The Rule Markup Initiative: KR Principles and DTD Modularization

1
The Rule Markup Initiative KR Principles and
DTD Modularization

• Harold Boley(joint work with
• Benjamin Grosof and
• Said Tabet)

• 23th Meeting of the DFKI Scientific Advisory
  Board
• February 19-21, 2001

2
Motivation (I)

• Rules in (and for) the Web have become a
  mainstream topic since
• inference rules were
• marked up for E-Commerce
• identified as a Design Issue of the Semantic Web
• transformation rules were used for document
  generation from central XML repository

• Rule interchange is becoming more important in
  Knowledge Representation (KR), especially in
• Intelligent Agents
• AI shells for knowledge-based systems

3
Motivation (II)

• The Rule Markup Initiative has taken initial
  steps towards defining a shared Rule Markup
  Language (RuleML) for interoperation between
  Participants
• RuleML permits forward (bottom-up) and backward
  (top-down) rules in XML for
• deduction
• rewriting
• further inferential-transformational tasks
• Current version uses backward notation, which can
  be applied in both directions

4
Initial Example Forward-Rule Notation
Challenge hypertext as one XHTML paragraph
ltpgtIf you want to review rule principles, you may
look at lta href"http//www.cs.brandeis.edu/
..."gtRule-Based Systemslt/agtlt/pgt
Original RuleML markup with XHTML in prem/conc
(English premise and semiformal conclusion)

• ltrulegt
• ltpremgt
• ltpgtYou want to review rule principleslt/pgt
• lt/premgt
• ltconcgt
• ltpgtYou may look at
• lta href"http//www.cs.brandeis.edu/..."gtRul
  e-Based Systemslt/agt
• lt/pgt
• lt/concgt
• lt/rulegt

5
Initial Example Backward-Rule Notation
Further formalized RuleML markup (still
unanalyzed English relation and
individual-constant names)

• ltifgt
• ltatomgt
• ltrelgtmay look atlt/relgt
• ltvargtyoult/vargt
• ltur label"Rule-Based Systems"gthttp//www.cs.b
  randeis.edu/...lt/urgt
• lt/atomgt
• ltatomgt
• ltrelgtwant to reviewlt/relgt
• ltvargtyoult/vargt
• ltindgtrule principleslt/indgt
• lt/atomgt
• lt/ifgt

6
RuleML Elements of Datalog DTD Clocksin/Mellish
Sample Prolog Clauses
Rule (Non-unit clause)
Fact (Unit clause)
ltifgt ltatomgt ltrelgtlikeslt/relgt
ltindgtMarylt/indgt ltindgtwinelt/indgt lt/atomgt
ltand/gt lt/ifgt likes(mary,wine).
ltifgt ltatomgt ltrelgtlikeslt/relgt likes(
ltindgtJohnlt/indgt john, ltvargtxlt/vargt X )
lt/atomgt - ltatomgt ltrelgtlikeslt/relgt
likes( ltvargtxlt/vargt X,
ltindgtwinelt/indgt wine ) . lt/atomgt lt/ifgt
Empty and ? true premise ? factual rule
7
RuleML Element of UR-Hornlog DTD Proposed
W3C-Page Authentication Rule
Tim Berners-Lee Any person who was some time in
the last 2 months an employee of an organization
which was some time in the last 2 months a W3C
member may register
ltatomgt ltrelgtmember inlt/relgt
ltvargtorglt/vargt ltur label"W3C"gthttp//
www.w3.org/lt/urgt ltctermgt
ltctorgtlastlt/ctorgt ltctermgt
ltctorgtmonthlt/ctorgt ltindgt2lt/indgt
lt/ctermgt lt/ctermgt lt/atomgt
lt/andgt lt/ifgt
ltifgt ltatomgt ltrelgtmay registerlt/relgt
ltvargtanylt/vargt lt/atomgt ltandgt ltatomgt
ltrelgtpersonlt/relgt ltvargtanylt/vargt
lt/atomgt ltatomgt ltrelgtorganizationlt/relgt
ltvargtorglt/vargt lt/atomgt
ltatomgt ltrelgtemployee inlt/relgt
ltvargtanylt/vargt ltvargtorglt/vargt
ltctermgt ltctorgtlastlt/ctorgt
ltctermgt ltctorgtmonthlt/ctorgt
ltindgt2lt/indgt lt/ctermgt lt/ctermgt
lt/atomgt
8
RuleML Elements of UR-Equalog DTD Equations for
URI Expansion
URLs/URIs or URs as 1st-class citizens
uriexp(daml) http//www.daml.org/
ltifgt lteqgt ltnanogt ltfungturiexplt/fungt
ltindgtdamllt/indgt lt/nanogt
lturgthttp//www.daml.org/lt/urgt lt/eqgt
ltand/gt lt/ifgt
uriexp(oil) http//www.ontoknowledge.org/oil/
ltifgt lteqgt ltnanogt ltfungturiexplt/fungt
ltindgtoillt/indgt lt/nanogt
lturgthttp//www.ontoknowledge.org/oil/lt/urgt
lt/eqgt ltand/gt lt/ifgt
Empty and ? true premise ? unconditional equatio
n
. . .
9
RuleML Element of URC-Bin-Data-Ground-Fact
DTDRDF Triple as Very Special Rule
RDF triple (predicate, subject, object) as
atom predicate(subject, object) or rel(ur,
urind) "http//www.w3.org/Home/Lassila has
creator Ora Lassila." (Creator,
http//www.w3.org/Home/Lassila, Ora Lassila)
ltifgt ltatomgt ltrelgtCreatorlt/relgt
lturgthttp//www.w3.org/Home/Lassilalt/urgt
ltindgtOra Lassilalt/indgt lt/atomgt ltand/gt lt/ifgt
10
Modularization of DTDs XHTML and KR

• Advantages
• Leads to reusable subDTDs and DTD interoperation
• Complex DTDs built with 'plug-and-play'
  technology
• (RuleML) Sublanguages determined by validations!
• For (rulebase) export find most precise
  sublanguage!

• Modular DTDs still mostly used outside KR
• First used for XHTML and described in XML Bible
• W3C Working Draft 5 January 2000
  Building XHTMLtm Modules
• W3C Candidate Recommendation 20 October 2000
  Modularization of XHTMLtm

11
Structure of the RuleML DTD Hierarchy

• Our system of DTDs (current version 0.7) uses a
  modularization approach similar to XHTML in order
  to accomodate the various rule subcommunities
• The evolving hierarchy of RuleML DTDs forms
  a partial order with ruleml as the greatest
  element (a ruleml-rooted DAG) -- many
  smallest elements
• Each DTD node in the hierarchy corresponds to a
  specific RuleML sublanguage
• Union (join) of sublanguages reached via
  outgoing links to smaller or equal nodes below
• Intersection (meet) of sublanguages via
  incoming links from greater or equal nodes above

12
The Module Hierarchy of RuleML DTDs
ruleml
ur-equalog
Rooted DAG will be extended with branches for
further sublanguages
equalog
ur-hornlog
hornlog
ur-datalog
ur-datalog join(ur,datalog)
datalog
bin-datalog
urc-datalog
ur
URL/URI-like ur-objects
urc-bin-datalog
urc-bin-data-ground-log
urc-bin-data-ground-fact
RDF-like triples
13
DTDs From Well-Formedness to Validity
XML principles for a document being
well-formed
XML principle for a document
being valid with respect to a DTD

• Open and close all tags
• Empty tags end with /gt
• There is a unique root element
• Elements may not overlap
• Attribute values are quoted
• lt and are only used to start tags and entities
• Only the five predefined entity references are
  used

• Matches the type-like constraints listed in the
  DTD (or, can be generated from DTD as linearized
  CF grammar-derivation tree)

Checked by validators such as http//www.stg.brown
.edu/service/xmlvalid/
14
A Relational Languageruleml-datalog.dtd (I)
lt!-- An XML DTD for a Datalog RuleML Sublanguage
--gt lt!-- Last Modification 2001-01-25 --gt lt!--
ENTITY Declarations --gt lt!-- in this
ruleml-datalog.dtd, parameter entities set two
.module switches to INCLUDE --gt lt!ENTITY
datalog.module "INCLUDE"gt lt!ENTITY
datalog-and-hornlog.module "INCLUDE"gt lt!-- hence
all conditional sections "lt!.module" . . .
"gt" activate their content --gt lt!-- in a
stand-alone use of the current DTD
"lt!.module" and "gt" are thus no-ops
--gt lt!datalog-and-hornlog.module lt!-- a
conclusion and premise will be usable within 'if'
implications --gt lt!-- in datalog and hornlog,
conc element uses an atomic formula --gt lt!-- in
datalog and hornlog, prem element uses an atomic
formula or an 'and' --gt lt!ENTITY conc
"atom"gt lt!ENTITY prem "(atom and)"gt gt
15
A Relational Languageruleml-datalog.dtd (II)
lt!-- ELEMENT and ATTLIST Declarations --gt lt!--
'rulebase' root element uses 'if' implications as
top-level rules --gt lt!-- label attribute allows
naming of an entire individual rulebase --gt lt!--
e.g., this can help enable forward inferencing of
selected rulebase(s) --gt lt!-- direction
attribute indicates the intended direction of
rule inferencing --gt lt!-- it is a preliminary
design choice and has a 'neutral' default value
--gt lt!ELEMENT rulebase (if)gt lt!ATTLIST
rulebase label CDATA IMPLIEDgt lt!ATTLIST rulebase
direction (forward backward bidirectional)
"bidirectional"gt lt!-- 'if' implications are
usable as top-level rules --gt lt!-- 'if' element
uses a conclusion followed by a premise --gt lt!--
"ltifgtconc premlt/ifgt" stands for "conc if prem",
i.e., "conc is true if prem is true" --gt lt!--
label attribute is a handle for the rule for
various uses, including editing --gt lt!ELEMENT
if (conc, prem)gt lt!ATTLIST if label CDATA
IMPLIEDgt
16
A Relational Languageruleml-datalog.dtd (III)
lt!datalog-and-hornlog.module lt!-- an 'and'
is usable within premises --gt lt!-- 'and' uses
zero or more atomic formulas --gt lt!--
"ltandgtatomlt/andgt" is equivalent to "atom"--gt lt!--
"ltandgtlt/andgt" is equivalent to "true"--gt
lt!ELEMENT and (atom)gt gt
lt!datalog.module lt!-- atomic formulas are
usable within conc's, prem's, and 'and's --gt lt!--
atom element uses rel(ation) symbol followed by a
sequence of --gt lt!-- zero or more arguments,
which may be ind(ividual)s or var(iable)s --gt
lt!ELEMENT atom (rel, (ind var))gt gt
17
A Relational Languageruleml-datalog.dtd (IV)
lt!-- there is one kind of fixed argument --gt
lt!-- individual constant, as in predicate logic
--gt lt!ELEMENT ind (PCDATA)gt lt!-- there is
one kind of variable argument --gt lt!-- logical
variable, as in logic programming --gt lt!ELEMENT
var (PCDATA)gt lt!-- there are only fixed
(first-order) relations --gt lt!-- relation or
predicate symbol --gt lt!ELEMENT rel (PCDATA)gt
18
A URL/URI Languageruleml-ur.dtd
lt!-- An XML DTD for a 'UR' RuleML Sublanguage
--gt lt!-- Last Modification 2001-01-23 --gt
lt!-- ENTITY Declarations --gt lt!-- a Uniform
Resource Identifier is currently PCDATA, but see
W3C's RFC2396 --gt lt!ENTITY URI "PCDATA"gt
lt!-- ELEMENT and ATTLIST Declarations --gt lt!--
there is an additional kind of fixed argument
--gt lt!-- objects (resources) use a URL/URI as
their OID, as in SHOE or RDF (cf. URML) --gt lt!--
however, unlike for XHTML anchors etc. URI used
as content, not as attribute --gt lt!-- 'label'
attribute, unlike URI, need not be unique
--gt lt!-- if no 'label' attribute is given,
browser must highlight the URI itself
--gt lt!ELEMENT ur (URI)gt lt!ATTLIST ur label
CDATA IMPLIEDgt
19
The Join Languageruleml-urdatalog.dtd
lt!-- An XML DTD for a 'UR' Datalog RuleML
Sublanguage --gt lt!-- Last Modification
2001-01-23 --gt lt!-- ENTITY Declarations --gt
lt!ENTITY urdatalog.module "INCLUDE"gt lt!ENTITY
datalog.module "IGNORE"gt lt!ENTITY datalog
SYSTEM "ruleml-datalog.dtd"gt datalog lt!ENTITY
ur SYSTEM "ruleml-ur.dtd"gt ur lt!--
ELEMENT and ATTLIST Declarations --gt
lt!urdatalog.module lt!-- atomic formulas
are usable within conc's, prem's, and 'and's
--gt lt!-- atom element uses rel name followed by
three kinds of arguments --gt lt!ELEMENT atom
(rel, (ur ind var))gt gt
20
Conclusions

• RuleML DTD 0.7, a system of 12 DTDs, is available
  at http//www.dfki.de/ruleml/indtd.html
• Sample files -- each referring to the most
  specific DTD still validating them -- are at
  http//www.dfki.de/ruleml/exa
• Further rule categories (e.g. ICs and triggers)
  and DTD updates will be available via main RuleML
  page at http//www.dfki.de/ruleml
• Distributed KR can already be based on current
  DTDs -- using (XSLT) transformations to reach
  follow-up and Participants DTDs