Typically, a computer on the Internet is characterized as either a "client" or "server" depending on the role that the computer is playing with respect to requesting information or providing information.

Client computers are computers that typically request information from a server computer which provides the information.

For this reason, servers are usually larger and faster machines that have access to many data files, programs, etc., in a large storage associated with the server.

However, the role of a server may also be adopted by a smaller machine depending on the transaction.

That is, user 110 may request information via their computer 108 from server A.

At a later time, server A may make a request for information from computer 108.

In the first case, where computer 108 issues a request for information from server A, computer 108 is a "client" making a request of information from server A.

Server A may have the information in a storage device that is local to Server A or server A may have to make requests of other computer systems to obtain the information.

User 110 may also request information via their computer 108 from a server, such as server B located at a remote geographical location on the Internet.

However, user 110 may also request information from a computer, such as small computer 124, thus placing small computer 124 in the role of a "server."

For purposes of this specification, client and server computers are categorized in terms of their predominant role as either an information requestor or provider.

Clients are generally information requestors, while servers are generally information providers.

Referring again to FIG. 1, data objects such as distributed hypermedia documents 10, 12 and 14, image 16 and sound data file 40, may be located at any of the computers shown in FIG. 2.

Since these data objects may be linked to a document located on another computer the Internet allows for remote object linking.

For example, hypertext document 10 of FIG. 1 may be located at user 110's client computer 108.

When user 110 makes a request by, for example, clicking on hypertext 20 (i.e., the phrase "hypermedia"), user 110's small client computer 108 processes links within hypertext document 10 to retrieve document 14.

In this example, we assume that document 14 is stored at a remote location on server B.

Thus, in this example, computer 108 issues a command that includes the address of document 14.

This command is routed through server A and Internet 100 and eventually is received by server B.

Server B processes the command and locates document 14 on its local storage.

Server 14 then transfers a copy of the document back to client 108 via Internet 100 and server A.

After client computer 108 receives document 14, it is displayed so that user 110 may view it.

Similarly, image object 16 and sound data file 40 may reside at any of the computers shown in FIG. 2.

Assuming image object 16 resides on server C when user 110 clicks on image icon 22, client computer 108 generates a command to retrieve image object 16 to server C.

Server C receives the command and transfers a copy of image object 16 to client computer 108.

Alternatively, an object, such as sound data file 40, may reside on server A so that it is not necessary to traverse long distances via the Internet in order to retrieve the data object.

The Internet is said to provide an "open distributed hypermedia system."

It is an "open" system since Internet 100 implements a standard protocol that each of the connecting computer systems, 106, 130, 120, 132 and 134 must implement (TCP/IP).

It is a "hypermedia" system because it is able to handle hypermedia documents as described above via standards such as the HTTP and HTML hypertext transmission and mark up standards, respectively.

Further, it is a "distributed" system because data objects that are imbedded within a document may be located on many of the computer systems connected to the Internet.