The W3 project has deined a number
of common practices which allow all
the clients and servers to communicate.
When you are reading a document,
behind every link there is the network-wide
address of the document to which
it refers. The design of these addresses
(URLs) is as fundamental to W3 as
hypertext itself. The addreses allow
any object anywhere on the internet
to be described, even though these
objects are accesed using a variety
of different protocols. This flexibility
allows the web to envelop all the
existing data in FTP archives, news
arcticles, and WAIS and Gopher servers.
The web uses a number of protocols,
then, but it also has its own Hypertext
Transfer Protocol (HTTP). This protocol
includes a number of facilities which
we needed: it is fast, stateless
and extensable. It also allows the
web to surmount the problems of different
data types using negotiation of the
data represeentation as already described
.
The other protocols which W3 clients
can speak include FTP, WAIS, Gopher,
and NNTP, the network news protocol.
Although W3 uses many different formats,
this is one basic format which every
W3 client understands. It is a simple
SGML document type allowing structured
text with links. The fact that
HTML is valid SGML opens the door
to interchange with other systems,
but SGML was not chosen for any particular
technical merit. HTML describes
the logical structure of the document
instead of its formatting. This
allows it to be displayed optimally
on different platforms using different
fonts and conventions.
Part of the W3 seminar . On to W3
software
Tim BL