- Количество слайдов: 49
Legal Data Markup Software CS 501 Design Presentation November 9 th, 2000
Project Team Sponsors Developers n Professor William Arms n Ju Joh n Professor Thomas Bruce n Sylvia Kwakye n Jason Lee n Nidhi Loyalka Reviewer n Omar Mehmood n Amy Siu n Charles Shagong n Brian Williams
Introduction n Objective: US Code (ASCII) Wellformed, valid XML output used as input to other applications Goal of end-use: Making law available for general public use
Overview n n n Development Environment Execution Environment Software Design DTD Design Packaging
Development Environment n Hardware n Server n n 233 MHz Intel PII processor 128 MB memory 28 GB hard disk Notebook Computers n n n 400 MHz Intel Celeron processor 96 MB memory 4. 7 GB hard disk
Development Environment n Software n n n Red Hat Linux 6. 2 Perl 5. 6 SSH Secure Shell 2. 3 CVS 1. 10. 7 Emacs 20. 5. 1 VIM 5. 6
Execution Environment n Caveat Client upgrades execution hardware and software environment at own risk. LDMS not guaranteed to work under new conditions.
Execution Environment n Naming Standards n General Rule n n n Filename Naming Convention n n Must start with a word in lower case. First letter of addition words in upper case. Example: the. Perl. File. pl File Name Length n Maximum of 20 characters.
Execution Environment n Naming Standards n Function Names n n n Must begin with a verb Example: initialize. Module Variable names n n Must begin with qualifiers Example: $error_Last. Error. Message
Execution Environment n Naming Standards n Filehandle Names n n Xml Output File Names n n Must be all capital letters Same as input file name with “. xml” extension DTD Element Names n n Element names in capital letters Nested element names start with DIV
Execution Environment n Coding Standards n n n A function shall not exceed 100 lines. A function shall have preceding comments on its purpose, pre- and postcondition. A variable shall have a purpose comment. Each loop shall have begin and end comments. A 3 -space indentation shall be used for each block of code.
Execution Environment n Coding Standards n n Perl contractions shall not be used. Each file shall have a modification history log. Each file shall include a copyright and license notice. Version number shall correspond to major and minor revisions to software
Software Design n System Architectural Components n n n Modules and their descriptions Design Constraints Error Handling Application Environment User Interfaces
System Architecture Program Read and Parse File Language Parsing Output Figure 1: Top-level diagram of major architectural components.
UML Component Diagram
File Parser Component
Natural Language Component
Figure 2: UML class diagram for LDMS
Design Constraints n n n 8 -bit ASCII input files. Non-uniform title structure. Unattended operation.
Title Variation Example -CITE 11 USC Sec. 506 01/23/00 -EXPCITETITLE 11 - BANKRUPTCY CHAPTER 5 - CREDITORS, THE DEBTOR, AND THE ESTATE SUBCHAPTER I - CREDITORS AND CLAIMS -HEADSec. 506. Determination of secured status
Title Variation (cont’d) -CITE 46 USC Sec. 13102 -EXPCITETITLE 46 - SHIPPING Subtitle II - Vessels and Seamen Part I - State Boating Safety Programs CHAPTER 131 - RECREATIONAL BOATING SAFETY -HEADSec. 13102. Program acceptance 01/05/99
Error Handling n n n Handled at topmost level. Processed by Store. And. Output. Errors module. Standard report format:
Application Environment n Preconditions n n Input files must exist in a known path. Required hardware and software must be available. Sufficient system resources must be free. Postconditions n A valid, well-formed XML document conforming to our DTD will be produced.
User Interface Design n Very little runtime interactivity required. Command-line operation. Allows batch processing.
Status Reporting n n Frequency of status reports controlled by -L parameter. Default is no status reporting.
Module Diagrams n Diagrams can be divided into two categories: n Structural diagrams. n n Flow diagram. Behavioral diagrams. n n Culture diagram. Context diagram.
Flow Diagram U. S. Code (ASCII) House Cornell LII U. S. Code Public U. S. Code (ASCII) LDMS U. S. Code (XML)
Culture Diagram House Format of code is not negotiable. “Why does publishing take so long? ” Cornell LII Seriously faulty input must be manually resolved. LDMS XML should be double-checked. Public
Context Diagram House of Representatives Produces Legal Data Markup System Uses as Input Produces XML Executes U. S. Code Downloads Cornell Legal Information Institute Publishes
Navigational Tags n n n
Content Tags n n
Data Tags n n n
LDMS Tags in Action
LDMS Tags in Action
Packaging n Release package will include: n n Documentation Source Code Executable Files Data Files
Documentation n Source-level documentation. Program design document. DTD design document.
Source-Level Documentation n Required for inclusion in each build. Source code comments. Separate text files.
Program Design Document n n n Intended as developer/maintainer resource. High-level view of processing engine. Individual processing components. Component interfaces. Updated as development progresses.
DTD Design Document n n Resource for DTD developers and maintainers. List of all elements and use. List of all attributes and use. Modified as development progresses.
Source Code n n n Source code for prototypes will not be considered deliverables. Testing harnesses will not be considered deliverables. All source code for release version will be provided.
Executables and Data Files n n n One executable script file. No other executables will be included. DTD will be considered a deliverable.
Installation n n No installation script is planned. Path to Perl binary must be specified at head of executable script. Project directory must be copied in its entirety to desired location. Relative paths within directory must remain unchanged. User must have write permission