As mentioned in the introduction to this chapter, Hyper-G should really be seen as an extension of WWW. Hyper-G is a more sophisticated WWW tool for more professional users: it has been designed to manage big amounts of hypermedia data. Thus it would be a real ``overkill'' to set up a Hyper-G server just to store your own home page and the pictures of your last holidays on it. But if you are going to start a real information system (having to deal with about 100 or even up to millions of nodes) then you should really consider Hyper-G to be your Web server. One of the most important features of Hyper-G is that a growing number of people think that it is indeed the best WWW server for dealing with large amounts of data.
In the first part of this chapter we discussed the user' s point of view on accessing a Hyper-G server using a WWW client like Mosaic or Netscape. In fact, the user has fewer advantages when using Hyper-G than the system administrator and all service providers. Although you might be able to set up a WWW server that mimics most features of a Hyper-G server, this would be lots of work! The system administrator would have to write scripts to check in and out links of HTML documents to maintain some kind of weak (only on the local server) link consistency. It would be necessary to install two search engines (title and full text) just to provide the two most frequently used search facilities of Hyper-G and a gateway to some kind of relational database to store attributes (name/value pairs) in a similar style to the way a Hyper-G server would. Also, among other things, structuring tools available in Hyper-G would have to be emulated in a tedious way.
Another advantage of using Hyper-G as a sophisticated WWW server is Hyper-G's own security system. For administrators this means that they can find out because all accesses to the server are logged what each person did to the database at a particular time. Of course, anonymous users are allowed in Hyper-G to avoid the big-brother syndrome. An important feature is also that the administrator can give write access to some parts of the collection hierarchy to some persons or groups: the information management itself can be decentralized in this fashion.
Identified users can restrict read access to those parts of the server they have write access to so that (parts of) the database are ``for their eyes only.'' Hyper-G is more secure than WWW, because you can have usernames and passwords on Hyper-G which differ from those on your UNIX host. Thus it is not as easy to break into Hyper-G servers as with most WWW servers: if you want someone to access only parts of an ordinary WWW server, you have to use the UNIX security system, since WWW documents are stored directly using the UNIX file system.
The Hyper-G server concept has many advantages over ordinary HTTP/WWW servers that are not immediately obvious if you have not really worked with both systems. The main advantage of course is that you have an advanced Web server with data management facilities, search capabilities, link management including bidirectional hyperlinks, a UNIX-like security system and multi-language support without having to write a lot of software on your own. All this is provided with your Hyper-G server software.