Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.adass.org/adass/proceedings/adass97/mckayn.html
Дата изменения: Fri May 15 22:17:28 1998 Дата индексирования: Tue Oct 2 01:31:32 2012 Кодировка: Поисковые слова: легирование |
Next: Teaching Astronomy via the Internet
Up: Education and Public Outreach
Previous: Cyber Hype or Educational Technology? What is being learned from all those BITS?
Table of Contents -- Index -- PS reprint -- PDF reprint
N.P.F. McKay1 and D.J. McKay1
Nuffield Radio Astronomy Laboratories, University of
Manchester, Jodrell Bank, Macclesfield, Cheshire SK11 9DL, UK
1Formerly: Australia Telescope National Facility, CSIRO, Paul
Wild Observatory,
Locked Bag 194, Narrabri NSW 2390, Australia
To help astronomers and engineers at the various institutions experiment with and visualise the effects of different positions of the new stations, a design and simulation software package was required. This program had to be intuitive to use, widely available and quickly developed. The resulting program was called the Virtual Radio Interferometer (VRI).
As well as exploring the properties of several existing instruments, the user has control of a number of telescope parameters, allowing the creation of hypothetical interferometers. These include the latitude of the site and the number of array elements which form the interferometer, together with their diameter, elevation limit and position on a two dimensional plane. The declination of the source, as well as the frequency and bandwidth of observation can also be specified. Once these parameters are set, a corresponding uv-plot is produced by the program, showing the aperture-plane coverage.
Given the plot of the uv-coverage, we can simulate its effect on various objects. An image of the source is loaded onto one of four display panels, a Fourier transform performed on it, and also displayed. This can be then masked with the uv-coverage. When the inverse Fourier transform is applied, the original image of the source is replaced by the image obtained by ``observing'' it with the simulated interferometer. Finally, uv-plots generated by a number of antenna configurations, or even different observatories, may be combined to generate an image composed of multiple ``observations''.
Java was chosen as the coding language, because it satisfied these requirements. As a language, it is high-level, strictly object-oriented, and easy to learn. Classes which facilitate the writing of software for Internet distribution form part of the standard Java Development Kit. Hence, setting a program up to run on the Web is not a complex exercise. The source code is partially compiled into a set of byte codes before released for use. These are platform independent, hence eliminating the need for multiple versions of the software. The byte code is downloaded onto the client computer and, at run time, is interpreted into native instructions and executed by the host machine. This alleviates congestion on the organisation's HTTP server, and once the program is downloaded onto a local machine, provides a faster program for users.
Being a semi-interpreted language, programs run slower than their native code counterparts. For example, a point Fourier transform takes 15 seconds to be executed by the Java interpreter, on a Sun Ultra-1 running at 140 MHz, which can be compared with figures of less than one second for native code. Other potential reasons for avoiding the language include the fact that Java is still somewhat immature and evolving, the lack of complete control over registers, memory, etc., provided by other languages such as C and C++ and the problems posed by integrating a new language with legacy code.
Nevertheless, the portability of the language, the rich vocabulary of standard class functions, as well as Java's easily extended object-oriented design, greatly outweigh these disadvantages.
Java Development Kit Version | 1.0.2 | |
Software classes | 26 | |
Java instructions | 1565 | |
Development time | 80 | hours |
Size of byte code | 1.3 | Mbyte |
The authors are of the opinion that the Java programming language has a useful role to play in general astronomical software. We would not, however, recommend its use for all applications. Its strength lies in the regime of off-line applications, interactive tools and, of course, user interfaces.
The Australia Telescope is funded by the Commonwealth of Australia for operation as a National Facility operated by CSIRO.
Next: Teaching Astronomy via the Internet
Up: Education and Public Outreach
Previous: Cyber Hype or Educational Technology? What is being learned from all those BITS?
Table of Contents -- Index -- PS reprint -- PDF reprint