Some times ago I’ve seen an interesting project, called Co-link

This project is about the possibility to edit and insert a link where the user wants in the page that he is reading on the Web.

This technology was developed with the objective of letting any interactant create new associative links in a pre-existent text and/or add new destinations to a previous link.

The idea is nice: the user can insert in any esistent word in the text a link and also insert a new link in a pre-esistent link…

Another idea is this: multilink

A multilink, or mlink, is a link that goes to one or more destinations.
A multilink is rendered as a popup menu that presents the targets as a list of menu items.
An item can be selected from the menu to follow the URL of the target, or a special item called “edit” can be selected to add, remove, or modify the targets of the multilink.
Targets are not embedded in the markup, but rather, stored in an external metafile linked to but exists independently from the page where the multilinks are defined.

Now the idea is not totally bad, but IMHO it’s dangerous…

There is an interesting discussion on Slashdot on it:

-> Multiple-Target Hyperlinks for the Masses

In fact when you find useful a tool like this, you have to understand that you include a semantic effort in the link concept…without using a standard way to make this semantic effort understandable by machines..

We can do it using RDF…

The RDF hyper-link scenario

We can have a multiple-link using an RDF model, upon its “open world“ model…

Using the property rdf:SeeAlso

There is a wonderful presentation of hyperlinks in RDF:

-> RDF Hyper-linking

This last statement, ‘seeAlso’ gets us from Semantics to Semantic Web.
The property rdfs:seeAlso is a relationship between something and a document which describes it further. This is all we need to link RDF files into a Web.

And with some technical stuff:

  • provenance: RDF stores need to keep track of where they found these files
    • open world: descriptions are scattered, incomplete, partial
    • identity reasoning: tools need to automatically figure out when two files talk about the same thing

So using this background we can do some identical stuff, but using web standards and semantically correct…

And this is the summary:

W3C’s RDF Schema Specification defines a relationship type called ‘seeAlso’

It relates things to other things that describe them (typically RDF/XML documents)

By traversing these references, we can re-discover Web crawling in the RDF world

It is entirely domain-neutral; while we prototyped this for FOAF people-descriptions, it works for companies, documents, images, …

RDF’s graph data model is designed for information merging; hyper-links are a great way of finding information to merge.

If your RDF documents tell just part of the story, add some seeAlso references to further reading.

This is a correct direction to resolve the issue IMHO…

Maybe in the future some code stuff about this solution…


-> Co-link e interazione nel Web
-> Multiple-Target Hyperlinks for the Masses
-> About Multilinks
-> RDF Hyper-linking