Sharon C. Salveter

Hysterical B-trees

In an order d B-tree with n nodes, a single insert or delete operation may cause B(logd n) splits and merges. Moreover, since an insert operation can exactly invert the effect of a delete operation, and vice versa, a sequence of m inserts and deletes can cause 8 (m logd n) splits and merges. We. present a scheme that reduces the number of splits and merges required to process m inserts and deletes to 0(m), while only modestly increasing the depth of the tree. On the surface it may not seem advantageous to limit the number of splits and merges. Before performing an insert or delete operation, it is generally necessary to perform a tree search involving 8(logd n) nodes. There are situations, however, where the number of splits and merges is the bounding factor in the time to insert and delete. If the application involves locality of reference that is, inserts and deletes tend to occur in close proximity in the tree maintaining ‘fingers’ that point into the tree can reduce search time substantially below 8(logd n) [2]. Furthermore, searching involves only reading the contents of nodes, while splitting and merging require modifying nodes. With certain storage strategies, writing into the tree is more costly than reading from the tree. B-trees are often proposed as index structures in databases. In a concurrent database environment, multiple processes can search the tree simultaneously,

PIQUE: A relational query language without relations

Abstract Relational database systems have gone far towards providing users with physical data independence. To use a relational database, users need not know the physical storage structures of relations, and are protected from changes in these structures. However, a user must still navigate among relations. In other words, if the information needed to answer a question spans several relations, he must explicitly specify how these relations are to be combined. Physical data independence is not enough. A user should also be afforded some degree of structural data independence. More specifically, he should be able to pose queries without having to explicitly navigate among relations in the database. Instead, the system should do the navigation for him. We consider universal scheme interfaces as a means for automatic database navigation, and introduce the concept of a generator as central to such navigation. We describe a particular generator based on the semantic notions of reducible and irreducible facts, and present PIQUE, an attribute-based query language designed to work with this generator. PIQUE is a concise, yet powerful, language with natural semantics. A distinguishing feature of PIQUE is that tuple variables in queries are bound implicitly and that the logical connectives “and”, “or,” and “not” can affect the binding, and, therefore, take on “semantic overtones.” Furthermore, the semantic interpretation PIQUE gives to these connectives is more natural than the one given by most other query languages. In the Appendix, we present PIQUEs formal syntax and semantics, along with the proof that PIQUE is relationally complete.

A transportable natural language database update system

Although natural language (NL) querying of databases has been an active research area for many years [Ha77, Da78, W076, Ka79, Wa78, Wz75] and at least one commercial system is available that supports NL DB querying [Ha79], little effort has been expended in support of NL DB update, as noted by Wiederhold[WiSl]. Salveter and Maier [SM82] have shown that because NL DB update is a fundamentally different problem from query, it is not possible to naturally extend NL query systems to also support update. The goal of this research is the design of a computer system that will support both NL query and update access to DBs. There are two reasons for addressing this problem. First, it is an interesting theoretical problem. Second, with the proliferation of personal computers and on-line DBs, we can no longer assume that end-users only query the DB and a professional programming staff provides update application programs.

ON THE EXISTENCE OF PRIMITIVE MEANING UNITS

Knowledge representation schemes are either based on a set of primitives or not. The decision of whether or not to have a primitive-based scheme is crucial since it affects the knowledge that is stored and how that knowledge may be processed. We suggest that a knowledge representation scheme may not initially have primitives, but may evolve into a primitive-based scheme by inferring a set of primitive meaning units based on previous experience. We describe a program that infers its own primitive set and discuss how the inferred primitives may affect the organization of existing information and the subsequent incorporation of new information.

Natural language updates

A great deal of research effort has been expended in support of natural language (NL) database querying. English and English-like query systems already exist, such as ROBOT[Ha77], TQA[Da78], LUNAREWo76] and those described by Kaplan[Ka79], Walker[Wa78] and Waltz[Wa75]. Little effort has gone to NL database update [KD81, Br81, Sk80, CHSI]. We want to extend l~ interaction to include data modification (insert, delete, modify) rather than simply data extraction. The desirability and unavailability of NL database modification has been noted by Wiederhold, et al[WiSl]. Database systems currently do not contain structures for explicit modelling of real world changes.

Transportable natural language database update

Natural language database access requires support of both query and update capabilities. Although a great deal of research effort has gone to support natural language database query, little effort has gone to support update. We describe a model of action that supports natural language database update, as well as query, and the implementation of a system that supports the model. A major goal of this research is to design a system that is easily transportable both to different database domains and different database management systems.

Research at SUNY: Stony Brook

The goal of the SYMCHEM project is the automatic discovery of synthesis routes for organic compounds. Future plans call for a natural language interface that will allow experts (chemists) to directly transfer their expertise to SYMCHEM. However, this is not an active research effort at this time.