RSS is not hypertext. Isn't that great?

by Bob DuCharme

(An early draft of this was originally posted to the xml-hypertext list in response to Dave Pawson's question "What is the difference between hypertext and rss?", but this is more of a bully pulpit so I'm putting a revision of it here.)

Hypertext is is a way to present links and RSS is a way to mark up a
class of links for a particular domain.

People can argue about the definition of "hypertext," but XLink's
definition of "hyperlink" works for me: "a link that is intended primarily
for presentation to a human user." The word "presentation" is key; I think we can also assume that an interactive, electronic medium is what puts the "hyper" in "hyperlink." A printed footnote reference is a link, not a hyperlink.

RSS is a way to describe links for a particular domain: relationships
between a resource (or subresource--in this case, a specific element storing the
title of a story) and another resource (the web page storing the complete story) in
the world of news. Everyone's free to call anything they want "news" if it can fit into the RSS structure, so this has been a boon to weblogs.

If we distinguish between link presentation systems (for example, hypertext, endnotes,
and sidebars) and classes of links specialized for particular domains (such as RSS, legal
case citations, and bibliographic references) we see a many-to-many relationship,
which is a very good thing. People can mix and match link presentation styles
with link domain classes. Hardcoding presentation styles to link domains loses
this flexibility. If I ship all header-story links, case citations, and
bibliographic references marked up <a href="http://whatever">like this</a>,
I lose that flexibility--what if I want to represent bibliographic reference
links differently from case citation links?

RSS links aren't inherently hypertext links; that's only one justifiably
popular way to present them to the user. Thinking of them as "links" and not
just "hypertext links" opens up the possibilities of what you can do with
them. For example, you could write a stylesheet that reads an RSS file and
writes each header, description, and linked story together in a XSL-FO file
for printing as a hot sheet report. The "link" isn't actively traversed by the
reader, but the link's existence still benefits the reader, because a
production step traverses the link for the reader. Maybe I sound like an old SGML hack, but I still believe that separating presentation information from content increases the value of the content because it lets you do more with that content--even when the "content" is data about link relationships.

What do you think?


2003-04-15 15:47:48
This is insightful?
This "insight" more accurately refers to the difference between HTML and XML, something that's been clear to anyone involved in Web development for at least a couple years. Why bring it up again, and why talk about RSS and hypertest instead of what they represent?
2003-04-17 12:58:52
This is insightful?
>This "insight" more accurately refers to the difference between HTML and XML

No; there are other hypertext markup languages (less popular, obviously) than HTML, and I was referring to all of them in general. Also, when I said that "RSS is a way to mark up a class of links for a particular domain," that can't be applied to XML. XML is more generalized than that, and can be used for presentation-specific metadata as well as presentation-independent metadata--compare DocBook with XSL-FO.

The main points of of my posting are not necessarily as obvious to everyone out there as you might think. (I was responding to an asked question, after all, from a guy with long, serious experience in XML technology.) Take a look at the thread which I was responding to. It begins at