This page is a proposal to build semantic linking into wiki extending a wiki to be a SemanticNetwork.
Social tagging sites are quite successfull using arbitrary words for tagging. IMO this does not scale. One very important thing about tags is that they should be suitable to allow searching by narrowing space. → One have to have a directed network of tags (SemanticNetwork).
A Wiki with its pages and links is already a network of tags, but missing direction and different releations. In a wiki you can only tell: pageA links to pageB.
To make a semantic network out of this, the links should get names, thus becoming relations.
pageA is_a_subconcept_of pageB pageC is_a_kind_of page
One possible solution to do this is to give names to relations in a wiki, by using a kind of InetWiki? linking to the same wiki:
This Page is a is:proposal to build t:semantics into t:wiki.
where “is” is the name of the relation “is a kind of”, and “t” is the relation “talks about topic”.
In the rendered output the relation-names (is, t) could be added as relation attributes
This page is <a href="?id=proposal" rel="is">proposal</a> ...
Is someone aware of experiments about semantic linking in Wikis ?
Karlsruhe [[:is located in::Germany]].
But, it doesn’t show up in the rendered output as a rel.
You can see an example of real use: Abraham, on the Bible wiki. You can see that they’ve noted the children of Abraham, and the semantics system recognizes it, and lets you perform queries using that information.
One thing I’ve done for right now is to create a SemediaWiki? site at http://socialsynergyweb.com/semanticwiki/index.php?title=Main_Page
I’d prefer not to use Media Wiki, but this is the only Semantic Wiki engien that my hosting can run right now. What I plan on doing is to transfer my open business wiki activities from http://socialsynergyweb.com/wikispace/doku.php, to http://socialsynergyweb.com/semanticwiki/index.php?title=Main_Page either temporarily or permanently.
I have a kind of loose theory that if I start experimenting with the ideas put forth by hansWobbe about TransClusion, and MicroContribution, but also including semantic annotation, I’ll be able able to create something worth copying to a Communityrepository.
The idea here is that MicroContribution is based around Semantic elements, and that content is re-useable in the form of RDF feed content. I want to see how difficult this is in practice with current SemanticWiki technology. I have a theory that semantic content will be the core basis for the ideas/application that we talked about at CommunityRepository and CommunityWikiBusinessDiscussion (those ideas to be refactored here, or somewhere. Still meditating on that).
In the late 1980s & early 1990s, Wikipedia:Expert_system(s) were quite the rage. Since I got a Corporate “penalty box” assignment at that time, I had about 18 months to explore their intricacies and was quite impressed by the relative advantages an Wikipedia:Inference_engine offers over the traditional Hierarchical approach (all be it, only) in specific situations. I don’t know enough yet to know if your proposal to meld the current (fad?) SemanticProgramming? with MicroContent concepts, is one of these, but I’m prepared to spend a bit of time investigating this.
In the process, I may also be able to generate a bit of useful content for your site, if you are interested.
Hans, your insights about this are most helpful, and please do not hesitate to experiment to your heart’s content on the SemanticWiki site. I know that KnowledgeManagement was a big thing back in the 1990’s, although it seems to have ground to a halt, and the semantic web seems to have reached a kind of limbo (except for the new growth in AdHoc? MicroFormats).
One of the interesting things to me, that seems to relate to what you’ve been doing, Hans, is that the content in SemanticWiki produces unique URI’s (as discussed by ChrisPurcell and others in ResourceDescriptionFramework). And, these URI’s are connected to statements about meaning. I’ve noted that your (HansWobbe) MicroBlocks? of content seem to be based around statements about things, at least in a way. So, SemanticWiki can work for both traditional article and essay or “page style content, and for TransClusions? of connected (or networked) MicroBlocks? of content. That’s my theory, anyway. I also believe that content can be recombined by TransClusion that is based on RDF and/or XML. Right now, though, I just want to immerse myself in the actual usage of SemanticWiki to get a feel for it.
In the very short time that I’ve used SemanticWiki, it’s opened my eyes to some different possibilities, and I am laying some of those out on this page:
Another very reasoned argument against some elements of the current MicroFormats approach.
I was very enthusiastic about MicroFormats, and still am, but once I started exploring the power of RDF, I began to wonder why MicroFormats emerged when RDF existed. Looking even deeper into the mechanics of the two standards, I see some real problems for the scalability and long-term decentralized use of MicroFormats on the web.
Yet, I do see the utility that MicroFormats serves as a replicable standard that sites can employ for making data exportable.
I predict that, mostly because quite a few RDF proponents are now working on MicroFormats, that these issues will likely disappear over time, and the two projects will merge on some level, to take advantage of the positive benefits of each idea.
I meant to respond to this two months ago, when you originally wrote. I thought that I did, but it must have just been some text lost in a buffer somewhere…
I track the SemanticWeb community, and I disagree very strongly; That was true for a couple of years there, but in this last, say, year, they’ve really re-crystallized their vision, and they’ve been doing an amazing amount of data interconnection.
They’ve realized that they are about NetworkedData?, rather than AI theory, and they’ve made that not only the focus of their message, but also the focus of their activity.
Let me dig up some of the names of the projects…
Yes, I now see the error of my thinking there, but only after spending a couple of months studying current SemanticWeb activity. There is actually a lot going on outside of MicroFormas?. The more I learn about SemanticWeb, the more interested I am in it, and NetworkedData?, even more so personally than MicroFormats.
Also, I am becoming more and more interested in thinking about how OneBigSoup-type NetworkedData? might be enabled by different aspects of SemanticWiki, at least between wiki communities. This is something I am exploring with others at http://semanticwiki.socialsynergyweb.com
I also looking at extensions like http://www.oddmuse.org/cgi-bin/wiki/LinkTagMap_Module and thinking about how something like this could be adapted into a full-blown Semantic extension for OddMuse. Perhaps with the aid of CPAN modules? Thinking similar things about MoinMoin, and other commonly used WikiEngine(s). If they can all output SemanticWeb data, there could be a toolset, like http://simile.mit.edu/ that could be used to navigate the the data. I am looking at this mostly from the perspective of CommunityToCommunity? cooperation. Also, beyond wiki, it’d be neat to see Semantic annotation in tools like Drupal, Moodle, ELGG, etc.
It is interesting to come back and read what I was talking about and thinking over a year ago. To update: I have been following http://linkeddata.org/ for about a year now, and I can agree with Lion’s comment above that there is a vibrant SemanticWeb community.
Sam, Can you provide a few hints regarding those aspects that you’ve found most interesting?
Hans, one of the ideas I’ve found to be compelling is related to sourcing ingredients for food recipes, using semantic markup. I will dig up a link, there are some folks in Australia exploring this. I read about this on http://linkeddata.org/ email list. I am also interested in the idea of using xmlhttprequest and semantic data on small devices that are part of environment. Or, even placing info like this into chips that are part of machines or devices. The data can report about what the machine is made of, where to get parts to make or repair the machines, where to download plans to build the machine, where to download plans to make machines that allow you to make the machine, etc etc.
Sam, I really would appreciate any further hints you might be able to add since I’m having a bit of trouble thinking “outside of the box” (I am currently in) on this SemanticWeb stuff. Perhaps that I sign that I’ve been spending far too much time focusing on the uses on Tags and their related NameConventions? lately, but so far, I seem to be able to achieve most of the touted benefits by simple using Tags to create “context(s)” within which I can then recognize some specific insights I am interested in. On the assumption that I’ve just drilled down into this particular “rut”, I’d like to see if anyone has some different perspectives that I can use to alter my current perspectives.
Hans, I think that the examples that I give are actually just some uses for SemanticWeb standards that people discovered after the fact, that are not necessarily connected to everything that was “promised” from the SemanticWeb back in 2001 in his Scientific American article.
Particularly, http://linkeddata.org/ is an example, of practical uses. I have an assumption, on a fundamental level, that if you connect objects into a network, that they will become more valuable than if you do not connect them. Or, even if design something to have the potential to connect. SemanticWeb standards like http://www.w3.org/DesignIssues/LinkedData.html are one way to do this. (see also http://linkeddata.org/guides-and-tutorials ).
There could be a different standard, or an interoperability of as many standards as possible. The standard to me is not as important as the underlying concept of linking and affordance to create machine-usable networks from recognizable objects.