Different Kinds of Server

First on our tour of the web , let's see what the different types of existing server look like using a W3 browser. (for the fancy demos, skip ahead )

FTP

An FTP server , for example, is simply presented as a hypertext list. The user does not have to know anything about FTP. This example is a server which has postscriptr copies of preprints of physics papers. We can fllow through the directory structure to the preprints directory , to the preprints from Rice , into the ps directory which I happen to know contains postscript files, and find a particular preprint of a physics paper .

Gopher

Looking at a Gopher server is very similar. Gopher is like WWW except that there is no hypertext: anything is either a plain document or a menu. Here we see an example of a gopher menu, and you can see that there are extra menu items which have been introduced to look like headings.

News

When a WWW browser reads news, it uses hypertext to give instant links between related articles and newsgroups. Looking for example at WWW's own newsgroup, comp.infosystems.www , let's select a reply message -- one starting with "Re:". [I can't leave you a link to this as news messages are deleted with time, so you have to pick one.] Notice that at the top of the article there is a link back to the newsgroup. Also, there is a link to any article references. If you are lucky with your choice, you will see a link within the text "in article <blah>, somebody writes". This is a good example of W3 software making links out of existing information.

W3 servers

Lastly, let us look at some real hypertext sepcifically written for W3. Here is the "STING" Software technology Interest Group service . We are presented with various options as to how we would like to search for data. Let's suppose we want to know what "objective" means in "objective C". We first select STING and give "Objective" as a search word. We get back a summary of what STING has found for us . As STING is relatively intelligent, it gives us pointers to other good places to look too. We chose the software engineering glossary entry , . This, like the search results, is a "virtual" hypertext document, in that it has been generated on the fly by a server. It isn't written in a disk file anywehere. It is relatively easy to write servers to generate hypertext spaces like this, allowing complex information to be very simply presented to a user with no special skills.

We have seen how a W3 client gives access to a vast range of existing data from existing servers, but how the hypertext data from HTTP (native W3) servers is a much richer and easier to explore.

(Part of the W3 seminar . on to "what can be done" )

Tim BL