Naturally, our prototype does not incorporate any actual functionality of the
system, but it does present our vision of how it will look when it is
constructed.
2. Clarification of Critical Issues
The prototype runs on SCF's web server. When the actual system is developed, CGI scripts will be used for authentication, the bulletin board, and query processing. The web page authoring operations and will be implemented as a Java applet.
The main purpose of the ICTM is to give professors with little or no technical background the ability to easily create web pages for their classes. Thus, it is of utmost importance that the system have a simple user interface, as is reflected in our prototype. This was critical when designing the web page authoring interface, since we did not want to present the instructor with a lot of confusing menus and icons. Instead, we have a simple array of buttons for the creation of new pages, selection of web pages, addition and deletion of links and images, and fetching files. There is also a text editor/"dummy page" where the user will modify and preview their edited page.
Unfortunately, the downside of a simple user interface is that it limits the web authoring capabilities of the instructor. However, it is not expected that the professor will need any sophisticated web page authoring operations. Hence, our system is designed to support only those functions which are essential to the construction of a simple web page. If the customer later requests additional authoring capabilites (creation of tables, lists, etc.), those operations could be incorporated into the system.
Another issue is how to minimize the overhead of translating web pages. Our solution is to create translations of web pages as soon as they are updated by the instructor. Therefore, after a page is modified and saved by the professor, the updated page is sent to the translator which creates a French and English version of the web page. These translations are saved in the student site along with the original version (which could be a mixture of both French and English), so that they can be immediately accessed whenever a user requests a specific translation. Thus, all the overhead incurred due to translation takes place right after a web page is modified, and each web page always has its translated versions ready to be loaded as soon as a user clicks on the link to view it. Handling translations in this manner does require extra disk space to store the different versions of each page, but this cost is of little consequence considering the potential improvements in performance.
Some of the files stored on the student site may be scanned images of newspaper or journal articles. Although it would be desirable to have the ability to search for keywords on text contained within images, we would need to incorporate an OCR(optical character recognition) program into our system to convert the scanned documents into text files which we could do searches on. Unfortunately, we could not find a feasible way to do the conversion from scanned articles to text files. One potential method would be to send articles to a company called JSTOR that specializes in OCR, and have them do the conversions. However, this process would be both time-consuming and costly. Another solution would be for USC to purchase OCR software, so that image to text conversions could be done on campus, or even integrated into the ICTM. These ideas are all possiblities for future expansion and evolution of our system.