CHANGES (Examiners' Edition -> Special Edition)
===============================================

In the Special Edition, the term "incomparable" is retained throughout the
thesis.


Title-Page
----------

"Comparative\\Code Understanding\\of Information Models"
->
"Comparative\\Code Understanding\\of Information Models (Special Edition)"


Acknowledgements
----------------

Delete "(1994--1997)"

Delete "(1997--2001)"


"Monica Seles, Jelena Doki\'{c}, Iva Majoli, Karina Hab\v{s}udov\'{a}
 and Daniela Hantuchov\'{a}"
->
"Monica Seles, Daniela Hantuchov\'{a}, Jelena Doki\'{c}, Iva Majoli,
 Karina Hab\v{s}udov\'{a} and Iroda Tulyaganova"


Glossary
--------

New entry:
\item[completeness] A correspondence heuristic (or a decision procedure in
general) is \emph{complete} if it always says that two objects correspond when
they do (i.e. no false negatives) (Section~\ref{sec:4.1.1}).


\item[frame-class model]: "(Appendix A)" -> "(Appendix D)"


\item[plausible correspondence]: "can whether" -> "can decide whether"


New entry:
\item[soundness] A correspondence heuristic (or a decision procedure in
general) is \emph{sound} if it never says that two objects correspond when
they do not (i.e. no false positives) (Section~\ref{sec:4.1.1}).


Section 1.0
-----------

"`indeterminate' as the value of nationality"
->
"`indeterminate' as the value of \texttt{nationality}"


Section 1.2
-----------

"at least at one" -> "at least one" (delete-bar)


Section 1.4.2.0
---------------

End list with "." instead of ";"


Section 1.5.3
-------------

"from the code-fragments covered by the corresponding HLCs"
->
"between the code-fragments covered by the corresponding HLCs"


Section 1.7
-----------

"semantic equivalence/strength cases" -> "semantic-equivalence/strength cases"


Section 2.2.3
-------------

"semantic equivalence/strength cases" -> "semantic-equivalence/strength cases"


Section 2.5
-----------

Table 2.1: Context Interchange: Section 2.4.3 -> 2.4.4


Section 3.2.1
-------------

Footnote 5: "semantic equivalence/strength cases"
         -> "semantic-equivalence/strength cases"


Section 3.2.2
-------------

"(e.g. both a+b and b+a can be represented as a bag ({a, b})"
->
"(e.g. both a+b and b+a can be represented as a bag {a, b})"


"([Gra97])" -> "[Gra97]"


Section 3.3
-----------

In the CCUS output (p.104), the instance-numbers for entities and for explicit
attributes don't agree with their instance-numbers in the rest of the chapter,
and must therefore be changed as follows:
"entity_type#188" -> "entity_type#189"
"entity_type#187" -> "entity_type#188"
"explicit_attribute#299" -> "explicit_attribute#293"
"explicit_attribute#298" -> "explicit_attribute#292"
"explicit_attribute#297" -> "explicit_attribute#291"
"explicit_attribute#296" -> "explicit_attribute#290"


Add the following lines to the CCUS output:
>>>
.....Correspondence between integer_literal#97 and integer_literal#98
     confirmed.
....Correspondence between comparator#8 and comparator#9 confirmed.
...Correspondence between domain_rule#3 and domain_rule#4 confirmed.
<<<


"model_frame#4 (test)" -> "model_frame#4"


End list at top of p.106 with "." instead of ";"


Section 3.4.1
-------------

"every element in the aggregation from the case must match
 every element in the aggregation from the retrieval prompt"
->
"every element in the aggregation from the case must be matched by
 an element in the aggregation from the retrieval prompt"


Section 3.4.2, Footnote 14
--------------------------

"If the CCUS has to match an ordered aggregation to an unordered one"
->
"If the CCUS has to compare an ordered aggregation with an unordered one"


Section 4.1.1
-------------

"the semantics of EXPRESS does not" -> "the semantics of EXPRESS do not"


Section 4.1.2
-------------

End list with "." instead of ";"


"it also enables the CCUS to avoid repeating the comparison of a given pair of
 frames, since it first checks whether they already correspond."
->
"it also enables the CCUS to avoid repeating the comparison of a given pair of
 frames, since it first checks whether they already correspond, or have
 already been compared but failed to correspond.\footnote{The CCUS frame-
 comparison algorithm maintains a hash-table for this purpose: each key is a
 frame from the first model, and its associated value is the set of frames
 from the second model with which it has been compared but failed to
 correspond.}"


Section 4.2.0
-------------

"Pseudocode algorithms for inferring correspondences are given in Appendix C,
 Section C.3."


Section 4.2.2
-------------

"frame-class thesaurus" -> "frame-class-thesaurus"


Section 4.2.5
-------------

On p.140, "literal#1" -> "integer_literal#1",
          "literal#2" -> "integer_literal#2".


"equivalence/\-strength" -> "equivalence/\linebreak strength"
(i.e. "/-" -> "/" at end of line)


Section 4.3
-----------

"a case is indexed by the top-level frame in its problem part"
->
"a case is indexed by the frame-class of the top-level frame in its problem
 part"


"Section E.3" -> "Section E.3.3"


Section 5.2.2.1
---------------

Move linebreak after "OPTIONAL {masculine," to after the second "{masculine,"


"federated schemata" -> "federated database schemata"


"c.f." -> "cf." (twice)


"dependencies" -> "dependences"


Section 5.2.2.4
---------------

"Zaremski and Wing" -> "Moormann Zaremski and Wing" (twice)


Section 5.2.3.0
---------------

"$\phi$ is stronger than $\psi$ if $\phi$ is true whenever $\psi$ is true"
->
"$\phi$ is stronger than $\psi$ if $\psi$ is true whenever $\phi$ is true"


Section 5.3: "Semantic-Equivalence/Strength Cases"
----------- -> "Semantic-Equivalence/Sttrength Cases"

N.B. The indentation of the first paragraph is a bug with changebars.


"because that constraint already holds on the target schema"
->
"because that constraint already holds on the source schema"


"A SE/S case" -> "An SE/S case"


"." instead of ";" at the end of Case 4.


Figure 5.8: "semantic equivalence/strength case"
         -> "semantic-equivalence/strength case"


"no SE/S case matches both frames" -> "no SE/S case is matched by both frames"


Insert new paragraph before "The other SE/S cases implemented are..." (p.186):
"Pseudocode algorithms for using SE/S cases are given in Appendix C,
 Section C.4."


Figure 5.10: "semantic equivalence/strength case"
          -> "semantic-equivalence/strength case"

             "2000:11:22 12:41" -> "2002:05:08 18:50"

             INSTANCE semantic_strength_slot#2:
                "influence = semantic_strength_influence(positive);"
             -> "influence = semantic_strength_influence(negative);"


Section 6.3.0
-------------

New paragraph at the end:
"Pseudocode algorithms for the HLC CCUS are given in Appendix C, Section C.5."


Section 6.3.1
-------------

Footnote 12: "an semantic equivalence/strength case"
          -> "a semantic-equivalence/strength case"


Figure 6.6, "INSTANCE range#2":
  "extracted_from = {bound#3, bound#4}"
  ->
  "extracted_from = {interval#1}"


"navigation cases (Section 6.3.1)" -> "navigation cases"
(i.e. remove self-reference)


"\texttt{together}" -> "\emph{together}"


Section 6.3.2
-------------

"previous section" -> "previous subsection"


Section 7.2.2
-------------

"compares h1 with r2', and compares r1' with h2"
->
"compares h1 with r2', and compares `r1 with h2"


Section 7.2.3
-------------

"compares h1 with r2', and compares r1' with h2"
->
"compares h1 with r2', and compares `r1 with h2"


Section 8.2.0
-------------

"extraction cases (Appendices A)" -> "extraction cases (Appendices A and B)"


Section 8.2.1
-------------

"(by the fact that a word is in multiple subsets)"
->
"(by the fact that a word is in multiple synsets)"


"Zaremski and Wing" -> "Moormann Zaremski and Wing"


Section 8.2.2
-------------

"corresponds to (age >= 1) AND (age < 101) correspond"
->
"corresponds to (age >= 1) AND (age < 101)" (delete-bar)


"As noted in Section 8.2.2" -> "As noted in Section 8.2.1"


Section 8.3.1.3
---------------

"frame-comparison);" -> frame-comparison;" (delete-bar)


"extraction case for adder" -> "extraction case for \texttt{adder}"


Section 8.3.2.0
---------------

"This section" -> "This subsection"


Section 8.3.2.1
---------------

"if-statements^7," -> "if-statements,^7" (footnote-reference after comma)


Section A.2
-----------

"(Section E.1.8" -> "(Section E.1.8)"


"3 <= SELF <= 6" -> "{3 <= SELF <= 6}"


Section A.3
-----------

"Sections E.1.9" -> "Section E.1.9" (delete-bar)


"sum z" -> "sum $z$" (i.e. a mathematical expression, in italics)


"product z" -> "product $z$"


Section A.5
-----------

"(Figure A.7" -> "(Figure A.7)"


"Inverse attributes were not always present in the EXPRESS language, and some
 real models (such as the Electronic Design Interchange Format information
 model \cite{lau}) express this constraint as a domain rule instead
 (Figure~\ref{fig:A-8}):"
->
"Inverse attributes were not always present in the EXPRESS language.
 Furthermore, some real models (such as the Electronic Design Interchange
 Format information model \cite{lau}) choose to express this constraint as a
 domain rule instead. Figure~\ref{fig:A-8} shows how an inverse attribute
 can be simulated by a domain rule:"


Section A.6
-----------

"there_exists (Section E.1.23)" -> "there_exists (Section E.1.24)"


"QUERY (x <* s | p(x)" -> "QUERY (x <* s | p(x))"


"this set has to be greater than or equal to one"
->
"the cardinality of this set has to be greater than or equal to one"


Section A.7
-----------

"the first model" -> "the first model in Figure A.10"


Section B.0
-----------

"semantic equivalence/strength cases" -> "semantic-equivalence/strength cases"

Section B.2
-----------

"semantic equivalence/strength cases" -> "semantic-equivalence/strength cases"


Section C.0
-----------

"This appendix presents pseudocode algorithms for the main comparison
algorithm of the Comparative Code-Understanding System (CCUS) (Section C.1)
and its aggregation-matching policies (Section C.2). Shortage of time meant
that the pseudocode algorithms pertaining to inferring correspondences
(Chapter 4), assessing semantic equivalence and semantic strength (Chapter 5),
and the Higher-Level-Constraint-extracting CCUS (Chapter 6) had to be left
out."
->
"This appendix presents pseudocode algorithms for:
* the main comparison algorithm of the Comparative Code-Understanding System
  (CCUS) (Section C.1);
* its aggregation-matching policies (Section C.2);
* inferring correspondences (Section C.3);
* assessing semantic equivalence and semantic strength (Section C.4);
* the Higher-Level-Constraint-extracting CCUS (HLC CCUS) (Section C.5)."


New sentence:
"This appendix does not include details of the Code-Understanding System
 (Appendix B); see Appendix D of [Bro99] for algorithmic details), nor of the
 underlying frame system (Section 3.2.1)."


Section C.1.0
-------------

"(which compares two frames)"
->
"(which compares two frames -- Section C.1.1)"


"(which compares two slot-fillers)"
->
"(which compares two slot-fillers -- Section C.1.2)"


New sentence:
"The method \texttt{processDifference} (Section C.1.3) provides the hook to
 assessing semantic equivalence (Section C.4) and to the HLC CCUS
 (Section C.5)."


New paragraph:
"The CCUS starts by calling \texttt{matchFrames} on the two top-level
 \texttt{model\_frame}s."


Section C.1.1
-------------

Add intro:
"This method compares two frames, on the precondition that they correspond or
plausibly correspond. It calls \texttt{matchFillers} (Section C.1.2),
\texttt{processDifference} (Section C.1.3), \texttt{recordCorrespondence}
(Section C.3.4) and \texttt{retractCorrespondence} (Section C.3.5)."
"This method is responsible for inferring abstract correspondences back at the
code level when two HLCs are inferred to correspond (Section 6.3.3) -- it
records correspondence between their fillers of the \texttt{covers} slot."


"If the frames have been matched" -> If the frames have been compared"


"are in each other's corresponds_to slots" -> "correspond (see Section 4.1.2)"


Remove ";"


after
  if frame1 and frame2 correspond (cf. Section 4.1.2)
  then return true
insert:
  if frame1 and frame2 have already been compared but failed to correspond
  then return false


"record correspondence between frame1 and frame2
  (i.e. add them to each other's corresponds_to slots)"
->
"recordCorrespondence(frame1, frame2, mode)"


"retract the correspondence between frame1 and frame2
 (i.e. remove them from each other's corresponds_to slots)"
->
"retractCorrespondence(frame1, frame2)"


Remove ";"


"}" -> "} // end for"


Insert at the end of the algorithm:
"if frame1's "covers" slot is filled and frame2's "covers" slot is filled
 then recordCorrespondence(frame1."covers", frame2."covers", mode)
      if they have not already been recorded to correspond"


Section C.1.2
-------------

New paragraph:
"This method compares two slot-fillers, or the underlying values of defined-
 type-instances, or elements of aggregations (vectors). It calls
 \texttt{sanctionCorrespondence} (Section C.3.1), \texttt{matchFrames}
 (Section C.1.1) and \texttt{matchVectors} (Section C.2.1)."


"slot filler" -> "slot-filler"


"defined-type instances" -> "defined-type-instances"


Section C.1.3: "processDifference: Algorithm to process a difference"
-------------

New subsection!


Section C.2.0
-------------

"matchVectors method" -> "matchVectors method (Section C.2.1)"


"matchUnorderedVectors or matchUnorderedVectors"
->
"matchUnorderedVectors (Section C.2.2) or matchOrderedVectors (Section C.2.3)"


Section C.2.1
-------------

"vector 2" -> "vector2"


Add return-type ": boolean"


Section C.2.2
-------------

"algorithm" -> "method"


New sentence: "It calls \texttt{matchFillers} (Section C.1.2)."


"vector 2" -> "vector2"


Add return-type ": boolean"


"then matched:= true
remove element2 from vector2clone"
->
"then {
   matched:= true
   remove element2 from vector2clone
}"


"}" -> "} // end for" (twice)


Section C.2.3
-------------

"algorithm" -> "method"


New sentence: "It calls \texttt{matchFillers} (Section C.1.2)."


"vector 2" -> "vector2"

Add return-type ": boolean"


Section C.3: "Inferring Correspondences"
-----------

New section!


Section C.4: "Assessing Semantic Equivalence and Semantic Strength"
-----------

New section!


Section C.5: "Combining Comparison and Extraction"
-----------

New section!


Section D.1
-----------

Comment for "TYPE simple_type" duplicated for "TYPE indeterminate_type".
Replace with:
  (********************************************************************)
  (* TYPE indeterminate_type                                          *)
  (********************************************************************)
  (* The type of indeterminate aggregation-bounds (`?').              *)
  (********************************************************************)


Section D.3
-----------

"(e.g. age >= 0, (age >= 0) AND (age <= 130)"
->
"(e.g. age >= 0, (age >= 0) AND (age <= 130))"


Section E.0
-----------

"semantic equivalence/strength cases" -> "semantic-equivalence/strength cases"


Section E.3: "Semantic-Equivalence/Strength Cases"
----------- -> "Semantic-Equivalence/Sttrength Cases"

N.B. The indentation of the first paragraph is a bug with changebars.


Section E.3.10
--------------

"2000:11:22 12:41" -> "2002:05:08 18:50"

INSTANCE semantic_strength_slot#2:
"influence = semantic_strength_influence(positive);"
->
"influence = semantic_strength_influence(negative);"


Section E.5
-----------

"textual differences" -> "textual descriptions"


"``explicit_attribute#293''" -> "explicit_attribute#293"


"comparator#8''" -> "comparator#8"


"1999:11:19 15:09" -> "2002:05:11 19:56"

"bound of ?limit on ?variable" -> "bound of ?limit on ?variable"


Bibliography
------------

[AG96] "http://symphony.clarku.edu/atkinson/papers/icse-96.ps (last modified"
       -> "http://www.cse.scu.edu/~atkinson/papers/icse-96.pdf (modified"


[AP94] "last modified" -> "modified"


[Bai95] "last modified" -> "modified"


[BF99] "last modified" -> "modified"


[BF00] "last modified" -> "modified"


[Bro] "last modified 25th September 2001" -> "modified 21st September 2002"


[Bro99] "last modified 16th June 1999" -> "modified 15th May 2002"


[CC98] "last modified" -> "modified"


[CLM95a] "October 1995" -> "17th-20th October 1995"
         Delete URL (404)


[CLM95b] Delete URL (404)


[Cog] "(accessed on 23rd May 2001)" -> "(accessed 9th October 2002)"


[CTL97] "(13th June 1999)" -> "(modified 13th June 1999)


[Ede97] "1997" -> "2(1), Spring 1997"
        "http://www.db2mag.com/9701edel.htm (last modified 23rd April 1997)"
        -> "http://www.db2mag.com/db_area/archives/1997/q1/9701edel.shtml
            (accessed 9th October 2002)"


[GB99] "http://inchinn.infj.ulst.ac.uk/htdocs/pgomes.pdf"
       -> "http://eden.dei.uc.pt/~pgomes/Documents/ijcai99ws.pdf
           (accessed 9th October 2002)"


[Hak98] " (accessed 9th October 2002)"


[LTV95] "last modified" -> "modified"


[Mil98] "last modified" -> "modified"


[Pro] Delete URL.


[Rat] "UML Modeling Language, Standard Software Notation: Resource Centre"
      -> "Unified Modeling Language Resource Center"
      "http://www.rational.com/uml/index.shtml (modified 18th November 1998)"
      -> "http://www.rational.com/uml/ (accessed 9th October 2002)"


[RG98] "Golgolla" -> "Gogolla"


[Spi] "P. Spiby. EXPRESS Information modelling language: ISO WD 10303-11
       (edition 2 home page). EuroSTEP. http://www.eurpc2.demon.co.uk/
       part11.htm (last modified 10th July 1998)."
->
[SS99] "P. Spiby and D. Sanderson. Industrial Automation Systems and
        Integration -- Product Data Representation and Exchange -- Part 11:
        Description Methods: The EXPRESS Language Reference Manual,
        Second Edition (ISO TC184/SC4/WG11-N081), 1999. http://www.nist.gov/
        sc4/wg_qc/wg11/n081/n81.pdf (modified 8th October 1999)."


[Sun] Delete URL.


[VHG] "last modified" -> "modified"


[xml97] force capitals on xml {XML};
        "December 1997 1997" -> "December 1997"
        "9th December 1997" -> "8th September 2000"


[ZW97] "A.M. Zaremski" -> "A. Moormann Zaremski"