The sucess of Rumour will depend on the number of Web servers that have Rumour installed. Since
Rumour is a distributed database system, the more wide spread Rumour is, the more
information it will store. Currently as only some Monash University Web servers have Rumour,
the number of topics stored in the Rumour system is limited.
Most of the Rumour components are written in the Perl scripting language and wwwlib, therefore it is possible
to implement Rumour on any platforms that supports Perl (for example Macintosh and Window NT).
Collector currently loads Cindex from a pre-defined list of servers. This will be extended so that it can
extract the list of servers from a Web server log. In the future, Rumour could be supported by Web browsers
so that the browser would invoke Collector whenever a link is followed.
Because of the way Rumour is designed, the only item that is exchanged between Rumours on different servers
is the Cindex. Hence, it is possible to extend a particular installation of Rumour without affecting
other installations. For example, different language recognizing modules can be developed to extract keywords
from documents.
The next step for the immediate future is to make Rumour ready to be distributed, so more people can install it
on their Web servers. Thus the effectiveness of Rumour can be assessed.