O. Deux. The story of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91-108 (1990).  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... modeling and processing complex objects in many applications, including information retrieval (IR) has led many researchers to study the NF 2 relational model, i.e. non first normal form relations [8] object oriented da tabases, deductive databases and deductive object oriented databases (e.g. [9, 10, 11]) Bib liographic data are naturally modeled as NF 2 relations [12, 13] Moreover, some terminologi cal network structures, e.g. thesauri and classifications, and citation networks are not com plex objects but rather represent transitive relationships and cannot therefore be modeled by the NF ....

Deux, O. The story of 02. IEEE Transactions on Knowledge and Data Engineering, 2(1), 1990, 91-108.

....December 1995. The work of Brodsky was supported in part by NSF RIA grant IRI 9409770, and by Office of Naval Research under prime grant No. N00014 94 1 1153. The work of Wang was supported in part by NSF RIA grant IRI 9409769, and by NSF grant BIR 9404831 [S86] ENCORE [Z89] O2 [D90], ORION [JWKL90] POSTGRESS [SRH90] and Illustra) Object Store [OHMS92] Open OODB [BKG93] and UniSQL. Surprising is that while these systems targeted to a large extent the above mentioned application realms, little was done to address the problem of optimization in the presence of expensive ....

O. Deux et al., The Story of O 2 . IEEE Transaction on Knowledge and Data Engineering, pp. 91-108. 1990.

....databases. The work done in object oriented databases is more closely related to our work. Such systems start with a general object oriented programming language and add support for fast associative access to collections of objects. Some representative systems are GemStone[12, 47, 48] O 2 [23, 24], Orion[4, 36, 37] and ObjectStore[40, 52] Work specifically dealing with 91 indexing falls into two categories: 1. Indexing schemes based on the structure of an object (that is, the values of instance variables) We will call these path based indexing schemes. 2. Indexing schemes based on ....

....interested in comparing space requirements. Various strategies have been proposed for index maintenance for path based indexes; a survey can be found in [9] Of these the multi index is the most commonly used scheme; it was originally proposed for GemStone[47] and is also used in the O 2 [24, 23] and ObjectStore[40, 52] databases. We will cover the GemStone scheme in some detail, then briefly cover the nested index and path index of Bertino and Kim[11] and the join index of Valduriez[60] emphasizing their differences from the GemStone scheme and each other. 4.1.1 Multi indexes ....

O. Deux et al. The story of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, March 1990.

....levels. Given the architecture of KRISYS and the different representational frameworks in client and server, our approach to knowledge processing can be placed somewhere in between OODBMS pursuing mere server processing, leaving the maintenance of the client buffer to the application programs [10], and OODBMS performing main memory query processing, having, however, the same data model in client and server [21] Finally, it is important to point out that our approach not only fits the special characteristics of KRISYS and of the implementational environment (LISP) but that it is feasible ....

Deux,O. et al: The Story of O2: IEEE Transactions on Knowledge and Data Engineering, Vol. 2, No. 1, March 1990, 91-108.

....semantics which do not address the issue directly. The authors of [3] address encapsulation but do not include other important object oriented features, like inheritance. Non monotonic multiple behavioral inheritance is a fundamental feature of object oriented data models such as O 2 [5] and Orion [10] The user can explicitly redefine (or override) the inherited attributes or methods and stop (or block) the inheritance of attributes or methods from superclasses. Ambiguities may arise when an attribute or method is defined in two or more superclasses, and the conflicts need to be ....

O. Deux and others. The Story of O2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, 1990.

....for relational DBMS. However, there is no comparable results in the context of object data models. The TEMPOS project 1 (Temporal Extension Model for Persistent Object Servers) investigates the feasability of a temporal server through the implementation of a prototype on top of the O 2 DBMS [8]. In this paper, we present preliminary results of the TEMPOS project: they include the definition of a multigranular temporal data model which gathers in an unified way the necessary concepts to model time and data histories. This model is formalized through the functional specification of type ....

O. Deux. The story of O2 . IEEE Transaction on Knowledge and Data Engineering, 2(1), March 1990.

....protocol to drive its operation and the well defined semantics of the objects operated on are exploited to reduce the amount of data transferred to maintain consistency. LOTEC builds indirectly on earlier work on distributed object systems (including [17, 18] and objectbases (including [14, 10]) LOTEC is most closely related to work by Fleisch and Hyde [11] and Itzkovitz and Schuster [12] 3. The Problem and Environment 3.1. The Environment As with LOTEC [20] we assume an environment based on distributed persistent objects. Users interact with the environment by invoking methods on ....

O. D. et al. The Story of # # . IEEE Transactions on Knowledge and Data Engineering, 2(1), 1990.

....the effectiveness of LOTEC andtoexplore the desirable operating characteristics of a network environment in whichitshould operate. Our work builds indirectly on earlier work on distributed object systems(including [RB94, SF95, BHAB96] and, in some ways, on objectbase systems (including [LLOW91, Deu90] Itismost closely related to work by Fleischand Hyde [FH98] and Itzkovitz and Schuster [IS99] The rest of this paper is organized as follows. Section 2 describes the problem domain and motivates our solution strategy. Section 3 describes nested object transactions and our correctness ....

O. Deux et al. TheStory of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108,1990.

....is done by including the OID of the referenced object in the value of the referencing object. For example, in Fig. 1 two parent objects, object1 and object2 are referencing object3, and object3 is being referenced by object1 and object2. 6 In recent object oriented databases such as O 2 [14], as mentioned in Section II.D, a complex object and a complex value are clearly distinguished# the former can be shared by other objects, whereas the latter cannot. Note that the described complex object model can deal with such a distinction by giving OID s for sharable complex objects of a ....

....3 updates the objects found bya selection operation. When a set specified by t.lectures is not clustered with the Teacher object, Query 3 requires navigation operations from the Teacher object to t.lecture. 1 We assume that the evaluated value of a query is a value, not an object like in O2 [14]. Although of course it is more general that the evaluated value is treated as true object which has its own OID, type, and value, such treatment requires an closed algebra of complex objects. Unfortunately there exists no agreed general algebra to our knowledge. 8 Indeed, navigation operation ....

O. Deux et al. The story of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, Mar. 1990.

....but provide little means of descriptive set operations [GM88] This reflects their origin from the programming language realm. Varieties of research prototypes and commercial products are available. For examples, Gemstone [MSOP86] which is the first OODBMS, extends Smalltalk [GR83] O2 [Deu90] and ObjectStore [LLOW91] evolved from C [Str86] The DBPLs developed using the programming language approach provide the ability to manipulate database relations or objects but all at the tuple or object level. The interface to the storage manager ends up with a tuple or an object at atime ....

O. Deux. The story of O2 system. IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, Mar. 1990.

....with monoid comprehension queries. For example, SAT(D) where D is of type Gen Disj T , is represented as SELECT SAT(c) INTO Some satisfflag FROM D AS T c 5 Optimization by Approximation based Filtering and Indexing General optimization of object oriented queries (e.g. ENCORE [Zdo] O2 [ea90] POSTGRESS [SRH90] and monoid comprehensions in particular, e.g. FM] as well as optimization in presence of expensive predicates [CS, HS] is outside the scope of this paper; We concentrate here on approximation based filtering, regrouping and indexing [BW95] that C 3 is designed to ....

O.Deux et. al. The story of o2. IEEE Transactions on Knowledge and Data Engineering, 1990.

....by NSF Grant CCR 90 57570 and ONR Contract NOOO14 93 11284. 1 Scaleup measures the running time when both the database size and the number of nodes on a parallel architecture increase by the same factor. 1 meet these new demands, new data models have been proposed. Object oriented databases [25, 20, 4] view the data as a collection of interlinked objects. The focus of query evaluation switches from processing massive, but uniform data sets to link traversal in a highly interconnected collection of objects. On a parallel shared nothing architecture, when a link traverses node boundaries, either ....

....either the query or the data has to migrate between nodes [33, 18] As a consequence, efficient parallel evaluation is harder to achieve for this model than for the relational one. An additional complication was caused by the absence of genuine query languages. With the important exception of O2 [25], OODBMS programming has been done only in general purpose object oriented languages extended with some collection datatypes and with a persistence mechanism (e.g. ObjectStore [44] Any attempt to parallelize such a language would face the usual problems of dealing with side effects, etc. The ....

[Article contains additional citation context not shown here]

O. Deux. The story of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91-- 108, March 1990.

.... 1 Model 2 Grid intersects Biotopes 100 Cities 1,000 same Biotopes 1,000 Cities 10,000 same northwest Biotopes 100 Cities 1,000 same Table 3: Test suite 1 ffl Nested Loop (NL) ffl Scan and Index (SI) ffl Synchronized Tree Traversal (STT) We chose for our experiment an O 2 DBMS platform [Deu89] on which the algorithms were implemented and run. The O 2 version 4.5 was run under Sun OS 4.1.3 on a Sparc station Sun System 10. For SI and STT, we used an efficient secondary memory implementation of a special quadtree data structure [Sam84] as an index developed at CNAM, Paris. The approach ....

O. Deux. The Story of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, March 1989.

....an example. Section 4 introduces the algebra. The algebraic equivalences used to unnest nested algebraic expressions are presented in Section 5 and applied to some representative nested queries. Section 6 concludes the paper. 2 Preliminaries The data model we are working on is similar to the O 2 [9], GOM [14] or Exodus [5] model. It features objects that have an identity, that are manipulated through user defined methods, whose structures are complex and that belong to classes that may be refined into subclasses. Each class has an extent which is a set containing all the objects belonging to ....

O. Deux. The story of o2. IEEE Transaction on Knowledge and Data Engineering, 2(1), March 1989.

.... strict inheritance . Namely, a class definition can only add new properties or restrict the value of the property inherited from a superclass. Now the question is how to indicate to the system which of the two options gives the desired semantics for Car factory. Conventional data models, such as [1, 3, 13, 15, 27, 28, 29] provide specialization mechanisms that support only one of the two options (see also Section 7) We believe that both options should be supported, as the appropriate option depends on the desired class semantics. Moreover, we believe that an a priori choice of one or the other option can lead to ....

....same origin then refinement data models usually adopt one of the following approaches: Approach 1: Only one of p 0 and p 00 is inherited. The selection of the inherited property is done automatically or is left to the user. Such data models are the ORION data model [3] and the O 2 data model [13]. Approach 2: One or both of p 0 and p 00 are inherited. Such data model is the ODE data model [1] Approach 3: The constraint is enforced that p 0 and p 00 must have identical values. If this constraint is satisfied then the properties inherited by class C from p 0 and p 00 coincide ....

O. Deux. The story of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91-108 (1990).

....membership is based on the fact that the object state contains less components than those of the class. Such notion can be extended to the case of objects whose state contains additional components with respect to those specified in the class, in the same spirit of the O 2 exceptional instances [10]. In this way we can achieve a more accurate classification method. Another possible extension could be that of allowing components to be dynamically added, or deleted, to the state of objects in the database. This could require a re classification of the object, that is, a migration of the object ....

O. Deux et al. The Story of o2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, 1990.

....promising ways to meet the demands of many advanced database applications. The object oriented philosophy creates a powerful synergy throughout the development life cycle by combining abstraction, encapsulation, and modularity. In the past decade, various object oriented data models were developed [3, 7, 8, 13]. Also see an review in [25] But there is a major problem with the objectoriented approach that is the lack of logical or mathematical foundations that, traditionally, has been playing an important role in database research. Such a foundation is essential for de ning the semantics of databases ....

....management systems. The objective of the OLOG system is to develop techniques for advanced intelligent information systems that directly support e ective storage, ecient access and inference of large amount of data with complex structures. The OLOG language [18] is based on IQL [2] and O2 [7]. It overcomes the problem associated with IQL. It e ectively integrates useful features in other deductive languages with a well de ned logical semantics. The OLOG system has been developed in C mainly on a SUN SPARCstation running Solaris 2.5. It is implemented as a persistent database system ....

O. Deux and others. The Story of O2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91-108, 1990.

....problems for both implementation and consistency. Chimera supports both object migration and dynamic addition of classes, leading to multiple class direct membership. By contrast, Chimera does not support exceptional instances, which is a peculiarity of the O 2 objectoriented database system [14]. Moreover, Chimera supports derived (or predicative) classes. That is, classes whose extents are not explicitly manipulated; rather those classes are implicitly populated in that a population predicate is associated with the class specifying sufficient and necessary conditions for an object to ....

....other than Chimera, namely in O 2 , GemStone and Iris. We then briefly discuss approaches to role modeling. Roles can be used, among other things, to provide alternative and or additional views of objects and thus can be used to support object evolution. 6. 1 Object Evolution in OODBMSs In O 2 [14] instances can be specialized. This enables attributes and methods to be added and redefined for an individual object. In redefining an attribute or a method, the new definition must be compatible with the definition given by the class. In this respect, rules governing subtype definitions are ....

[Article contains additional citation context not shown here]

O. Deux et al. The Story of 0 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, 1990.

.... visit Exhibition WHERE Artist.name = Romy Schneider AND Exhibition.address = Le Louvre 6 Implementation Some of the above mechanisms have been implemented with the Gram model and validated on top of the hypermedia system Multicard (Rizk and Sauter 1992) and the object oriented DBMS O 2 (Deux 1989). The overall system architecture is shown in Figure 5. Hypermedia Application O2C O2Engine O2Graph O2SQL Hypermedia Basic Classes Schema Browser Hypermedia Toolkit O2 Multicard MultiTalk Figure 5: Multicard O 2 Architecture In a first step, we have connected Multicard with O 2 by ....

Deux, O. (1989, March). The Story of O 2 . IEEE Transactions on Knowledge and Data Engineering 2 (1), 91--108.

No context found.

O. Deux. The story of O 2 . IEEE Transactions on Knowledge and Data Engineering, 2(1):91-108 (1990).

No context found.

O. Deux. The Story of O2. IEEE Transactions on Knowledge and Data Engineering, 2(1):91--108, March 1990.

No context found.

Deux, O., et al., "The Story of O2", IEEE Transaction on Knowledge and Data Engineering, 2(1), March 1990

No context found.

O. Deux et al. The story of O 2 , IEEE Transactions on Knowledge and Data Engineering, 2(1), 91-108, (1990).

No context found.

O. Deux et al. The story of O 2 , IEEE Transactions on Knowledge and Data Engineering, 2(1), 91-108 (1990).

No context found.

Deux, O. et al #1990#. The Story of 0 2 . IEEE Transactions on Knowledge and Data Engineering, 2#1#:91#108.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC