End-User Programming for Scientific Databases

本项目将开发一个面向最终用户的构建科学数据库的工具。该工具的目的是通过从声明性的、易于理解的高级规范自动生成数据库和用户界面软件，从根本上减少设计、实现和维护复杂科学数据库的时间和费用。具体来说，该工具将生成带有用c++代码编写的X-Window接口的SYBASE数据库。该工具将分两个阶段开发：初始内核和高级阶段。该内核将支持数据库设计的开发和维护，以及从中自动生成实现。这一阶段的重点是自动化开发一种通用类型的数据库应用程序，该应用程序具有许多专门针对科学需求的特性，包括用引用或权威注释事实的能力、处理同义词和技术同义词典的功能，以及自记录数据模型。该内核将根据本提案作者先前为科学数据库应用程序开发的软件进行改编和推广，包括所建议类型的原型应用程序生成器。在第二阶段，我们将使用内核作为框架来研究更高级的问题，特别是如何扩展使用该工具可以轻松生成的应用程序的范围。在这个阶段，我们将确定当前不支持的特定功能类，并开发扩展来支持它们。我们期望这些扩展将包含指定功能片段的受限语言，例如用于构造名称或默认值的规则。最后，我们将尝试在参与数据库开发的生物社区的那些部分中传播该工具，并将他们的反馈纳入设计中。

This project will develop an end-user oriented tool for constructing scientific databases. The purpose of the tool is to radically reduce the time and expense of designing, implementing and maintaining complex scientific databases by generating database and user-interface software automatically from declarative, easily understood high-level specifications. Specifically, the tool will generate SYBASE databases with X-Window interfaces written in C++ code. The tool will be developed in two stages: an initial kernel and an advanced stage. The kernel will support the development and maintenance of database designs and the automatic generation of implementations from them. This phase will focus on automating the development of a generic type of database application with a number of features specifically oriented towards scientific needs, including the ability to annotate facts with references or authorities, facilities for handling synonyms and technical thesauri, and self-documenting data models. The kernel will be adapted and generalized from software developed previously by the authors of this proposal for scientific database applications, including a prototype application generator of the type proposed. In the second phase we will use the kernel as a framework in which to investigate more advanced problems, specifically, how to expand the scope of the applications which can easily be generated using the tool. In this phase we will identify specific classes of functionality not currently supported and develop extensions to support them. We expect that these extensions will involve the incorporation of restricted languages for specifying fragments of functionality, such as rules for constructing names or default values. Finally we will attempt to circulate the tool among those segments of the biological community involved in database development, and incorporate their feedback into the design.