Re: HTML spec. version identifiers

Dave_Raggett (dsr@hplb.hpl.hp.com)
Tue, 20 Jul 93 9:59:56 BST


Kevin Altis asks:

> Will HTML+ and future HTML versions have version identifier tags so that
> backward compatibility can be maintained without complicating the parsing
> process. It isn't necessary for current HTML documents to have any kind of
> identifier, but if HTML+ and future versions had a version tag to go at the
> beginning of the document, then that would simplify extending HTML in the
> future and also allow a document to be parsed differently based on its
> version. A document without a version would be parsed as the current HTML
> spec or what appears in the upcoming RFC.

Tim has been asked by folks in the IETF for something to identify HTML docs
as being such, and I have included in my spec the recommendation that authors
include a version statement. The standard way to do this in SGML is via
the DOCTYPE element, e.g.

<!DOCTYPE htmlplus PUBLIC "-//Internet/RFC xxxx//EN">

This states that the document conforms to the htmlplus DTD as defined by
the named publicly available specification. The name used here will be
instantiated to something more formal by the time the HTML+ spec becomes an
RFC, and would change with subsequent revisions.

This may seem like overkill for small manually keyed documents, but with
WYSIWYG editors almost upon us, it will be entered automatically as a matter
of course when you create a new document.

Dave