The Cindex is a smaller version of Lindex. it contains the name of the Web server, and the subjects
that are contained within this Web server.
So in a sense, the Cindex is a "content" page on the related server.
It does not contain individual file URLs.
It was designed to be small, so retrieving it would be fast.
During the
index making process, the Lindex is constructed first,
then the Cindex is made from the Lindex.
The format of the Cindex does not follow any previously known format because the Cindex was designed to contain no redundant information. The format of Cindex is as follows :
Rumour'Cindex:\000ServerName: <related Web server name>\000 Keywords:<keywords><CR>
an example of this is shown below :
Rumour'Cindex:\000ServerName: http://ip4.cs.monash.edu.au:8000/\000Keywords: Internet Radio (1), Star Wars Links (1), Monash University (1), RTHK (1), X-Files Links (1), Star Trek Links (1), Distributed Indexing WWW System (1), Hong Kong (3), Home Page (1), algorithm animation (1),
See Appendix I for an example.
Each keyword also has a keyword counter, which indicates the
number of times that keyword has
appeared in the Lindex.
The search is not case sensitive, so the case of the keywords does
not matter.
The whole Cindex should not contain any CR until the end of the record,
and each item is
separated by a 
000 character.
Because the Cindex will be collected by other machines,
it must be placed in the Rumour index directory,
Rumour/ind/Cindex.ind.