1. Ahn Ilsoo, Snodgrass Richard T. Partitioned Storage Structures for Temporal Databases. Information Systems. 1988;13.
The partitioning referred to in the title of this paper is akin to our idea of horizontal decomposition. As the title suggests, however, the emphasis in the paper is more on the use of such decomposition as a technique for physical design. (To quote reference [108]: “Temporal partitioning is … in the domain of physical design.”)
2. Allen James F. Maintaining Knowledge about Temporal Intervals. CACM. 1983;26 November.
The source of Allen’s operators (see Chapter 7). For purposes of reference, we give below a complete list of the operators from Allen’s paper, together with our own and SQL’s analogs of those operators. Note: The “i” in ALLEN’s operators mi, oi, di, si, and fi stands for inverse. (The m, o, and d operators in particular—unlike our analogs of those operators—aren’t commutative; thus, if, e.g., m is pronounced meets, then mi might reasonably be pronounced is met by, or just met by.) Observe that Allen has no direct analog of our MERGES operator, nor of “⊂” or “⊃”, nor of the negated forms (NOT OVERLAPS, etc.) of any of our other operators.
Aside: Reference [76] gives the following somewhat more extensive list of operators and helpful explanatory diagram:
We leave it as an exercise to interpret these operators in terms of the operators defined in Chapter 7. End of aside.
3. Al-Taha Khaled K, Snodgrass Richard T, Soo Michael D. Bibliography on Spatiotemporal Databases. ACM SIGMOD Record. 1993;22 March.
Here’s a lightly edited extract from the preamble to this bibliography: “Only recently have issues concerning the simultaneous support of both space and time in databases been considered. While this new area of research was presaged by Thrift’s observation in 1977 that time could be considered to be an additional dimension in a two- or three-dimensional space [Thrift 1977], little was done until Nikos Lorentzos’s and Gail Langran’s doctoral dissertations [Langran 1989A, Lorentzos 1988].” To elaborate briefly: Reference [Thrift 1977] is N. Thrift: “An Introduction to Time Geography,” in Concepts in Modern Geography 13, Geo-Abstracts Ltd., London, U.K. (1977); reference [Langran 1989A] is G. Langran: “Time in Geographic Information Systems,” Ph.D. Dissertation, University of Washington (1989); reference [Lorentzos 1988] is our reference [67].) It’s particularly interesting to see that the authors of this bibliography are here on record as agreeing that “time is just another dimension”—a position we agree with, though we haven’t used the term dimension (in this context) much in the present book.
4. Ariav Gad. A Temporally Oriented Data Model. ACM TODS. December 1986;11.
The approach described in this paper differs from just about every other in the literature (with the possible exception of reference [17]). In effect, it timestamps entire relations, instead of individual attributes—more precisely, individual attribute values—or individual tuples (see the annotation to reference [18]). An algebra and an SQL extension are defined. However, the algebra is incomplete (only three operators are defined), and the author states explicitly that further work is needed.
5. Ben-Zvi J. The Time Relational Model Computer Science Department, University of California at Los Angeles 1982; Ph.D. Dissertation.
Ben-Zvi was one of the earliest workers in the temporal database field. A useful brief summary of his ideas and contributions can be found in a short paper by Gadia [59].
6. Claudio Bettini, Curtis E. Dyreson, William S. Evans, Richard T. Snodgrass, and X. Sean Wang: “A Glossary of Time Granularity Concepts,” in reference [55].
This paper is an attempt to inject some rigor into the granularity issue. It’s intended as a formal addendum to reference [62].
7. Böhlen Michael H. Temporal Database System Implementations. ACM SIGMOD Record. December 1995;24.
A survey of 13 implemented systems, with a high level analysis of system features supported (or not, as the case may be). The following observations regarding the systems examined are worthy of note: “Transaction time support is negligible … The focus is on queries. Updates, rules, and integrity constraints are neglected … Temporal database design has barely found its way into products.” (Of course, the paper is rather old by now, and this last sentence is no longer as true as it used to be.)
8. Böhlen Michael H, Jensen Christian S, Snodgrass Richard T. Temporal Statement Modifiers. ACM TODS. December 2000;25.
Among other things this paper defines ATSQL, an SQL-based temporal language that’s similar in spirit to TSQL2 [105-106].
9. Böhlen Michael H, Snodgrass Richard T, Soo Michael D. Coalescing in Temporal Databases. In: Proc. 22nd Int. Conf. on Very Large Data Bases, Mumbai (Bombay), India. September 1996.
The operation of coalescing is analogous, more or less, to our PACK operation. This paper presents a set of algebraic transformation rules for expressions involving such coalescing and investigates approaches to the implementation of such expressions.
As an aside, we note that the PACK operator and the related UNPACK operator are known by a variety of names in the literature.1 Reference [98] calls them CONTRACT and EXPAND, respectively. The previous edition of reference [37] called them COALESCE and UNFOLD. PACK has also been called COMPRESS [94]. And reference [70] (as well as several others by Lorentzos and various coauthors) defines a pair of operators called FOLD and UNFOLD, but those operators aren’t quite the same as our PACK and UNPACK; loosely, FOLD “packs” a set of points into a set of intervals, and UNFOLD “unpacks” a set of intervals into a set of points (the same paper uses NORMALIZE—originally spelled NORMALISE—for what we call PACK). Finally, there’s at least one paper in the literature (namely, reference [95]) that uses PACK and UNPACK to mean something quite different: namely, operators that are akin, somewhat, to the regular GROUP and UNGROUP operators of relational algebra.
10. Casanova Marco A, Fagin Ronald, Papadimitriou Christos H. Inclusion Dependencies and Their Interaction with Functional Dependencies. In: Proc. 1st ACM SIGACT-SIGMOD Symposium on Principles of Database Systems, Los Angeles, Calif. March 1982.
The origin of the concept of inclusion dependencies (INDs). The paper provides a sound and complete set of inference rules for such dependencies.
11. Cattell RGG, Barry Douglas K, eds. The Object Data Standard: ODMG 3.0. San Francisco, Calif.: Morgan Kaufmann; 2000.
12. Clifford James. A Model for Historical Databases. In: Proc. Workshop on Logical Bases for Data Bases, Toulouse, France. December 1982.
See reference [13], also the annotation to reference [18].
13. James Clifford and Albert Croker: “The Historical Relational Data Model (HRDM) Revisited,” in reference [110].
Revises and extends the proposals of reference [12].
14. Clifford James, Croker Albert, Tuzhilin Alexander. On Completeness of Historical Relational Query Languages. ACM TODS. March 1994;19.
A quote: “In this paper we define temporally grouped and temporally ungrouped historical data models and propose two notions of historical relational completeness, analogous to Codd’s notion of relational completeness, one for each type of model” (italics in the original). The paper additionally considers to what extent various approaches described in the literature satisfy the proposed completeness criteria. The approaches in question are HRDM [13], Gadia’s approach [58], TQuel [104], and Lorentzos’s approach [68].
15. J. Clifford and A. Rao: “A Simple General Structure for Temporal Domains,” in reference [96].
Proposes a formalism for defining a variety of temporal point types, associated interval types, and operators on values of those types.
16. Clifford J, Tuzhilin A, eds. Recent Advances in Temporal Databases. New York, N.Y.: Springer-Verlag; 1995; (Proc. Int. Workshop on Temporal Databases, Zurich, Switzerland, September 17th-18th, 1995).
17. Clifford James, Warren David S. Formal Semantics for Time in Databases. ACM TODS. June 1983;8.
Describes an approach based on the idea that any given relvar (our term) can be thought of as a sequence of timestamped relations, where the timestamps in question correspond to logged time points (not intervals) as discussed in Chapter 17. The major contribution of the paper is a formal definition of the approach in terms of intensional logic.
18. Clifford James, Tansel Abdullah Uz. On An Algebra for Historical Relational Databases: Two Views. In: Proc. ACM SIGMOD Int. Conf. on Management of Data, Austin, Tex. May 1985.
In the first part of this paper, Clifford proposes an algebra for dealing with “temporal relations” in which attribute values can be relations and the tuples within those inner relations can be timestamped with time points representing stated (i.e., valid) time. In the second part, Tansel proposes another algebra in which the timestamps are intervals rather than points (see also references [109] and [111]). Note: Timestamping tuples within values of relation valued attributes is known, loosely, as “timestamping attributes”; it’s contrasted with “timestamping tuples,” which refers to the idea of timestamping tuples in the containing instead of the contained relation, an idea that seems to have originated in reference [12]. The approach we advocate in the present book is somewhat akin to “timestamping tuples,” except that the “timestamps” in question are actually part of the tuples in question; in other words, they’re represented by means of relational attributes in the usual way. Also, of course, a given tuple in our approach can contain any number of such “timestamps.”
19. Clifford James, Dyreson Curtis, Isakowitz Tomás, Jensen Christian S, Snodgrass Richard T. On the Semantics of ‘Now’ in Databases. ACM TODS. June 1997;22.
Presents arguments in favor of allowing database tuples to contain the variable referred to in Chapter 12 as “the NOW marker” (and several other similar variables as well). To quote: “This article … provides a formal basis for defining the semantics of databases with variables … [It] demonstrates that existing variables, such as now and until changed, are indispensable [sic] in databases. It also … introduces new now-relative and now-relative indeterminate [variables].” Note: The term variable here doesn’t refer to our relation variables (relational databases always contain variables of that kind, of course); rather, it refers to “the NOW marker” and other similar constructs.
20. Codd EF. Derivability, Redundancy, and Consistency of Relations Stored in Large Data Banks. IBM Research Report RJ599 (August 19th, 1969). Republished in ACM SIGMOD Record. March 2009;38.
Codd’s very first published paper on the relational model.
21. Codd EF. A Relational Model of Data for Large Shared Data Banks. CACM. June 1970;13 Republished in Milestones of Research—Selected Papers 1958-1982 (CACM 25th Anniversary Issue), CACM 26, No. 1 (January 1983) and elsewhere.
A revised and extended version of reference [20]. The major technical difference between the two is that reference [20] permits relation valued attributes while reference [21] prohibits them.
22. Codd EF. Relational Completeness of Data Base Sublanguages. In: Rustin Randall, ed. Data Base Systems: Courant Computer Science Symposia 6. Englewood Cliffs, N.J.: Prentice-Hall; 1972.
23. Codd EF, Date CJ. Much Ado about Nothing. In: Date CJ, ed. Relational Database Writings 1991-1994. Reading, Mass.: Addison-Wesley; 1995.
Codd was probably the best known advocate of nulls and many-valued logics as a basis for dealing with missing information. This article contains the text of a debate between Codd and one of the present authors on the subject.
24. Darwen Hugh. Into the Unknown. (writing as Andrew Warden) In: Date CJ, ed. Relational Database Writings 1985-1989. Reading, Mass.: Addison-Wesley; 1990.
25. Darwen Hugh. The Nullologist in Relationland. In: Date CJ, Darwen Hugh, eds. Relational Database Writings 1989-1991. Reading, Mass.: Addison-Wesley; 1992.
26. Hugh Darwen: “Valid Time and Transaction Time Proposals: Language Design Aspects,” in reference [55].
Among other things, this paper advocates the language design principle known as syntactic substitution: “A language definition should start with a few judiciously chosen primitive operators … Subsequent development is [done], where possible, by defining new operators in terms of … previously defined [ones]. Most importantly, syntactic substitution does not refer to an imprecise principle such as might be expressed as “A is something like, possibly very like, B,” where A is some proposed new syntax and B is some expression using previously defined operators. If A is close in meaning to B but cannot be specified by true syntactic substitution, then we have a situation that is disagreeable and probably unacceptable, in stark contrast to true syntactic substitution, which can be very agreeable and acceptable indeed.”
27. Darwen Hugh. The Decomposition Approach. In: Date CJ, Darwen Hugh, eds. Database Explorations: Essays on The Third Manifesto and Related Topics. Bloomington, Ind.: Trafford Publishing; 2010; Note: This paper is based on an earlier presentation by Darwen titled “How to Handle Missing Information Without Using Nulls,” the slides for which can be found at www.thethirdmanifesto.com (May 9th, 2003; revised May 16th, 2005).
28. Darwen Hugh. An Introduction to Relational Database Theory Frederiksberg, Denmark: Ventus; 2012; Free download available at http://bookboon.com; 2012.
29. Darwen Hugh, Date CJ. An Overview and Analysis of Proposals Based on the TSQL2 Approach. In: Date CJ, ed. Date on Database: Writings 2000-2006. Berkeley, Calif.: Apress; 2006.
Of previously published proposals for dealing with the temporal database problem, TSQL2 [105-106] is probably the best known, and several other proposals, including in particular those of reference [107], have been based on it. This paper provides an overview and critical analysis of such proposals, comparing and contrasting them with the approach advocated in the present book.
30. Darwen Hugh, Date CJ, Fagin Ronald. A Normal Form for Preventing Redundant Tuples in Relational Databases. In: Proc. 15th Int. Conf. on Database Theory, Berlin, Germany. 2012; March 26th-29th.
31. Darwen Hugh, Sykes Mike. Concerns about the TSQL2 Approach to Temporal Databases May 1996; Kansas City, Mo. ftp://sqistandards.org/SC32/WG3/Meetings/MCI_1996_05_KansasCity_USA/mci071.ps; May 1996; (This is an internal working document, not generally available at the time of writing.).
A precursor to reference [32], criticizing among other things the notion of “temporal upward compatibility” advocated in reference [107] (also in reference [8]). Briefly, temporal upward compatibility means it should be possible to convert an existing nontemporal database into a temporal one by just “adding temporal support,” thereby allowing existing nontemporal applications to run unchanged against the now temporal database. See reference [29] for further discussion.
32. Darwen Hugh, Sykes Mike, et al. On Proposals for Valid-Time and Transaction-Time Support January 1997; Madrid, Spain ftp://sqlstandards.org/SC32/WG3/Meetings/MAD_1997_01_Madrid_ESP/mad146.ps; January 1997; (This is an internal working document, not generally available at the time of writing.).
A response to reference [107].
33. Date CJ. Dates and Times in IBM SQL: Some Technical Criticisms. In: Relational Database Writings 1985-1989. Reading, Mass: Addison-Wesley; 1990.
We include this reference merely to draw attention to the fact that the support in IBM’s DB2 product for datetime data types—not to be confused with its temporal support—differs significantly, in both syntax and semantics, from that of the SQL standard [61]. For a detailed tutorial on this latter, see reference [47].
34. Date CJ. NOT Is Not “Not”! (Notes on Three-Valued Logic and Related Matters). In: Relational Database Writings 1985-1989. Reading, Mass: Addison-Wesley; 1990.
35. Date CJ. A Note on Orthogonality. In: Relational Database Writings 1994-1997. Reading, Mass: Addison-Wesley; 1998.
36. Date CJ. Encapsulation Is a Red Herring. Database Programming & Design. September 1998;12.
37. Date CJ. An Introduction to Database Systems 8th edition Boston, Mass.: Addison-Wesley; 2004.
38. Date CJ. “On the Notion of Logical Difference”; “On the Logical Difference Between Model and Implementation”; and “On the Logical Differences Between Types, Values, and Variables,”. all In: Date on Database: Writings 2000-2006. Berkeley, Calif.: Apress; 2006.
39. Date CJ. The Theory of Bags: An Investigative Tutorial. In: Logic and Databases: The Roots of Relational Theory. Victoria, B.C.: Trafford Publishing; 2007.
40. Date CJ. Image Relations. In: Date CJ, Darwen Hugh, eds. Database Explorations: Essays on The Third Manifesto and Related Topics. Bloomington, Ind.: Trafford Publishing; 2010.
41. Date CJ. Inclusion Dependencies and Foreign Keys. In: Date CJ, Darwen Hugh, eds. Database Explorations: Essays on The Third Manifesto and Related Topics. Bloomington, Ind.: Trafford Publishing; 2010.
42. Date CJ. SQL and Relational Theory: How to Write Accurate SQL Code 2nd edition Sebastopol, Calif.: O’Reilly Media; 2012.
43. Date CJ. Database Design and Relational Theory: Normal Forms and All That Jazz Sebastopol, Calif.: O’Reilly Media; 2012.
44. Date CJ. View Updating and Relational Theory: Solving the View Update Problem Sebastopol, Calif.: O’Reilly Media; 2013.
45. Date CJ. Relational Theory for Computer Professionals: What Relational Databases Are Really All About Sebastopol, Calif.: O’Reilly Media; 2013.
46. C. J. Date: The New Relational Database Dictionary (to appear).
This reference is the third in a series. The first was titled simply The Relational Database Dictionary (O’Reilly, 2006). The second was titled The Relational Database Dictionary, Extended Edition (Apress, 2008).
47. Date CJ, Darwen Hugh. A Guide to the SQL Standard 4th edition Reading, Mass.: Addison-Wesley; 1997.
A comprehensive tutorial on SQL:1992, this book also contain details of the Call Level Interface feature CLI (added to the standard in 1995), the Persistent Stored Modules feature PSM (added in 1997), and a few features that became part of SQL: 1999. Note: Although the book is now quite old, most of what it has to say is still applicable to the current version of the standard, SQL:2011 [61].
48. Date CJ, Darwen Hugh. Multiple Assignment. In: Date CJ, ed. Date on Database: Writings 2000-2006. Berkeley, Calif.: Apress; 2006.
49. Date CJ, Darwen Hugh. The Third Manifesto. In: Database Explorations: Essays on The Third Manifesto and Related Topics. Bloomington, Ind.: Trafford Publishing; 2010.
See the annotation to reference [52].
50. Date CJ, Darwen Hugh. Tutorial D. In: Database Explorations: Essays on The Third Manifesto and Related Topics. Bloomington, Ind.: Trafford Publishing; 2010.
See the annotation to reference [52].
51. Date CJ, Darwen Hugh. The Inheritance Model and The Inheritance Model: What Was Changed and Why. In: Database Explorations: Essays on The Third Manifesto and Related Topics. Bloomington, Ind.: Trafford Publishing; 2010.
52. Date CJ, Darwen Hugh. Databases, Types, and the Relational Model: The Third Manifesto 3rd edition Boston, Mass.: Addison-Wesley; 2007.
This book introduces and explains The Third Manifesto, a precise though somewhat formal definition of the relational model and a supporting type theory (including a comprehensive model of type inheritance—see reference [51]). Note that the Manifesto as such constitutes just one comparatively short chapter in the book (twelve pages, out of a total of well over 500). The book also contains a definition of the language Tutorial D. (Please observe that Tutorial D isn’t part of the Manifesto as such but is merely one possible language that can be used to illustrate and explain the ideas of the Manifesto.) Note: Upgraded definitions of both the Manifesto itself and Tutorial D can be found in references [49] and [50], respectively. The most recent versions can be found at the website www.thethirdmanifesto.com, which also gives information regarding a variety of existing Tutorial D implementations, as well as other projects related to the Manifesto.
53. Date CJ, Darwen Hugh, Lorentzos Nikos A. Temporal Data and the Relational Model San Francisco, Calif.: Morgan Kaufmann; 2003.
The present book’s predecessor.
54. Christina Davies, Brian Lazell, Martin Hughes, and Leslie Cooper: “Time Is Just Another Attribute—Or at Least Another Dimension,” in reference [16].
To quote: “[The] flexibility and simplicity of the relational model are too valuable to be jettisoned without good reason … [A] much stronger case must be made against the unextended relational model before it is rejected or modified for temporal applications.” (We agree strongly with these remarks.) The paper presents three case studies and identifies some limitations of TSQL2 [105-106]. It also argues that if SQL is to be extended, then the extensions in question should not be limited to temporal intervals only.
55. Etzion Opher, Jajodia Sushil, Sripada Suryanarayana, eds. Temporal Databases: Research and Practice. New York, N.Y.: Springer-Verlag; 1998.
This book is an anthology giving “the state of the temporal database art” as of about 1997. It’s divided into four major parts, as follows:
1. Temporal Database Infrastructure
2. Temporal Query Languages
3. Advanced Applications of Temporal Databases
4. General Reference
Several of the references listed in this appendix—in particular reference [26]—are included in this book.
56. Fagin Ronald. Normal Forms and Relational Database Operators. In: Proc. 1979 ACM SIGMOD Int. Conf. on Management of Data, Boston, Mass. May/June 1979.
This paper, which is the one that introduced projection/join normal form (also known as 5NF), is the definitive reference on what might be called classical normalization theory.
57. Gadia Shashi K, Vaishnav Jay H. A Query Language for a Homogeneous Temporal Database. In: Proc. 4th ACM SIGACT-SIGMOD Symposium on Principles of Database Systems, Portland, Ore. March 1985.
Sketches a language for dealing with “temporal relations” in which, as in reference [18], attributes can be relation valued and tuples within those inner relations can be timestamped (see also reference [58]). The timestamps are sets of intervals in packed form, and they represent stated times, not logged times. The term homogeneous in the paper’s title refers to the restriction that—as the paper itself puts it—“the temporal domain within a tuple does not vary from one attribute to another … [As a consequence,] there are several weaknesses in the model. For example, … information about persons, their parents and dates of birth [can’t be represented in the same relation].”
58. Gadia Shashi K. A Homogeneous Relational Model and Query Languages for Temporal Databases. ACM TODS. 1988;13 December.
A more formal and complete treatment of the model underlying the proposals of reference [57].
59. Shashi K. Gadia: “Ben-Zvi’s Pioneering Work in Relational Temporal Databases,” in reference [110].
60. Hall Patrick, Owlett John, Todd Stephen. Relations and Entities. In: Nijssen GM, ed. Modelling in Data Base Management Systems. Amsterdam, Netherlands: North-Holland; 1976.
61. International Organization for Standardization (ISO). Database Language SQL 2011; Document ISO/IEC 9075:2011.
The current version, or edition, of the SQL standard (“SQL:2011”).
62. Christian S. Jensen and Curtis E. Dyreson (eds.): “The Consensus Glossary of Temporal Database Concepts—February 1998 Version,” in reference [55].
This glossary follows on from, and subsumes, two earlier ones which appeared in ACM SIGMOD Record 21, No. 3 (September 1992) and ACM SIGMOD Record 23, No. 1 (March 1994), respectively. We strongly agree with the article’s opening remarks: “A technical language is an important infrastructural component of any scientific community. To be effective, such a language should be well defined, intuitive, and agreed upon.” The authors then go on to give, not just their own preferred terms and definitions as such, but also discussions of alternative terms, justifications for their preferences, and so forth. Note, however, that several terms and concepts discussed in the present book seem to have no counterpart in this glossary, the following among them:
■ Point types
■ Ordinal types
■ Cyclic types
■ Successor functions (NEXT_T), etc.
■ Interval types and the interval type generator
■ BEGIN, END, PRE, POST, and POINT FROM
■ Allen’s operators
■ Interval UNION, INTERSECT, and MINUS
■ EXPAND and COLLAPSE
■ Equivalence (for sets of intervals)
■ “The moving point now”
63. Jensen Christian S, Snodgrass Richard T, Soo Michael D. Extending Existing Dependency Theory to Temporal Databases. IEEE Transactions on Knowledge and Data Engineering. August 1996;8.
This paper is one of many to define “temporal” versions of certain familiar relational concepts: specifically (in the case at hand), temporal functional dependencies, temporal primary keys, and temporal normal forms. However, the definitions in question all rely on an assumption that timestamps aren’t represented by regular relational attributes.
64. Jones S, Mason PJ. Handling the Time Dimension in a Data Base. In: Proc. Int. Conf. on Data Bases, Aberdeen, Scotland. London, U.K.: Heyden & Son; 1980; (July 1980).
One of the very earliest proposals (possibly the earliest).
65. Kulkarni Krishna, Michels Jan-Eike. Temporal Features in SQL:2011. ACM SIGMOD Record. September 2012;41.
66. Levene Mark, Loizou George G. A Guided Tour of Relational Databases and Beyond London, U.K.: Springer-Verlag; 1999.
Includes one chapter (Chapter 7, “Temporal Relational Databases”) on temporal matters. Section 4 of that chapter (“A Historical Relational Algebra”) discusses Lorentzos’s work as described in references [68] and [70].
67. Lorentzos Nikos A. A Formal Extension of the Relational Model for the Representation and Manipulation of Generic Intervals U.K.: Birkbeck College, University of London; August 1988; Ph.D. Dissertation.
The approach to temporal data described and advocated in the present book has its origins in Lorentzos’s research as reported in his PhD dissertation (as well as in numerous subsequent publications by Lorentzos and other researchers; see, e.g., references [68], [70], and [76]). As far as we’re aware, Lorentzos’s dissertation was (a) the first to propose the interval abstraction as such; (b) the first, or one of the first, to propose a truly relational approach to the problem; and (c) the first to propose the idea that temporal support should be at least partly just a special case of support for intervals in general. (We note in passing that reference [89] effectively supports these claims.)
Note: We need to say something regarding the use of the term extension in the title of this reference. That term refers primarily to the introduction of:
a. A generic interval data type (or rather, as we would now say, an interval type generator), along with
b. A variety of new operators that can be applied to relations with attributes of some interval type.
As explained in Chapter 4, however, the question of which types and type generators are supported is orthogonal to the question of support for the relational model as such. And support for “new operators” is part of (and is implied by) support for “new types.” In other words, features a. and b. above can certainly be regarded as an extension to something, but they’re not an extension to the relational model. It’s this fact that enables us to claim, legitimately, that our approach to temporal database issues is a truly relational one and involves no changes or extensions to the classical relational model.
Remarks analogous to the foregoing apply equally to the titles of several other references in this appendix—for example, references [70] and [74]—and we won’t bother to repeat them every time, instead letting the foregoing paragraphs do duty for all.
68. Nikos A. Lorentzos and R. G. Johnson: “TRA: A Model for a Temporal Relational Algebra,” in reference [96].
This paper and reference [69], which overlap somewhat, were the first to discuss the FOLD and UNFOLD operators (predecessors of PACK and UNPACK, respectively). Significantly, the paper includes an example of a relation with two distinct “valid time” attributes, something the SQL standard wouldn’t permit. It also discusses what it calls “periodic” (i.e., cyclic) temporal data, such as hours of the day. Note: Despite the reference in the paper’s title to “a temporal relational algebra,” the authors are careful to state explicitly that they’re not proposing any “new elementary relational algebra operations.” In other words, their “temporal relational algebra” is just the classical relational algebra, enhanced (as in the present book) with certain useful shorthands.
69. Lorentzos Nikos A, Johnson RG. Extending Relational Algebra to Manipulate Temporal Data. Information Systems. 1988;13.
See the annotation to reference [68].
70. Nikos A. Lorentzos: “The Interval-Extended Relational Model and Its Application to Valid-Time Databases,” in reference [110].
71. Lorentzos Nikos A. DBMS Support for Time and Totally Ordered Compound Data Types. Information Systems. September 1992;17.
In this book we’ve discussed the nature of interval types at considerable length, but we’ve said comparatively little about the nature of the point types in terms of which those interval types are defined. This paper and reference [72] address this latter issue. In essence, they propose a new type generator that would allow users to define temporal point types that almost certainly wouldn’t be available as system defined types. An example might be time points measured in terms of week and day numbers (e.g., “week 5, day 2”). Another might be time points measured using a cesium-based atomic clock, in which the time unit is 1/9,192,631,770th of a second (such a point type is needed in certain scientific applications). Nontemporal point types (e.g., weights measured in stones, pounds, and ounces) can also be defined by means of the proposed type generator. Note: Alternative proposals for dealing with some of these matters—in particular, with the question of units and granularity—can be found in Chapter 18 of the present book, also in reference [52].
72. Lorentzos Nikos A. DBMS Support for Nonmetric Measurement Systems. IEEE Transactions on Knowledge and Data Engineering. 1994;6 December.
See the annotation to reference [71].
73. Lorentzos Nikos A, Darwen Hugh. Extension to SQL2 Binary Operations for Temporal Data. (invited paper) In: Proc. 3rd HERMIS Conf., Athens, Greece. 1996; September 26th-28th.
This paper extends the work described in reference [76]. It proposes the addition of analogs of certain of our “U_” operators to “SQL2” (i.e., the SQL:1992 standard); to be specific, it discusses what we would now call U_JOIN, U_UNION, U_INTERSECT, and U_MINUS, and gives examples in each case.
74. Lorentzos Nikos A, Johnson RG. An Extension of the Relational Model to Support Generic Intervals. In: Schmidt Joachim W, Ceri Stefano, Missikoff Michel, eds. Extending Database Technology. New York, N.Y.: Springer-Verlag; 1988.
This paper was the first to show that temporal data management is actually a special case of a more general problem: namely, that of supporting a “generic interval type” (what we would now call an interval type generator), together with operators to deal with relations that include interval data.
75. Lorentzos Nikos A, Kollias Vassiliki J. The Handling of Depth and Time Intervals in Soil Information Systems. Comp Geosci. 1989;15.
Defines a set of operators for dealing with data related to changes in the composition of soil with depth and the way those changes themselves change with time. Note: We include this reference here primarily as a concrete example of an application area in which at least some of the pertinent intervals are nontemporal ones specifically.
76. Lorentzos Nikos A, Mitsopoulos Yannis G. IXSQL: An Interval Extension to SQL. In: Proc. Int. Workshop on an Infrastructure for Temporal Databases, Arlington, Tex. 1993; June 14th-16th.
An informal presentation of the ideas discussed more formally in reference [79].
77. Lorentzos Nikos A, Manolopoulos Yannis. Efficient Management of 2-Dimensional Interval Relations. In: Karagiannis D, ed. Lecture Notes in Computer Science 856. Proc. 5th DEXA Int. Conf., Athens, Greece. New York, N.Y.: Springer-Verlag; September 1994.
The source of the SPLIT operator described in Appendix E of the present book (in fact, Appendix E is heavily based on ideas first presented in this paper). Note: The term 2-dimensional interval relation in the title of this paper refers to relations with two distinct interval attributes, not to relations with a single attribute whose values are two-dimensional intervals.
78. Lorentzos Nikos A, Manolopoulos Yannis. Functional Requirements for Historical and Interval Extensions to the Relational Model. Data and Knowledge Engineering 17 1995.
Identifies a number of criteria that support for interval (and therefore temporal) data should satisfy. Two broad approaches to the problem are described, “nested” and “nonnested.” The authors show that all of the approaches intended for temporal data in particular can in fact be used for interval data in general. They then proceed to evaluate all known approaches in terms of the criteria identified in the paper.
79. Lorentzos Nikos A, Mitsopoulos Yannis G. SQL Extension for Interval Data. IEEE Transactions on Knowledge and Data Engineering. May/June 1997;9.
Describes IXSQL, “an Interval Extension to SQL.” IXSQL is based on the operators of the conventional relational algebra, together with two additional operators, FOLD and UNFOLD (predecessors of PACK and UNPACK, respectively). As its name suggests, the language is explicitly intended for the management of interval data—though it’s important to note that in fact it supports both interval and noninterval data in the same uniform way. Since the paper also shows that temporal data is just a special case of interval data in general, it follows that IXSQL actually supports both temporal and nontemporal data in that same uniform fashion. See also reference [76] and, for a further extension of the IXSQL approach, reference [119].
80. Lorentzos Nikos A, Poulovassilis Alexandra, Small Carol. Implementation of Update Operators for Interval Relations. BCS Comp J. 1994;37.
Describes optimized implementation algorithms for operations on relations with two or more interval attributes (see Appendix E).
81. Lorentzos Nikos A, Poulovassilis Alexandra, Small Carol. Manipulation Operations for an Interval-Extended Relational Model. Data and Knowledge Engineering. 1995;17.
This paper is a more formal version of reference [80]. Complexity and simulation results are included.
82. Lorentzos Nikos A, Tryfona Nektaria, Viqueira Jose R Rios. Relational Algebra for Spatial Data Management. In: Proc. Int. Workshop on Integrated Spatial Databases: Digital Images and GIS, Portland, Maine. 1999; June 14th-16th.
This paper presents initial results from the research described in more detail in reference [84].
83. Lorentzos NA, Viqueira JRR. On a Spatio-Temporal Relational Model Based on Quanta. In: Koubarakis M, ed. Spatiotemporal Databases: The Chorochronos Approach—Lecture Notes in Computer Science 2520. New York, N.Y.: Springer-Verlag; 2003.
See reference [119].
84. Lorentzos NA, Viqueira JRR. Relational Formalism for the Management of Spatial Data. BCS Comp J. January 2006;49.
This paper defines spatial data types in terms of “spatial quanta” and further generalizes the FOLD and UNFOLD operators of IXSQL [79] to make them applicable to spatial data. Additional operators based on FOLD, UNFOLD, and existing relational operators are also defined, such that the complete set of operators thus made available fully supports the management of spatial data. The new operators are also applicable to conventional data and have considerable practical interest. See also references [84] and [119].
85. Lorentzos NA, Rios Viqueira JR, Tryfona N. Quantum-Based Spatial Extension to the Relational Model. In: Fotiadis Dimitrios I, Nikolopoulos Stavros D, eds. Advances in Informatics. Singapore: World Scientific Publishing; 2000.
86. Lorentzos NA, Yialouris CP, Sideridis AB. Time-Evolving Rule-Based Knowledge Bases. Data and Knowledge Engineering. 1999;29.
This paper applies the relational operators defined in reference [79] to support temporal knowledge, meaning knowledge that changes over time.
87. Lum V. Designing DBMS Support for the Temporal Dimension. In: Proc. ACM SIGMOD Int. Conf. on Management of Data, Boston, Mass. June 1984.
This early paper recognizes the need for supporting both logged times and stated times, but proposes that only logged times be given special treatment by the system—stated times are to be dealt with by means of regular attributes in the usual way. The paper gives numerous reasons as to why an implementation that doesn’t give special internal treatment to logged times won’t perform well. Special storage structures and access methods are proposed to address such problems.
88. Martin NG, Navathe SB, Ahmed R. Dealing with Temporal Schema Anomalies in History Databases. In: Proc. 13th Int. Conf. on Very Large Data Bases, Brighton, U.K. September 1987.
To quote from the abstract: “Because history databases do not discard data, they cannot discard outdated database schemas. Thus, in any proposal for a practical history database system, some method must be provided for accessing data [described by] outdated, yet historically valid, schemas.” Note: We touched on this kind of problem very briefly at the end of Chapter 14.
89. Matchett Robert. Temporal Tables in DB2. IBM Data Magazine April 2012.
90. McGoveran David. Nothing from Nothing. (in four parts) In: Date CJ, Darwen Hugh, McGoveran David, eds. Relational Database Writings 1994-1997. Reading, Mass.: Addison-Wesley; 1998.
Part I of this paper explains the crucial role of logic in database systems. Part II shows why that logic must be two-valued logic (2VL) specifically, and why attempts to use three-valued logic (3VL) in this context are misguided. Part III examines the problems that three-valued logic (3VL) is supposed to solve. Finally, Part IV describes a set of pragmatic solutions to those problems—including in particular some recommended database design approaches—that don’t involve 3VL.
91. McKenzie E, Snodgrass R. Supporting Valid Time: An Historical Algebra. Tech. Report TR87-008 Chapel Hill, N.C.: Dept. of Computer Science, University of North Carolina; 1987.
Defines an algebra for the management of valid time (called stated time in the present book). A historical relation must have at least one valid time attribute (though the attribute in question is apparently not a regular attribute as such). Values of such an attribute are sets of time points (e.g., the set {d01,d02,d03,d04,d08,d09,d10}). Tuples that are identical except possibly for their valid time component are said to be “value equivalent”; relations aren’t allowed to contain distinct but value equivalent tuples. The existing operators of the relational algebra are adjusted in such a way as to enforce this requirement. New operators are also defined.
92. McKenzie Edwin, Snodgrass Richard. Extending the Relational Algebra to Support Transaction Time. In: Proc. ACM SIGMOD Int. Conf. on Management of Data, San Francisco, Calif. May 1987.
This paper provides a formalization of the concept of transaction time (called logged time in the present book) in terms of denotational semantics. The fundamental observation is that, as explained in Chapter 17, a database is really a variable; an update causes one value of that variable to be replaced by another, and the complete set of such values over time can be thought of as a chronologically ordered sequence. (Actually, the paper talks in terms of individual relations rather than the database as a whole, but the foregoing remarks are still applicable, mutatis mutandis.) The paper proposes extending the relational algebra to include an update operator called rollback.2
Note: In the introduction, the paper says that in “Codd’s relational algebra … the relations … model the current reality as is currently best known.” We believe we’ve shown in the present book how “Codd’s relational algebra” can be used to deal with all kinds of temporal data, not just “the current reality” but past and future states of affairs as well.
93. Edwin McKenzie Jr L, Snodgrass Richard T. Evaluation of Relational Algebras Incorporating the Time Dimension in Databases. ACM Comp Surv. December 1991;23.
To quote from the abstract: “In this paper we survey extensions of the relational algebra [to support queries on temporal databases] … We identify 26 criteria that provide an objective basis for evaluating temporal algebras … Twelve time-oriented algebras are summarized and then evaluated against the criteria.” Note: Another approach to evaluating alternative temporal proposals can be found in reference [78].
94. Shamkant B. Navathe and Rafi Ahmed: “Temporal Extensions to the Relational Model and SQL,” in reference [110].
Proposes a scheme in which intervals are represented by distinct “begin” and “end” attributes (i.e., the interval [b:e] is represented by distinct attributes with values b and e, respectively—more or less as in the SQL standard, in other words, except that SQL uses the closed:open interpretation). As for the relational operators: Projection remains unchanged. Restriction is extended to support interval comparisons such as OVERLAPS (expressed in the proposed SQL extension by means of a WHEN clause). Four distinct joins are defined, two giving a result with one pair of begin/end attributes and two giving a result with two such pairs. Finally, four new restriction-like operators are defined that involve only the begin and end attributes.
95. Özsoyoglu G, Özsoyoglu ZM, Matos V. Extending Relational Algebra and Relational Calculus with Set-Valued Attributes and Aggregate Functions. ACM TODS. December 1987;12.
96. Rolland C, Bodart F, Leonard M, eds. Temporal Aspects in Information Systems. Amsterdam, Netherlands: North-Holland; 1988.
97. Saracco Cynthia, Nicola Matthias, Gandhi Lenisha. A Matter of Time: Temporal Data Management in DB2 10 April 2012; In: http://www.ibm.com/developerworks/data/library/techarticle/dm1204db2temporaldata/; April 2012.
98. Sarda NL. Algebra and Query Language for a Historical Data Model. BCS Comp J. February 1990;33.
This reference, along with reference [99], appeared shortly after reference [68] (and they do resemble this latter reference, somewhat). It—i.e., reference [98]—defines a “temporal relational algebra” consisting of Codd’s original operators plus two more whose functionality resembles that of FOLD and UNFOLD as defined in reference [68].
99. Sarda NL. Extension to SQL for Historical Databases. IEEE Transactions on Knowledge and Data Engineering. February 1990;2.
Applies the ideas of the reference [98] to SQL specifically.
100. Schueler B-M. Update Reconsidered. In: Nijssen GM, ed. Architecture and Models in Data Base Management Systems. Amsterdam, Netherlands: North-Holland; 1977.
The source of the important idea, or shift in perspective, to the effect that “the log is the real database” (see Chapter 17). Schueler argues forcefully that destructive overwriting operations—i.e., UPDATE and DELETE operations as conventionally understood—should be outlawed. Instead, every item in the database should be thought of as a chronologically ordered stack: The top entry in the stack represents the current value of the item, and previous values are represented by entries lower down; thus, an INSERT or UPDATE places a new entry on the top of the stack and pushes all existing entries one place down. (We’re being deliberately vague here as to what exactly an “item” might consist of.) Each entry is timestamped, and all entries are accessible at all times. Schueler claims that such a scheme would dramatically simplify the structure of the system with respect to recovery, auditability, locking, archiving, understandability, and usability—not to mention the purely temporal issues that are the focus of the present book—and hence reduce system costs and improve system functionality in a variety of important ways.
101. Shoshani Arie, Kawagoe Kyoji. Temporal Data Management. In: Proc. 12th Int. Conf. on Very Large Data Bases, Kyoto, Japan. August 1986.
Sketches another nonrelational “temporal data model.” The model in question is motivated by research into scientific and statistical databases, which typically contain sets of measurements (e.g., results of experiments). The basic data object is the time sequence, which is a chronologically ordered sequence of <time:value> pairs for some given surrogate (where the surrogate in question denotes some entity). For example, the time sequence <2012:130>, <2013:138>, <2014:142> might represent the weight of some given person on three consecutive birthdays.
102. Erwin Smout: “Could Packing Be Specified in Terms of Closed-Open Interval Semantics Without Having to Commit to Granularity?”, http://www.thethirdmanifesto.com/papers.html/packing-without-granularity.pdf (to appear). Note: A draft of this paper dated January 14th, 2014 was offered as a contribution to the email discussion forum ttm@thethirdmanifesto.com, and the remarks that follow are based on that draft.
The formalism on which the approach to temporal data discussed in the body of this book is based assumes that a successor function (“NEXT_T”) always exists. As we know from Appendix B, however, that assumption has been the subject of some criticism. The present reference [102] shows that such an assumption might in fact be unnecessary. In other words, it presents an alternative formalism, one that (in effect) assumes that point types are ordered but not necessarily ordinal. In outline, what it does is this: First, it defines an interval as a pair of values, FROM and TO (necessarily of the same ordered type), such that FROM < TO (intervals are thus explicitly assumed to be in closed:open form). It then defines what it means for two such intervals to be equal, to meet, or to overlap, in all cases avoiding any reliance on a successor function. Using these operators, it then defines PACK—not UNPACK!—and various U_ operators (U_join, etc.). It also gives a precise definition of packed form (here called packed normal form),3 and algorithms for achieving that packed form.
We continue to believe that the UNPACK operator and unpacked form are conceptually useful tools for explaining our approach to temporal data in an intuitive fashion. It is, however, extremely interesting, and indeed encouraging, to note that those tools might not be totally necessary as a logical underpinning for that approach.
103. Snodgrass Richard, Ahn Ilsoo. A Taxonomy of Time in Databases. In: Proc. ACM SIGMOD Int. Conf. on Management of Data, Austin, Tex. May 1985.
The source of the terms transaction time, valid time, and user defined time. Note: Transaction time and valid time were discussed at length in the present book, but “user defined time” wasn’t mentioned. This latter term is used in the present reference to refer to temporal values and attributes that are “not interpreted by the DBMS”; examples might be date of birth, date of last salary increase, and time of arrival. Observe, however, that in the approach to temporal databases advocated in the present book, transaction times and valid times are also—like all other values and attributes!—“not interpreted by the DBMS.”
104. Snodgrass Richard. The Temporal Query Language TQuel. ACM TODS. June 1987;12 See also Richard Snodgrass: “An Overview of TQuel,” in reference [110].
TQuel was a version of the Ingres language QUEL, extended to support temporal data. This paper describes the TQuel language and a prototype implementation.
105. Snodgrass RT. TSQL2 Language Specification. ACM SIGMOD Record. March 1994;23.
106. Snodgrass Richard T, ed. The TSQL2 Temporal Query Language. Norwell, Mass.: Kluwer Academic Publishers; 1995; Note: A new edition of this book was published by Springer in 2012.
The definitive reference on the TSQL2 language. The following quote from a Kluwer brochure gives a sense of the book’s scope: “A consensus effort of eighteen temporal database experts has resulted in a new temporal database query language, TSQL2, which is upwardly compatible with the international standard [i.e., SQL:1992].4 TSQL2 provides comprehensive support for temporal applications. No other temporal query language covers as much ground. TSQL2 was designed to be compatible with existing database design and implementation methodologies. The complete specification of the language is included. TSQL2 is also an effective platform for teaching temporal database concepts, eliminating the need to constantly switch between incompatible language proposals.” See also references [8], [105], [107], and [108], and (for an alternative point of view) references [29] and [54].
107. Snodgrass Richard T, Böhlen Michael H, Jensen Christian S, Steiner Andreas. Adding Valid Time to SQL/Temporal and Adding Transaction Time to SQL/Temporal 1997; Madrid, Spain January ftp://sqlstandards.org/SC32/WG3/Meetings/MAD_1997_01_Madrid_ESP/mad146.ps; 1997; (This is an internal working document, not generally available at the time of writing.).
These two papers together constitute a TSQL2-based proposal for incorporating temporal functionality into the SQL standard [i.e., SQL:1992]. They were originally submitted to the January 1997 meeting of the ISO committee JTC1/SC21/WG3 Database Languages rapporteur group, but were subsequently withdrawn. See also reference [31].
108. Snodgrass Richard T. Developing Time-Oriented Database Applications in SQL San Francisco, Calif.: Morgan Kaufmann; 2000.
This book uses TSQL2 [105-106] as a basis for explaining in generic terms how to implement temporal applications and databases in SQL:1992 (in other words, using SQL DBMSs that have no built in temporal support at all, apart from SQL’s standard datetime types or something analogous). It also includes product specific information and suggestions for several different commercially available (but, at least in 2000 when the book appeared, nontemporal) products: Microsoft Access and SQL Server, IBM DB2 Universal Database, Oracle8 Server, Sybase SQL Server, and UniSQL.
109. Tansel Abdullah U. Adding Time Dimension to Relational Model and Extending Relational Algebra. Information Systems. 1986;11.
This paper can be seen as a more formal version of material from reference [18]. See also reference [112].
110. Tansel Abdullah Uz, Clifford James, Gadia Shashi, Jajodia Sushil, Segev Arie, Snodgrass Richard, eds. Temporal Databases: Theory, Design, and Implementation. Redwood City, Calif.: Benjamin/Cummings; 1993.
This book is a collection of papers. It’s divided into four parts, as follows:
1. Extensions to the Relational Data Model
2. Other Data Models
3. Implementation
4. General Language and Other Issues in Temporal Databases
Several of the other references mentioned in this appendix are included in this book.
111. Tansel AU, Garnett L. Nested Historical Relations. In: Proc. ACM SIGMOD Int. Conf. on Management of Data, Portland, Ore. June 1989.
See the annotation to references [18] and [112].
112. Abdullah Uz Tansel and Erkan Tin: “Expressive Power of Temporal Relational Query Languages and Temporal Completeness,” in reference [55].
The idea of trying to define some notion of “temporal completeness” seems like a worthy objective (in this connection, see also references [14] and [114]). This paper proposes that a temporal query language (a) should be relationally complete [22] but (b) should also support interval selectors and the “point e interval” operator (to use the terminology of the present book). The paper then evaluates a number of proposed approaches with respect to these criteria.
113. David Toman: “Point-Based Temporal Extensions of SQL and Their Efficient Implementation,” in reference [55].
As its title suggests, this paper proposes an extension to SQL based on time points instead of intervals. Some interesting questions are raised concerning implementation. Answers to those questions could be relevant to languages based on intervals too, because the unit intervals resulting from UNPACK are effectively isomorphic to points, in a sense.
114. Tuzhilin Alexander, Clifford James. A Temporal Relational Algebra as a Basis for Temporal Relational Completeness. In: Proc. 16th Int. Conf. on Very Large Data Bases, Brisbane, Australia. August 1990.
To quote from the abstract: “We define a temporal algebra that is applicable to any temporal relational data model … We show that this algebra has the expressive power of a safe temporal calculus … We propose [this] temporal algebra … and the equivalent temporal calculus as … alternative [bases for defining] temporal relational completeness.”
115. van Roessel JW. Conceptual Folding and Unfolding of Spatial Data for Spatial Queries. In: Robinson VB, Tom H, eds. Toward SQL Database Extensions for Geographic Information Systems. Gaithersburg, Md: National Institute of Standards and Technology; 1993; Report NISTIR 5258.
This reference uses Lorentzos’s FOLD and UNFOLD operators (predecessors of PACK and UNPACK, respectively) as the basis for defining an approach to dealing with spatial data. See also reference [116].
116. van Roessel JW. An Integrated Point-Attribute Model for Four Types of Areal GIS Features. In: Proc. 6th Int. Symp. on Spatial Data Handling, Edinburgh, U.K. 1994; September 5th-9th.
117. Vassilakis Costas. Design and Optimized Implementation of a System for the Management of Historical Data Greece: Department of Informatics, University of Athens; October 1995.
Describes some research based on Lorentzos’s FOLD and UNFOLD operators [67-68].
118. Vassilakis Costas, Lorentzos Nikos, Georgiadis Panagiotis. Implementation of Transaction and Concurrency Control Support in a Temporal DBMS. Information Systems. 1998;23.
Identifies transaction management problems (and describes solutions to those problems) in a temporal DBMS implemented as a separate software layer on top of a conventional DBMS. For example, an operator such as PACK, if implemented in such a separate layer, might map to a sequence of several operations—possibly including operations to write intermediate results into the database—at the level of the underlying DBMS.
119. Viqueira JRR, Lorentzos NA. SQL Extension for Spatio-Temporal Data. VLDB Journal. April 2007;16.
The work described in this paper is based on previous research reported in references [79] and [84]. It formalizes an SQL extension for the management of spatiotemporal data (see reference [3])—which is to say, spatial data that evolves with respect to time—and shows how a single set of relational operators can be made to apply uniformly to conventional, temporal, interval, spatial, and spatiotemporal data.
120. Yu Wu, Sushil Jajodia, and X. Sean Wang: “Temporal Database Bibliography Update,” in reference [55].
This paper is one in a cumulative series. Earlier contributions are as follows (in reverse chronological sequence):
■ Vassilis J. Tsotras and Anil Kumar: “Temporal Database Bibliography Update,” ACM SIGMOD Record 25, No. 1 (March 1996).
■ Nick Kline: “An Update of the Temporal Database Bibliography,” ACM SIGMOD Record 22, No. 4 (December 1993).
■ Michael D. Soo: “Bibliography on Temporal Databases,” ACM SIGMOD Record 20, No. 1 (March 1991).
■ Robert B. Stam and Richard T. Snodgrass: “A Bibliography on Temporal Databases,” Data Engineering Bulletin 11, No. 4 (December 1988).
■ Edwin McKenzie: “Bibliography: Temporal Databases,” ACM SIGMOD Record 15, No. 4 (December 1986).
■ A. Bolour, T. L. Anderson, L. J. Dekeyser, and Harry K. T. Wong: “The Role of Time in Information Processing: A Survey,” ACM SIGMOD Record 12, No. 3 (April 1982).
See also reference [3].
121. Zhang Y. Multi-Temporal Database Management with a Visual Query Interface Sweden: Department of Computer and Systems Services, Royal Institute of Technology and Stockholm University; October 1997; Ph.D. Dissertation.
Zhang’s dissertation is based on Lorentzos’s FOLD and UNFOLD operators [67-68].