2.2.2 The enumeratedMatch Matching Rule

   There is no existing matching rule that could be used to match a
   value of an arbitrary ENUMERATED type.  The enumeratedMatch matching
   rule is defined to fill this role.  It may be applied to values of
   any ENUMERATED type, or any INTEGER type (typically those with a
   named number list).

   The LDAP style definitions for enumeratedMatch and its assertion
   syntax are:

   ( <OID-enumeratedMatch> NAME 'enumeratedMatch'
       SYNTAX <OID-OpenType> )

   ( <OID-OpenType> DESC 'OpenType' )

   [ Editor's Note: I need a way to specify a matching rule with a



Legg                     Expires 23 April 2000                 [Page 17]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   variable assertion syntax.  In X.500 this is indicated by omitting
   the optional SYNTAX field in the MATCHING-RULE information object
   (the assertion syntax then defaults to the target attribute syntax in
   actual usage - unless the description of the matching rule says
   otherwise).  The SYNTAX in the LDAP string encoding of a
   MatchingRuleDescription is mandatory.  The problem would be solved if
   the SYNTAX field were made optional, however failing that, the
   OpenType syntax, which can be any other syntax in actual usage, is
   defined. ]

   [ Editor's Note: <OID-enumeratedMatch> and <OID-OpenType> to be
   supplied. ]

   In general, the LDAP string encoding for the OpenType syntax is
   described by <Value> in Section 8.  Since enumeratedMatch only
   applies to values of an ENUMERATED or INTEGER type, assertion values
   for enumeratedMatch are constrained to the string encoding given by
   <EnumeratedValue> or <IntegerValue> respectively, in Section 8.

   The X.500 style definition for enumeratedMatch is:

   enumeratedMatch MATCHING-RULE ::= {
       ID      <OID-enumeratedMatch> }

   The enumeratedMatch rule evaluates to true if the component value and
   assertion value are the same.

   The enumeratedMatch rule MAY be used as the equality matching rule
   for an attribute.


   4.2.2.3 The presentMatch Matching Rule

   At times it would be useful to test not if a specific value of a
   particular component is present, but whether any value of a
   particular component is present.  The presentMatch matching rule
   allows the presence of a particular component value to be tested.

   The LDAP style definitions for presentMatch and its assertion syntax
   are:

   ( <OID-presentMatch> NAME 'presentMatch' SYNTAX <OID-NULL> )

   ( <OID-NULL> DESC 'NULL' )

   [ Editor's Note: <OID-presentMatch> and <OID-NULL> to be supplied. ]

   The LDAP string encoding for a value of the NULL syntax is given by



Legg                     Expires 23 April 2000                 [Page 18]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   <NullValue> in Section 8.

   The X.500 style definition for presentMatch is:

   presentMatch MATCHING-RULE ::= {
       SYNTAX  NULL
       ID      <OID-presentMatch> }

   When used in a extensible match filter item, presentMatch behaves
   like the "present" case of a regular search filter.  In a
   ComponentAssertion presentMatch evaluates to TRUE if and only if the
   component reference identifies one or more component values,
   regardless of the actual component value contents.  Note that if
   useDefaultValues is TRUE then the identified component values may be
   (part of) a DEFAULT value.

   The notional count referenced by the <count> form of ComponentId is
   taken to be present if the SET OF value is present, and absent
   otherwise.  Note that in ASN.1 notation an absent SET OF value is
   distinctly different from a SET OF value that is present but empty.
   It is up to the specification using the ASN.1 notation to decide
   whether the distinction matters.  Often an empty SET OF component and
   an absent SET OF component are treated as semantically equivalent.
   If a SET OF value is present, but empty, a presentMatch on the SET OF
   component will return TRUE and the notional count will be present and
   equal to zero.


   4.2.3 Summary of Useful Matching Rules

   The following is a non-exhaustive list of useful matching rules and
   the ASN.1 types to which they can be applied, taking account of all
   the extensions described within Section 4.2.1, and the new matching
   rules defined in Section 4.2.2.

   +================================+==============================+
   | Matching Rule                  | ASN.1 Type                   |
   +================================+==============================+
   | bitStringMatch                 | BIT STRING                   |
   +--------------------------------+------------------------------+
   | booleanMatch                   | BOOLEAN                      |
   +--------------------------------+------------------------------+
   | caseIgnoreMatch                | NumericString                |
   | caseIgnoreOrderingMatch        | PrintableString              |
   | caseIgnoreSubstringsMatch      | VisibleString (ISO646String) |
   | caseExactMatch                 | IA5String                    |
   | caseExactOrderingMatch         | UTF8String                   |
   | caseExactSubstringsMatch       | BMPString (UCS-2, UNICODE)   |



Legg                     Expires 23 April 2000                 [Page 19]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   |                                | UniversalString (UCS-4)      |
   |                                | TeletexString (T61String)    |
   |                                | VideotexString               |
   |                                | GraphicString                |
   |                                | GeneralString                |
   |                                | any ChoiceOfStrings type     |
   +--------------------------------+------------------------------+
   | caseIgnoreIA5Match             | IA5String                    |
   | caseExactIA5Match              |                              |
   +--------------------------------+------------------------------+
   | distinguishedNameMatch         | DistinguishedName            |
   |                                | RDNSequence                  |
   +--------------------------------+------------------------------+
   | enumeratedMatch                | ENUMERATED                   |
   |                                | INTEGER                      |
   +--------------------------------+------------------------------+
   | generalizedTimeMatch           | GeneralizedTime              |
   | generalizedTimeOrderingMatch   |                              |
   +--------------------------------+------------------------------+
   | integerMatch                   | INTEGER                      |
   | integerOrderingMatch           |                              |
   +--------------------------------+------------------------------+
   | numericStringMatch             | NumericString                |
   | numericStringOrderingMatch     |                              |
   | numericStringSubstringsMatch   |                              |
   +--------------------------------+------------------------------+
   | objectIdentifierMatch          | OBJECT IDENTIFIER            |
   +--------------------------------+------------------------------+
   | octetStringMatch               | OCTET STRING                 |
   | octetStringOrderingMatch       |                              |
   | octetStringSubstringsMatch     |                              |
   +--------------------------------+------------------------------+
   | presentMatch                   | any ASN.1 type               |
   +--------------------------------+------------------------------+
   | rdnMatch                       | RelativeDistinguishedName    |
   +--------------------------------+------------------------------+
   | telephoneNumberMatch           | PrintableString              |
   | telephoneNumberSubstringsMatch | TelephoneNumber              |
   +--------------------------------+------------------------------+
   | uTCTimeMatch                   | UTCTime                      |
   | uTCTimeOrderingMatch           |                              |
   +--------------------------------+------------------------------+


   5. ComponentFilter

   The ComponentAssertion allows the value(s) of any one component type
   in a complex ASN.1 type to be matched, but there is often a desire to



Legg                     Expires 23 April 2000                 [Page 20]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   match the values of more than one component type.  A ComponentFilter
   is an assertion about the presence, or values of, multiple components
   within an ASN.1 value.

   The ComponentFilter assertion, an expression of ComponentAssertions,
   evaluates to either TRUE, FALSE or undefined for each tested ASN.1
   value.

   A ComponentFilter is described by the following ASN.1 type (assumed
   to be defined with "EXPLICIT TAGS" in force):

   ComponentFilter ::= CHOICE {
       item  [0] ComponentAssertion,
       and   [1] SEQUENCE OF ComponentFilter,
       or    [2] SEQUENCE OF ComponentFilter,
       not   [3] ComponentFilter }

   Note: despite the use of SEQUENCE OF instead of SET OF for the "and"
   and "or" alternatives in ComponentFilter, the order of the component
   filters is not significant.

   A ComponentFilter that is a ComponentAssertion evaluates to TRUE if
   the ComponentAssertion is TRUE, evaluates to FALSE if the
   ComponentAssertion is FALSE, and evaluates to undefined otherwise.

   The "and" of a sequence of component filters evaluates to TRUE if the
   sequence is empty or if each component filter evaluates to TRUE,
   evaluates to FALSE if at least one component filter is FALSE, and
   evaluates to undefined otherwise.

   The "or" of a sequence of component filters evaluates to FALSE if the
   sequence is empty or if each component filter evaluates to FALSE,
   evaluates to TRUE if at least one component filter is TRUE, and
   evaluates to undefined otherwise.

   The "not" of a component filter evaluates to TRUE if the component
   filter is FALSE, evaluates to FALSE if the component filter is TRUE,
   and evaluates to undefined otherwise.


   6. The componentFilterMatch Matching Rule

   The componentFilterMatch matching rule allows a ComponentFilter to be
   applied to an attribute value.  The result of the matching rule is
   the result of applying the ComponentFilter to the attribute value.

   The LDAP style definitions for componentFilterMatch and its assertion
   syntax are:



Legg                     Expires 23 April 2000                 [Page 21]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   ( <OID-componentFilterMatch> NAME 'componentFilterMatch'
       SYNTAX <OID-ComponentFilter> )

   ( <OID-ComponentFilter> DESC 'ComponentFilter' )

   [ Editor's Note: <OID-componentFilterMatch> and <OID-ComponentFilter>
   to be supplied. ]

   The LDAP string encoding of a ComponentAssertion is defined by the
   following ABNF:

   ComponentAssertion = "{" sp component ","
                          [ sp useDefaultValues "," ]
                            sp rule ","
                            sp assertion-value sp "}"
   component          = component-label msp
                            dquote component-reference dquote
   useDefaultValues   = use-defaults-label msp BooleanValue
   rule               = rule-label msp MatchingRuleId
   assertion-value    = value-label msp Value

   component-label    = %x63.6F.6D.70.6F.6E.65.6E.74 ; "component"
   use-defaults-label = %x75.73.65.44.65.66.61.75.6C.74.56.61.6C.75.65.73
                                                     ; "useDefaultValues"
   rule-label         = %x72.75.6C.65                ; "rule"
   value-label        = %x76.61.6C.75.65             ; "value"

   sp                 =  *%x20  ; zero, one or more space characters
   msp                = 1*%x20  ; one or more space characters
   dquote             =   %x22  ; " (double quote)

   The matching rule in a ComponentAssertion is described by
   <MatchingRuleId>, which is defined in [RFC2251].  The ABNF for
   <Value> and <BooleanValue> is defined in Section 8.

   The LDAP string encoding of a ComponentFilter is defined by the
   following ABNF:

   ComponentFilter = filter-item / and-filter / or-filter / not-filter

   filter-item     = item-chosen ComponentAssertion
   and-filter      = and-chosen  SequenceOfComponentFilter
   or-filter       = or-chosen   SequenceOfComponentFilter
   not-filter      = not-chosen  ComponentFilter

   item-chosen     = %x69.74.65.6D.3A  ; "item:"
   and-chosen      = %x61.6E.64.3A     ; "and:"
   or-chosen       = %x6F.72.3A        ; "or:"



Legg                     Expires 23 April 2000                 [Page 22]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   not-chosen      = %x6E.6F.74.3A     ; "not:"

   SequenceOfComponentFilter = "{" [ sp ComponentFilter
                                   *( "," sp ComponentFilter) ] sp "}"

   Note that the string encodings of ComponentAssertion and
   ComponentFilter conform to the generic string encodings defined in
   Section 8 (in the event that there is a discrepancy between the above
   ABNF and Section 8, Section 8 is to be taken as definitive).

   The X.500 style definition [X501] for componentFilterMatch is:

   componentFilterMatch MATCHING-RULE ::= {
       SYNTAX  ComponentFilter
       ID      <OID-componentFilterMatch> }

   A ComponentAssertion can potentially use any matching rule, including
   componentFilterMatch, so componentFilterMatch MAY be nested.  The
   component references in a nested componentFilterMatch are relative to
   the component corresponding to the containing ComponentAssertion.  In
   Section 9, an example search on the seeAlso attribute shows this
   usage.


   7. Equality Matching of Complex Components

   It is possible to test if an attribute value of a complex ASN.1
   syntax is the same as some purported (i.e. assertion) value by using
   a complicated ComponentFilter that tests if corresponding components
   are the same.  However, it would be more convenient to be able to
   present a whole assertion value to a matching rule that could do the
   component-wise comparison of an attribute value with the assertion
   value for any arbitrary attribute syntax.  Similarly, the ability to
   do a straightforward equality comparison of a component value that is
   itself of a complex ASN.1 type would also be convenient.

   It would be difficult to define a single matching rule that
   simultaneously satisfies all notions of what the equality matching
   semantics should be.  For example, in some instances a case sensitive
   comparison of string components may be preferable to a case
   insensitive comparison.  Therefore a basic equality matching rule,
   allComponentsMatch, is defined in Section 7.1, and the means to
   derive new matching rules from it with slightly different equality
   matching semantics is described in Section 7.2.

   The directoryComponentsMatch defined in Section 7.3 is a derivation
   of allComponentsMatch that suits typical uses of the directory.
   Other specifications are free to derive new rules from



Legg                     Expires 23 April 2000                 [Page 23]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   allComponentsMatch or directoryComponentsMatch, that suit their usage
   of the directory.

   The allComponentsMatch rule, the directoryComponentsMatch rule and
   any matching rules derived from them are collectively called
   component equality matching rules.


   7.1 The allComponentsMatch Matching Rule

   The LDAP style definition for allComponentsMatch is:

   ( <OID-allComponentsMatch> NAME 'allComponentsMatch'
       SYNTAX <OID-OpenType> )

   [ Editor's Note: <OID-allComponentsMatch> to be supplied. ]

   The X.500 style definition for allComponentsMatch is:

   allComponentsMatch MATCHING-RULE ::= {
       ID      <OID-allComponentsMatch> }

   When allComponentsMatch is used in a ComponentAssertion the assertion
   syntax is the same as the ASN.1 type of the identified component.
   Otherwise, the assertion syntax of allComponentsMatch is the same as
   the attribute syntax of the attribute to which the matching rule is
   applied.

   Broadly speaking, this matching rule evaluates to true if and only if
   corresponding components of the assertion value and the attribute or
   component value are the same.

   In detail, equality is determined by the following cases applied
   recursively.

   a) Two values of a SET or SEQUENCE type are the same if and only if,
   for each component type, the corresponding component values are
   either,

      i) both absent,

      ii) both present and the same, or

      iii) absent or the same as the DEFAULT value for the component, if
      a DEFAULT value is defined.

      Values of a REAL, EMBEDDED PDV, EXTERNAL, unrestricted CHARACTER
      STRING, or INSTANCE OF type are compared according to their



Legg                     Expires 23 April 2000                 [Page 24]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


      respective SEQUENCE type (see Section 4.1.2).

   b) Two values of a SEQUENCE OF type are the same if and only if, the
   values have the same number of (possibly duplicated) instances and
   corresponding instances are the same.

   c) Two values of a SET OF type are the same if and only if, the
   values have the same number of instances and each distinct instance
   occurs in both values the same number of times, i.e. both values have
   the same instances, including duplicates, but in any order.

   d) Two values of a CHOICE type are the same if and only if, both
   values are of the same chosen alternative and the component values
   are the same.

   e) Two BIT STRING values are the same if and only if the values have
   the same number of bits and corresponding bits are the same.

   f) Two BOOLEAN values are the same if and only if both are TRUE or
   both are FALSE.

   g) Two values of a string type are the same if and only if the values
   have the same number of characters and corresponding characters are
   the same.  Letter case is significant.  For the purposes of
   allComponentsMatch, the string types are NumericString,
   PrintableString, TeletexString (T61String), VideotexString,
   IA5String, GraphicString, VisibleString (ISO646String),
   GeneralString, UniversalString, BMPString, UTF8String,
   GeneralizedTime, UTCTime and ObjectDescriptor.

   h) Two INTEGER values are the same if and only if the integers are
   equal.

   i) Two ENUMERATED values are the same if and only if the enumeration
   item identifiers are the same (equivalently, if the integer values
   associated with the identifiers are equal).

   j) Two NULL values are always the same, unconditionally.

   k) Two OBJECT IDENTIFIER values are the same if and only if the
   values have the same number of arcs and corresponding arcs are the
   same.

   l) Two OCTET STRING values are the same if and only if the values
   have the same number of octets and corresponding octets are the same.

   m) Two values of an open type are the same if and only if both are of
   the same ASN.1 type and are the same according to that type.



Legg                     Expires 23 April 2000                 [Page 25]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   Tags and constraints, being part of the type definition and not part
   of the values, are ignored for matching purposes.

   The allComponentsMatch rule MAY be used as the defined equality
   matching rule for an attribute.


   7.2 Deriving Component Equality Matching Rules

   A new component equality matching rule with more refined matching
   semantics MAY be derived from allComponentsMatch, or any other
   component equality matching rule, using the convention described in
   this section.

   The matching behaviour of a derived component equality matching rule
   is specified by nominating, for each of one or more identified
   components, an equality matching rule that will be used to match
   values of that component.  This overrides the matching that would
   otherwise occur for values of that component using the base rule for
   the derivation.  These overrides can be conveniently represented as
   rows in a table of the following form.

     Component   |  Matching Rule
     ============+===============
                 |
                 |

   Usually, all component values of a particular ASN.1 type are to be
   matched the same way.  An ASN.1 type reference (e.g.
   DistinguishedName) or an ASN.1 built-in type name (e.g. INTEGER) in
   the Component column of the table specifies that the nominated
   matching rule is to be applied to all values of the named type,
   regardless of context.

   An ASN.1 type reference with a component reference appended
   (separated by a ".")  specifies that the nominated matching rule
   applies only to the identified components of values of the named
   type.  Other component values that happen to be of the same ASN.1
   type are not selected.

   Additional type substitutions as described in Section 4.2 are assumed
   to be performed to align the component type with the matching rule
   assertion syntax.

   Conceptually, the rows in a table for the base rule are appended to
   the rows in the table for a derived rule for the purpose of deciding
   the matching semantics of the derived rule.  Notionally,
   allComponentsMatch has an empty table.



Legg                     Expires 23 April 2000                 [Page 26]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   A row specifying values of an outer containing type (e.g.
   DistinguishedName) takes precedence over a row specifying values of
   an inner component type (e.g. RelativeDistinguishedName), regardless
   of their order in the table.  Specifying a row for component values
   of an inner type is only useful if a value of the type can also
   appear on its own, or as a component of values of a different outer
   type.  For example, if there is a row for DistinguishedName then a
   row for RelativeDistinguishedName can only ever apply to
   RelativeDistinguishedName component values that are not part of a
   DistinguishedName.  A row for values of an outer type in a table for
   the base rule takes precedence over a row for values of an inner type
   in the table for the derived rule.

   Where more than one row applies to a particular component value the
   earlier row takes precedence over the later row.  Rows in the table
   for the derived rule take precedence over any applicable rows in a
   table for the base rule.


   7.3 The directoryComponentsMatch Matching Rule

   The directoryComponentsMatch matching rule is derived from the
   allComponentsMatch matching rule.

   The LDAP style definition for directoryComponentsMatch is:

   ( <OID-directoryComponentsMatch> NAME 'directoryComponentsMatch'
       SYNTAX <OID-OpenType> )

   [ Editor's Note: <OID-directoryComponentsMatch> to be supplied. ]

   The X.500 style definition for directoryComponentsMatch is:

   directoryComponentsMatch MATCHING-RULE ::= {
       ID      <OID-directoryComponentsMatch> }

   The matching semantics of directoryComponentsMatch are described by
   the following table, using the convention described in Section 7.2.

     ASN.1 Type                               | Matching Rule
     =========================================+=========================
     RDNSequence                              | distinguishedNameMatch
     RelativeDistinguishedName                | rdnMatch
     TelephoneNumber                          | telephoneNumberMatch
     FacsimileTelephoneNumber.telephoneNumber | telephoneNumberMatch
     PresentationAddress                      | presentationAddressMatch
     NumericString                            | numericStringMatch
     GeneralizedTime                          | generalizedTimeMatch



Legg                     Expires 23 April 2000                 [Page 27]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


     UTCTime                                  | uTCTimeMatch
     DirectoryString{}                        | caseIgnoreMatch
     BMPString                                | caseIgnoreMatch
     GeneralString                            | caseIgnoreMatch
     GraphicString                            | caseIgnoreMatch
     IA5String                                | caseIgnoreMatch
     PrintableString                          | caseIgnoreMatch
     TeletexString                            | caseIgnoreMatch
     UniversalString                          | caseIgnoreMatch
     UTF8String                               | caseIgnoreMatch
     VideotexString                           | caseIgnoreMatch
     VisibleString                            | caseIgnoreMatch

   Notes.

   1) The DistinguishedName type is defined by assignment to be the same
   as the RDNSequence type.  Some types (e.g. Name and LocalName)
   directly reference RDNSequence rather than DistinguishedName.
   Specifying RDNSequence captures all these DN-like types.

   2) A RelativeDistinguishedName value is only matched by rdnMatch if
   it is not part of an RDNSequence value.

   3) The telephone number component of the FacsimileTelephoneNumber
   ASN.1 type [X520] is defined as a constrained PrintableString.
   PrintableString component values that are part of a
   FacsimileTelephoneNumber value can be identified separately from
   other components of PrintableString type by the specifier
   FacsimileTelephoneNumber.telephoneNumber, so that
   telephoneNumberMatch can be selectively applied.  The fourth edition
   of X.520 defines the telephoneNumber component of
   FacsimileTelephoneNumber to be of the type TelephoneNumber, making
   the row for FacsimileTelephoneNumber.telephoneNumber components
   redundant.

   The directoryComponentsMatch rule MAY be used as the defined equality
   matching rule for an attribute.


   8. String Encodings for Values of Arbitrary ASN.1 Types

   This section defines a human readable UTF8 string encoding for an
   ASN.1 value of any given ASN.1 type.  The primary use of this
   encoding method is for the string encoding of assertion values
   appearing in ComponentAssertions, however the encoding method can
   also be nominated to define the string encoding for new attribute
   syntaxes.




Legg                     Expires 23 April 2000                 [Page 28]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   The string encodings are based on ASN.1 value notation [X680], with
   changes to accommodate the notation's use as a transfer syntax, and
   to support well established ad-hoc string encodings for directory
   data types.

   Referencing this section is sufficient to define the string encoding
   of values of a specific ASN.1 type, though other specifications may
   wish to provide a condensed, customized version of the ABNF as a
   convenience for the implementor (for example, as has been done for
   ComponentAssertion and ComponentFilter in Section 6).  Such a
   specification SHOULD state that if there is a discrepancy between the
   customized ABNF and the encoding defined in this section, that the
   encoding described by this section takes precedence.

   The string encoding of a value of any ASN.1 type is described by the
   following ABNF:

   Value = BitStringValue /
           BooleanValue /
           ChoiceValue /
           EmbeddedPDVValue /
           EnumeratedValue /
           ExternalValue /
           GeneralizedTimeValue /
           IntegerValue /
           InstanceOfValue /
           NullValue /
           ObjectDescriptorValue /
           ObjectIdentifierValue /
           OctetStringValue /
           RealValue /
           SequenceOfValue /
           SequenceValue /
           SetOfValue /
           SetValue /
           StringValue /
           UTCTimeValue /
           VariantEncoding

   A value of a type with a defined type name is encoded according to
   the type definition on the right hand side of the type assignment for
   the type name.

   A value of a type defined by the use of a parameterized type with
   actual parameters is encoded according to the parameterized type with
   the DummyReferences substituted with the actual parameters.

   A value of a tagged or constrained type is encoded as a value of the



Legg                     Expires 23 April 2000                 [Page 29]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   type without the tag or constraint, respectively.  Tags do not appear
   in the string encodings defined by this document.

   A value of an open type defined by an ObjectClassFieldType is encoded
   according to the specific type of the value.

   A value of a fixed type defined by an ObjectClassFieldType is encoded
   according to that fixed type.

   A value of a selection type is encoded according to the type
   referenced by the selection type.

   A value of a type described by TypeFromObject notation is encoded
   according to the denoted type.

   A value of a type described by ValueSetFromObjects notation is
   encoded according to the governing type.

   ASN.1 identifiers figure prominently in the string encodings.  The
   ABNF for an identifier was first given in Section 4.1 but is repeated
   here for convenience.  The case of letters in an identifier is always
   significant.

   identifier    = lowercase *alphanumeric *(hyphen 1*alphanumeric)
   alphanumeric  = uppercase / lowercase / decimal-digit
   uppercase     = %x41-5A  ; "A" to "Z"
   lowercase     = %x61-7A  ; "a" to "z"
   decimal-digit = %x30-39  ; "0" to "9"
   hyphen        = "-"

   A value of the BIT STRING type is encoded according to the
   <BitStringValue> rule.  If the definition of the BIT STRING type
   includes a named bit list, the <bit-list> form of <BitStringValue>
   rule MAY be used.  If the number of bits in a BIT STRING value is a
   multiple of four the <hstring> form of <BitStringValue> MAY be used.
   The <bstring> form of <BitStringValue> is used otherwise.

   BitStringValue = bstring / hstring / bit-list

   The <bit-list> rule encodes the one bits in the bit string value as a
   comma separated list of identifiers.  Each <identifier> MUST be one
   of those in the named bit list.  An <identifier> MUST NOT appear more
   than once in the same <bit-list>.  The <bstring> rule encodes each
   bit as the character "0" or "1" in order from the first bit to the
   last bit.  The <hstring> rule encodes each group of four bits as a
   hexadecimal number where the first bit is the most significant.  An
   odd number of hexadecimal digits is permitted.




Legg                     Expires 23 April 2000                 [Page 30]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   bit-list          = "{" [ sp identifier *( "," sp identifier) ] sp "}"

   hstring           = squote *hexadecimal-digit squote %x48  ; '...'H
   hexadecimal-digit = %x30-39 /  ; "0" to "9"
                       %x41-46    ; "A" to "F"

   bstring           = squote *binary-digit squote %x42  ; '...'B
   binary-digit      = "0" / "1"

   sp                = *%x20  ; zero, one or more space characters
   squote            =  %x27  ; ' (single quote)

   A value of the BOOLEAN type is encoded according to the
   <BooleanValue> rule.

   BooleanValue = %x54.52.55.45 /   ; "TRUE"
                  %x46.41.4C.53.45  ; "FALSE"

   A value of the INTEGER type is encoded according to the
   <IntegerValue> rule.  If the definition of the INTEGER type includes
   a named number list, the <identifier> form of <IntegerValue> MAY be
   used, in which case the <identifier> MUST be one of those in the
   named number list.

   IntegerValue    = "0" /
                     positive-number /
                     ("-" positive-number) /
                     identifier

   positive-number = non-zero-digit *decimal-digit
   non-zero-digit  = %x31-39  ; "1" to "9"

   A value of the ENUMERATED type is encoded according to the
   <EnumeratedValue> rule.  The <identifier> MUST be one of those in the
   list of enumerations in the definition of the ENUMERATED type.

   EnumeratedValue = identifier

   A value of the NULL type is encoded according to the <NullValue>
   rule.

   NullValue = %x4E.55.4C.4C  ; "NULL"

   An OBJECT IDENTIFIER value is encoded using the string representation
   described by the <oid> rule in [RFC2252].

   ObjectIdentifierValue = oid




Legg                     Expires 23 April 2000                 [Page 31]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   [ Editor's Note: The <oid> rule allows either a dotted decimal
   representation of the OBJECT IDENTIFIER or an object descriptor name.
   I would prefer to allow only the dotted decimal representation for
   ObjectIdentifierValue instead of perpetuating the potentially
   ambiguous object descriptor names. ]

   A value of the OCTET STRING type is encoded according to the
   <OctetStringValue> rule.  The octets are encoded in order from the
   first octet to the last octet.  Each octet is encoded as a pair of
   hexadecimal digits where the first digit corresponds to the four most
   significant bits of the octet.  If the hexadecimal string does not
   have an even number of digits the four least significant bits in the
   last octet are assumed to be zero.

   OctetStringValue = hstring

   The contents of a string value are encoded as a UTF8 character string
   between double quotes.  Depending on the ASN.1 string type, and an
   application's internal representation of that string type, a
   translation to or from the UTF8 character encoding may be required.
   NumericString, PrintableString, IA5String, VisibleString
   (ISO646String) are compatible with UTF8 and do not require any
   translation.  BMPString (UCS-2) and UniversalString (UCS-4) have a
   direct mapping to and from UTF8 [RFC2279].  For the remaining string
   types see [X680].  Any embedded double quotes in the resulting UTF8
   character string are escaped by repeating the double quotes
   character.

   StringValue       = dquote *SafeUTF8Character dquote

   dquote            = %x22 ; " (double quote)

   SafeUTF8Character = %x01-21 / %x23-7F /   ; ASCII minus dquote
                       dquote dquote /       ; escaped double quote
                       %xCO-DF %x80-BF /     ; 2 byte UTF8 character
                       %xEO-EF 2(%x80-BF) /  ; 3 byte UTF8 character
                       %xFO-F7 3(%x80-BF) /  ; 4 byte UTF8 character
                       %xF8-FB 4(%x80-BF) /  ; 5 byte UTF8 character
                       %xFC-FD 5(%x80-BF)    ; 6 byte UTF8 character

   A value of the GeneralizedTime type, UTCTime type or ObjectDescriptor
   type is encoded as a string value.  GeneralizedTime and UTCTime use
   the VisibleString character set so the conversion to UTF8 is trivial.
   ObjectDescriptor uses the GraphicString type.

   GeneralizedTimeValue  = StringValue
   UTCTimeValue          = StringValue
   ObjectDescriptorValue = StringValue



Legg                     Expires 23 April 2000                 [Page 32]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   A value of a CHOICE type is encoded according to the <ChoiceValue>
   rule.  The <ChoiceOfStringsValue> encoding MAY be used if the
   corresponding CHOICE type has been declared a ChoiceOfStrings type.
   This document declares DirectoryString to be a ChoiceOfStrings type
   (see Section 4.2.1.1).  The <IdentifiedChoiceValue> form of
   <ChoiceValue> is used otherwise.

   ChoiceValue           = IdentifiedChoiceValue /
                           ChoiceOfStringsValue

   IdentifiedChoiceValue = identifier ":" Value
   ChoiceOfStringsValue  = StringValue

   For implementations that recognise the internal structure of the
   DirectoryString CHOICE type (e.g. X.500 directories), if the
   character string between the quotes in a <StringValue> contains only
   characters that are permitted in a PrintableString the
   DirectoryString is assumed to use the printableString alternative,
   otherwise it is assumed to use the uTF8String alternative.  The
   <IdentifiedChoiceValue> rule MAY be used for a value of type
   DirectoryString to indicate a different alternative to the one that
   would otherwise be assumed from the string contents.  No matter what
   alternative is chosen, the <Value> will still be a UTF8 encoded
   character string, however it is a syntax error if the characters in
   the UTF8 string cannot be represented in the chosen string type.

   Implementations that don't care about the internal structure of a
   DirectoryString value MUST be able to parse the
   <IdentifiedChoiceValue> form, though the particular identifier found
   will be of no interest.

   A value of a SEQUENCE type is encoded according to the
   <SequenceValue> rule.  The <ComponentList> rule encodes a comma
   separated list of the particular component values present in the
   SEQUENCE value, where each component value is preceded by the
   corresponding identifier from the SEQUENCE type definition.  The
   components are encoded in the order of their definition in the
   SEQUENCE type.

   SequenceValue = ComponentList

   ComponentList = "{" [ sp NamedValue *( "," sp NamedValue) ] sp "}"
   NamedValue    = identifier msp Value
   msp           = 1*%x20  ; one or more space characters

   A value of a SET type is encoded according to the <SetValue> rule.
   The components are encoded in the order of their definition in
   the SET type (i.e. just like a SEQUENCE value).



Legg                     Expires 23 April 2000                 [Page 33]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   This is a deliberate departure from ASN.1 value notation where
   the components of a SET can be written in any order.

   SetValue = ComponentList

   SEQUENCE and SET type definitions are sometimes extended by the
   inclusion of additional component types, so an implementation SHOULD
   be capable of skipping over any <NamedValue> encoding with an
   identifier that is not recognised, on the assumption that the sender
   is using a more recent definition of the SEQUENCE or SET type.

   A value of a SEQUENCE OF type is encoded according to the
   <SequenceOfValue> rule, as a comma separated list of the instances in
   the value.  Each instance is encoded according to the component type
   of the SEQUENCE OF type.

   SequenceOfValue = "{" [ sp Value *( "," sp Value) ] sp "}"

   A value of a SET OF type is encoded according to the <SetOfValue>
   rule, as a list of the instances in the value.  Each instance is
   encoded according to the component type of the SET OF type.

   SetOfValue      = "{" [ sp Value *( "," sp Value) ] sp "}"

   A value of an EMBEDDED PDV, EXTERNAL, unrestricted CHARACTER STRING,
   or INSTANCE OF type is encoded according to the corresponding
   SEQUENCE type (see Section 4.1.2).

   EmbeddedPDVValue = SequenceValue
   ExternalValue    = SequenceValue
   InstanceOfValue  = SequenceValue

   A value of the REAL type is encoded as "0" if it is zero, or as the
   corresponding SEQUENCE type otherwise (see Section 4.1.2).

   RealValue = "0" /          ; zero REAL value
               SequenceValue  ; non-zero REAL value

   The values of some named complex ASN.1 types have special string
   encodings.  These special encodings are always used instead of the
   encoding that would otherwise apply based on the type definition.

   VariantEncoding = RDNSequenceValue /
                     RelativeDistinguishedNameValue /
                     ORAddressValue

   A value of the RDNSequence type, i.e. a distinguished name, is
   encoded according to the <RDNSequenceValue> rule, as a quoted LDAPDN



Legg                     Expires 23 April 2000                 [Page 34]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   character string.  The character string is first derived according to
   the <distinguishedName> rule in Section 3 of [RFC2253], and then it
   is encoded as if it were a UTF8String value, i.e. between double
   quotes with any embedded double quotes escaped by being repeated.

   RDNSequenceValue = StringValue

   A RelativeDistinguishedName value that is not part of an RDNSequence
   value is encoded according to the <RelativeDistinguishedNameValue>
   rule as a quoted character string.  The character string is first
   derived according to the <name-component> rule in Section 3 of
   [RFC2253], and then it is encoded as if it were a UTF8String value.

   RelativeDistinguishedNameValue = StringValue

   A value of the ORAddress type is encoded according to the
   <ORAddressValue> rule as a quoted character string.  The character
   string is first derived according to the textual representation of
   MTS.ORAddress from [RFC2156], and then it is encoded as if it were an
   IA5String value.

   ORAddressValue = StringValue


   9. Component Matching Examples

   This section contains examples of search filters using the
   componentFilterMatch matching rule.  The filters are described using
   the string representation of LDAP search filters from [RFC2254].
   Additional line breaks and indenting have been added only as an aid
   to readability.

   The example search filters are all single extensible match filter
   items, though there is no reason why componentFilterMatch can't be
   used in more complicated search filters.

   The first examples describe searches over the objectClasses schema
   operational attribute, which has an attribute syntax described by the
   ASN.1 type ObjectClassDescription [X501], and holds the definitions
   of the object classes known to the directory server.  The definition
   of ObjectClassDescription is as follows:

   ObjectClassDescription ::= SEQUENCE {
       identifier       OBJECT-CLASS.&id,
       name             SET OF DirectoryString { ub-schema } OPTIONAL,
       description      DirectoryString { ub-schema } OPTIONAL,
       obsolete         BOOLEAN DEFAULT FALSE,
       information  [0] ObjectClassInformation }



Legg                     Expires 23 April 2000                 [Page 35]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   ObjectClassInformation ::= SEQUENCE {
       subclassOf       SET OF OBJECT-CLASS.&id OPTIONAL,
       kind             ObjectClassKind DEFAULT structural,
       mandatories  [3] SET OF ATTRIBUTE.&id OPTIONAL,
       optionals    [4] SET OF ATTRIBUTE.&id OPTIONAL }

   ObjectClassKind ::= ENUMERATED {
       abstract     (0),
       structural   (1),
       auxiliary    (2) }

   OBJECT-CLASS.&id and ATTRIBUTE.&id are equivalent to the OBJECT
   IDENTIFIER ASN.1 type.  A value of OBJECT-CLASS.&id is an OBJECT
   IDENTIFIER for an object class.  A value of ATTRIBUTE.&id is an
   OBJECT IDENTIFIER for an attribute type.

   The following search filter finds the object class definition for the
   object class identified by the OBJECT IDENTIFIER 2.5.6.18:

   (objectClasses:componentFilterMatch:=item:{
       component "identifier",
       rule objectIdentifierMatch,
       value 2.5.6.18 })

   A match on the "identifier" component of objectClasses values is
   equivalent to the objectIdentifierFirstComponentMatch matching rule
   applied to attribute values of the objectClasses attribute type.  The
   componentFilterMatch matching rule subsumes the functionality of the
   objectIdentifierFirstComponentMatch, integerFirstComponentMatch and
   directoryStringFirstComponentMatch matching rules.

   The following search filter finds the object class definition for the
   object class called foobar:

   (objectClasses:componentFilterMatch:=item:{
       component "name.*",
       rule caseIgnoreMatch,
       value "foobar" })

   An object class definition can have multiple names and the above
   filter will match an objectClasses value if any one of the names is
   "foobar".

   The component reference "name.0" identifies the notional count of the
   number of names in an object class definition.  The following search
   filter finds object class definitions with exactly one name:

   (objectClasses:componentFilterMatch:=



Legg                     Expires 23 April 2000                 [Page 36]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


       item:{ component "name.0", rule integerMatch, value 1 })

   The "description" component of an ObjectClassDescription is defined
   to be an OPTIONAL DirectoryString.  The following search filter finds
   object class definitions that have descriptions, regardless of the
   contents of the description string.

   (objectClasses:componentFilterMatch:=
       item:{ component "description", rule presentMatch, value NULL })

   The presentMatch returns TRUE if the description component is present
   and FALSE otherwise.

   The following search filter finds object class definitions that don't
   have descriptions.

   (objectClasses:componentFilterMatch:=
       not:item:{ component "description", rule presentMatch, value NULL })

   The following search filter finds object class definitions with the
   word "bogus" in the description:

   (objectClasses:componentFilterMatch:=
       not:item:{ component "description", rule caseIgnoreSubstringsMatch,
       value { any:"bogus" } })

   The assertion value is of the SubstringAssertion syntax, i.e.

   SubstringAssertion ::= SEQUENCE OF CHOICE {
       initial      [0] DirectoryString {ub-match},
       any          [1] DirectoryString {ub-match},
       final        [2] DirectoryString {ub-match} }

   The "obsolete" component of an ObjectClassDescription is defined to
   be DEFAULT FALSE.  An object class is obsolete if the "obsolete"
   component is present and set to TRUE.  The following search filter
   finds all obsolete object classes:

   (objectClasses:componentFilterMatch:=
       item:{ component "obsolete", rule booleanMatch, value TRUE })

   An object class is not obsolete if the "obsolete" component is not
   present, in which case it defaults to FALSE, or is present but is
   explicitly set to FALSE.  The following search filter finds all non-
   obsolete object classes.

   (objectClasses:componentFilterMatch:=
       item:{ component "obsolete", rule booleanMatch, value FALSE })



Legg                     Expires 23 April 2000                 [Page 37]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   The useDefaultValues flag in the ComponentAssertion defaults to TRUE
   so the componentFilterMatch rule treats an absent "obsolete"
   component as being present and set to FALSE.  The following search
   filter finds only object class definitions where the "obsolete"
   component has been explicitly set to FALSE, rather than implicitly
   defaulting to FALSE.

   (objectClasses:componentFilterMatch:=
       item:{ component "obsolete",
              useDefaultValues FALSE, rule booleanMatch, value FALSE })

   With the useDefaultValues flag set to FALSE, if the "obsolete"
   component is absent the component reference identifies no component
   value and the matching rule will return FALSE.  The matching rule can
   only return TRUE if the component is present and set to FALSE.

   The "information.kind" component of the ObjectClassDescription is an
   ENUMERATED type.  The following search filter finds object class
   definitions for auxiliary object classes.

   (objectClasses:componentFilterMatch:=
       item:{ component "information.kind",
              rule enumeratedMatch, value auxiliary })

   The following search filter finds auxiliary object classes with
   commonName (cn or 2.5.4.3) as a mandatory attribute:

   (objectClasses:componentFilterMatch:=and:{
       item:{ component "information.kind",
              rule enumeratedMatch, value auxiliary },
       item:{ component "information.mandatories.*",
              rule objectIdentifierMatch, value cn } })

   The following search filter finds auxiliary object classes with
   commonName as a mandatory or optional attribute:

   (objectClasses:componentFilterMatch:=and:{
       item:{ component "information.kind",
              rule enumeratedMatch, value auxiliary },
       or:{
           item:{ component "information.mandatories.*",
                  rule objectIdentifierMatch, value cn },
           item:{ component "information.optionals.*",
                  rule objectIdentifierMatch, value cn } } })

   Extra care is required when matching optional SEQUENCE OF or SET OF
   components because of the distinction between an absent list of
   instances and a present, but empty, list of instances.  The following



Legg                     Expires 23 April 2000                 [Page 38]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   search filter finds object class definitions with less than three
   names, including object class definitions with a present but empty
   list of names, but does not find object class definitions with an
   absent list of names.

   (objectClasses:componentFilterMatch:=
       item:{ component "name.0", rule integerOrderingMatch, value 3 })

   If the "name" component is absent the "name.0" component is also
   considered to be absent and the ComponentAssertion evaluates to
   FALSE.  If the "name" component is present, but empty, the "name.0"
   component is also present and equal to zero, so the
   ComponentAssertion evaluates to TRUE.  To also find the object class
   definitions with an absent list of names the following search filter
   would be used:

   (objectClasses:componentFilterMatch:=or:{
       not:item:{ component "name", rule presentMatch, value NULL },
       item:{ component "name.0", rule integerOrderingMatch, value 3 } })

   Distinguished names embedded in other syntaxes can be matched with a
   componentFilterMatch.  The uniqueMember attribute type has an
   attribute syntax described by the ASN.1 type NameAndOptionalUID.

   NameAndOptionalUID ::= SEQUENCE {
       dn        DistinguishedName,
       uid       UniqueIdentifier OPTIONAL }

   The following search filter finds values of the uniqueMember
   attribute containing the author's DN:

   (uniqueMember:componentFilterMatch:={ component "dn",
       rule distinguishedNameMatch, value "cn=Steven Legg, o=Adacel, c=au" })

   The DistinguishedName and RelativeDistinguishedName ASN.1 types are
   also complex ASN.1 types so the component matching rules can be
   applied to their inner components.

   DistinguishedName   ::= RDNSequence

   RDNSequence ::= SEQUENCE OF RelativeDistinguishedName

   RelativeDistinguishedName ::= SET SIZE (1 .. MAX) OF AttributeTypeAndValue

   AttributeTypeAndValue ::= SEQUENCE {
       type        AttributeType ({SupportedAttributes}),
       value       AttributeValue ({SupportedAttributes}{@type}) }




Legg                     Expires 23 April 2000                 [Page 39]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   AttributeType ::= ATTRIBUTE.&id

   AttributeValue ::= ATTRIBUTE.&Type

   ATTRIBUTE.&Type is an open type.  A value of ATTRIBUTE.&Type is
   constrained by the type component of AttributeTypeAndValue to be of
   the attribute syntax of the nominated attribute type.  Note: the
   fourth edition of X.500 extends and renames the AttributeTypeAndValue
   SEQUENCE type.

   The seeAlso attribute has the DistinguishedName syntax.  The
   following search filter finds seeAlso attribute values containing the
   RDN, "o=Adacel", anywhere in the DN:

   (seeAlso:componentFilterMatch:=
       item:{ component "*", rule rdnMatch, value "o=Adacel" })

   The following search filter finds all seeAlso attribute values with
   "cn=Steven Legg" as the RDN of the named entry (i.e. the "first" RDN
   in an LDAPDN or the "last" RDN in an X.500 DN).

   (seeAlso:componentFilterMatch:=
       item:{ component "-1", rule rdnMatch, value "cn=Steven Legg" })

   The following search filter finds all seeAlso attribute values naming
   entries in the DIT subtree of "o=Adacel, c=au":

   (seeAlso:componentFilterMatch:=and:{
       item:{ component "1", rule rdnMatch, value "c=au" },
       item:{ component "2", rule rdnMatch, value "o=Adacel" } })

   The following search filter finds all seeAlso attribute values
   containing the naming attributes commonName (cn) and telephoneNumber
   in the same RDN:

   (seeAlso:componentFilterMatch:=
       item:{ component "*", rule componentFilterMatch,
              value and:{
                  item:{ component "*.type",
                         rule objectIdentifierMatch, value cn },
                  item:{ component "*.type",
                         rule objectIdentifierMatch,
                         value telephoneNumber } } })

   The following search filter would find all seeAlso attribute values
   containing the attributes commonName and telephoneNumber, but not
   necessarily in the same RDN:




Legg                     Expires 23 April 2000                 [Page 40]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   (seeAlso:componentFilterMatch:=and:{
       item:{ component "*.*.type",
              rule objectIdentifierMatch, value cn },
       item:{ component "*.*.type",
              rule objectIdentifierMatch, value telephoneNumber } })

   The following search filter finds all seeAlso attribute values
   containing the word "Adacel" in any AttributeTypeAndValue of any RDN:

   (seeAlso:componentFilterMatch:=
       item:{ component "*.*.value", rule caseIgnoreSubstringsMatch,
              value { any:"Adacel" } })

   The component reference "*.*.value" identifies an open type.  If the
   actual ASN.1 type of a value of this component is not compatible with
   the caseIgnoreSubstringsMatch then the ComponentAssertion evaluates
   to FALSE.  Otherwise the substring assertion is evaluated against the
   component value.


   10. Security Considerations

   The component matching rules described in this document allow for a
   compact specification of matching capabilities that could otherwise
   have been defined by a plethora of specific matching rules, i.e.
   despite their expressiveness and flexibility the component matching
   rules do not behave in a way uncharacteristic of other matching
   rules, so the security issues for component matching rules are no
   different than for any other matching rule.  However, because the
   component matching rules are applicable to any attribute syntax,
   support for them in a directory server may allow searching of
   attributes that were previously unsearchable by virtue of there not
   being a suitable matching rule.  Such attribute types ought to be
   properly protected with appropriate access controls.

   The generic string encodings in Section 8 do not necessarily enable
   the exact octet encoding of values of TeletexString, VideotexString,
   GraphicString or GeneralString to be reconstructed, so a
   transformation from DER to generic string encoding and back to DER
   may not reproduce the original DER encoding.  This has consequences
   for the verification of digital signatures.


   11. Acknowledgements

   The author would like to thank Tom Gindin for private email
   discussions that clarified and refined the ideas presented in this
   document.



Legg                     Expires 23 April 2000                 [Page 41]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   12. References

   [BCP-11] - R. Hovey, S. Bradner, "The Organizations Involved in the
   IETF Standards Process", BCP 11, RFC 2028, October 1996.

   [RFC2119] - S. Bradner, "Key words for use in RFCs to Indicate
   Requirement Levels", RFC 2119.

   [RFC2156] - S. Kille, "MIXER (Mime Internet X.400 Enhanced Relay):
   Mapping between X.400 and RFC 822/MIME", RFC 2156, January 1998.

   [RFC2234] - D. Crocker, P. Overell, "Augmented BNF for Syntax
   Specifications:  ABNF", RFC 2234, November 1997.

   [RFC2251] - M. Wahl, T. Howes, S. Kille, "Lightweight Directory
   Access Protocol (v3)", RFC 2251, December 1997.

   [RFC2252] - M. Wahl, A. Coulbeck, T. Howes, S. Kille, "Lightweight
   Directory Access Protocol (v3): Attribute Syntax Definitions", RFC
   2252, December 1997.

   [RFC2253] M. Wahl, S. Kille, T. Howes. "Lightweight Directory Access
   Protocol (v3): UTF-8 String Representation of Distinguished Names",
   RFC2253, December 1997.

   [RFC2254] - T. Howes, "The String Representation of LDAP Search
   Filters", RFC 2254, December 1997.

   [RFC2279] - F. Yergeau, "UTF-8, a transformation format of ISO
   10646", RFC 2279, January 1998.

   [SYNTAX] - S. Legg, "ASN.1 Types for LDAP Syntaxes", draft-legg-
   ldapext-asn1-for-syntaxes-00.txt (to be published).

   [X500] - ITU-T Recommendation X.500 (1993) | ISO/IEC 9594-1:1994,
   Information Technology - Open Systems Interconnection - The
   Directory:  Overview of concepts, models and services

   [X501] - ITU-T Recommendation X.501 (1993) | ISO/IEC 9594-2:1994,
   Information Technology - Open Systems Interconnection - The
   Directory: Models

   [X520] - ITU-T Recommendation X.511 (1993) | ISO/IEC 9594-6:1994,
   Information Technology - Open Systems Interconnection - The
   Directory:  Selected attribute types

   [X680] - ITU-T Recommendation X.680 (1997) | ISO/IEC 8824-1:1998
   Information Technology - Abstract Syntax Notation One (ASN.1):



Legg                     Expires 23 April 2000                 [Page 42]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   Specification of basic notation

   [X681] - ITU-T Recommendation X.681 (1997) | ISO/IEC 8824-2:1998
   Information Technology - Abstract Syntax Notation One (ASN.1):
   Information object specification

   [X682] - ITU-T Recommendation X.682 (1997) | ISO/IEC 8824-3:1998
   Information Technology - Abstract Syntax Notation One (ASN.1):
   Constraint specification

   [X683] - ITU-T Recommendation X.683 (1997) | ISO/IEC 8824-4:1998
   Information Technology - Abstract Syntax Notation One (ASN.1):
   Parameterization of ASN.1 specifications


   13. Intellectual Property Notice

   The IETF takes no position regarding the validity or scope of any
   intellectual property or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; neither does it represent that it
   has made any effort to identify any such rights.  Information on the
   IETF's procedures with respect to rights in standards-track and
   standards-related documentation can be found in BCP-11. [BCP-11]
   Copies of claims of rights made available for publication and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementors or users of this
   specification can be obtained from the IETF Secretariat.

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights which may cover technology that may be required to practice
   this standard.  Please address the information to the IETF Executive
   Director.


   14. Copyright Notice

      Copyright (C) The Internet Society (2000). All Rights Reserved.

   This document and translations of it may be copied and furnished to
   others, and derivative works that comment on or otherwise explain it
   or assist in its implementation may be prepared, copied, published
   and distributed, in whole or in part, without restriction of any
   kind, provided that the above copyright notice and this paragraph are
   included on all such copies and derivative works.  However, this



Legg                     Expires 23 April 2000                 [Page 43]

INTERNET-DRAFT   LDAP & X.500 Component Matching Rules  October 23, 2000


   document itself may not be modified in any way, such as by removing
   the copyright notice or references to the Internet Society or other
   Internet organizations, except as needed for the purpose of
   developing Internet standards in which case the procedures for
   copyrights defined in the Internet Standards process must be
   followed, or as required to translate it into languages other than
   English.

   The limited permissions granted above are perpetual and will not be
   revoked by the Internet Society or its successors or assigns.

   This document and the information contained herein is provided on an
   "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
   TASK FORCE DISCLAIMS ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING
   BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION
   HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF
   MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.


   15. Author's Address

   Steven Legg
   Adacel Technologies Ltd.
   250 Bay Street
   Brighton, Victoria 3186
   AUSTRALIA

   Phone: +61 3 8530 7808
     Fax: +61 3 9596 2960
   EMail: steven.legg@adacel.com.au





















Legg                     Expires 23 April 2000                 [Page 44]


Html markup produced by rfcmarkup 1.107, available from http://tools.ietf.org/tools/rfcmarkup/