WikiVersusSemanticWeb

WikiWikiWeb, while true to its name (“fast” in Hawai’ian) and convenient for informal exchange, occupies a low rung on the ladder of digital-document semantic richness. In WikiWikiWeb, HTML (hardly the example of deep semantics from the start) undergoes a flaying that leaves for the most part a presentational shell of a document (a reverse-flaying?).

This process seems to be marching directly away from the SemanticWeb. (really? how so?) And WikiWikiWeb is growing more popular by the day (as of 2005).

Yes, yes, in the end it’s all about people and wikis connect people. But is there a way to reconcile WikiWikiWeb and the SemanticWeb? Can we offer thorough document-editing facilities and encourage rich markup while maintaining the ease that wikis provide?

Maybe the answer to all of this is that we need better user agents. WikiWikiWeb is a kludge, designed around the popular user agents of the 1990s. Yet, one observes with frustration, many of the limitations of those user agents haunt us today (2005). Amaya shows the path to victory, but most Web users lack software with Amaya’s capabilities.

So.

What do you think?

-- EtanWexler

HTML is no longer a human-editable data format, at least in the form commonly used on the Web (just try reading the source of just about any web page). In Wiki markup, many parts of HTML that apparently weren't so necessary after all for describing the actual content have been stripped away. What markup is used is usually used well, and it's much easier to "screen-scrape" information from Wiki content than from a random pseudo-HTML page. This is a step closer to the Semantic Web, by cleaning the slate first. It's not even easy to convince people of the importance of using tags well when they rarely or never see them.

Wiki is a user interface, even if an old one (basically a plain text editor). LaTeX lovers would tell you that text should be mostly plain text, with sparse and well-thought-out markup (not some verbose and inane garbage produced by an automaton). If you introduce new markup to Wiki that people see the usefulness of, they will naturally start adding that to both existing and new pages. If you can't convince people of the usefulness of some tags, maybe they are not all that useful at that moment. Providing useful tools, documentation and real-life examples would go a long way. You could start by introducing extra tags here, generating RDF from them and actually doing something useful with that RDF. See also http://www.emacswiki.org/cw/DublinCoreForWiki. (Maybe you do have such stuff, I don't know, I just happened to walk by. Linking it here would be natural, given that this page is named WikiVersusSemanticWeb.)　(Addition: At least there's something rudimentary at SemanticWiki.)

It seems like you are advocating editors that allow you to edit formats that are basically non-human-readable due to looking like HTML, RDF/XML, etc. look like. However, XML (and because of that RDF, too) is criticized of being unnecessarily verbose and unreadable by humans. See e.g. http://www.yaml.org and http://docutils.sourceforge.net/rst.html. Quoting from http://c2.com/cgi/wiki?YamlAintMarkupLanguage: "YAML has a lot of the same appeal for me as Wiki does. Comparing YAML to XML is roughly like comparing the WikiWiki MarkupLanguage to HTML."

People want the basic format itself to be one UI, apparently a GUI is no longer enough these days. I suppose this is because "normal" people (amateurs) are no longer mere consumers but also producers. They can't (or can but won't) take everything the pros will. I think this is also the reason why MUDs are alive, even though they are in many ways as primitive as games from the 80's. Consumers who also produce see things in a very different way. Many would also argue that human-unreadable formats are ugly kludges that exist only because they were either designed by a committee, or something had to be implemented and the implementer didn't quite have the time to come up with an elegant design.

-- A visitor