Laserfiche WebLink
CRDSS <br />TASK MEMORANDUM 1.05-23 <br />Study of System Integration Issues <br />Software - File Organization for CRDSS <br />1.0 ISSUE <br />Organizing files in a large system is a very time-consuming process for a number of reasons. First, <br />the sheer number of files that require a "location" may require a large amount of time. Second, as <br />the number of files on the system increases, whoever is responsible for the files may realize that a <br />better way of storing the files is necessary and must decide whether to spend time reorganizing the <br />files "now or later." This memorandum discusses the following issues related to file organization for <br />the CRDSS: <br />What is the basic file organization framework to be used for the CRDSS? <br />? <br />How will flexibility be built into the file organization structure? <br />? <br />2.0 DISCUSSION/ANALYSIS <br />File Structure <br />File organization can take advantage of two basic features of file systems: <br />(1)Directory structure. <br />(2)File (and directory) names. <br />Directories can be used to categorize files into reasonable groups. File and directory names can be <br />used to indicate the contents of files. If many similar files exist (e.g., time series files), file names <br />can follow a standard to indicate how the files are different. Directories can be utilized so that <br />information is more specific as the number of directory levels increases. <br />A good initial file structure for a system can help to limit the time spent managing files by <br />automatically providing a location for each file. In such a system, the data "grows into" the file <br />structure, rather than pushing the redesign of the file structure. <br />RTi has a great deal of experience with storing both software and data for large systems. This <br />experience will be used to put together the initial file structure for the CRDSS. However, it is RTi's <br />experience that it is very difficult to completely define the file structure for a large system up front <br />because many issues cannot be resolved until data collection begins. It is therefore important that the <br />system file structure by simple and flexible enough to allow for future changes without disrupting the <br />whole system. <br />One of the most important issues in system organization is that of separating data from programs and <br />source code. This is important because there will be a time when the executable programs (and <br />1 <br />A275 05.10.94 1.05-23 Malers <br />