Abstract:

HTML is the lingua franca for global hypertext: there are millions
and millions of pages in HTML. The Web has been successful beyond
the wildest dreams of the people involved in its early days! HTML is
a powerful enabler for commerce allowing businesses to reach out to
consumers and other businesses, as well as enabling effective
communication within businesses and other organizations, not
forgetting its role for personal Web pages and special interest
groups. 

The next generation of HTML will seek to reduce the cost of
developing Web content, whilst at the same time increasing the
ability of authors to provide rich browsing experiences on an ever
widening range of Web clients. It is likely to lead to a large range
of tools for authoring, managing, transforming and rendering HTML. 

While some content providers will be prepared to design for the
specific needs of a range of browsers, many will not. Work on the
next generation of HTML will seek to encourage the development of
authoring and related management tools that reduce the cost of
creating content that can be easily repurposed for a wide variety
of browser capabilities. 

HTML in this world view will belong to a family of XML tag sets
designed to work together and sharing a common namespace. Some of
these will be structural, e.g. for representing document idioms such
as headings and paragraphs, or for relational data and forms, while
others will describe presentation, e.g.  for scalable graphics and
temporal synchronisation of media elements, and aimed at particular
classes of browser capabilities, e.g. bitmapped color displays and
printers, or voice browsers.

Style sheets and scripts can be used to transform from structural
markup into that targetted for specific classes of browsers. A key
goal is to allow these transformations to be applied where
appropriate, e.g.  by the author working on the Website, at a
proxy server under the control of an ISP or in the browser itself. 

Web browsers have increasingly disparate capabilities. At the same
time the Web has become an increasingly important means for building
distributed applications. To cope with this, it is proposed that
W3C defines profiles specifying which tag sets can be combined to
match the needs of each class of browser.

If several different classes of browser use the same Web address
to access the same Web resource, this can create difficulties for
effective caching. The cache may unknowingly deliver the wrong
variant of the resource to the browser, a variant the browser
does not support. This can be solved by making caches smarter so
that they can identify whether the cache holds a variant that
matches the capabilities of a given browser.