- From: Reto Bachmann-Gmür <reto@gmuer.ch>
- Date: Tue, 10 Jan 2006 11:42:57 +0100
- To: "siebeneicher@oaklett.org" <siebeneicher@oaklett.org>
- CC: semantic-web@w3.org
siebeneicher@oaklett.org wrote: > > Reto Bachmann-Gmür schrieb: >> <SiteMapEntry> >> <rdfs:label>Flight - Asylum - Integration</rdfs:label> >> <page rdf:resource="http://www.osar.ch/education" /> >> <children rdf:parseType="Collection"> >> <SiteMapEntry> >> [...] >> </SiteMapEntry> >> </children> >> </SiteMapEntry> > > I am not introduced to the need of this example, could you tell me > what should it solve, or better asked, in what matter could users > "act" with a software that is using this example. Is it the way > Navibars tree do it? A simple tree, once all branches are opened by > the user, he has a look at the whole tree hierachy and can choose the > webpage he wants. The tree is presented to the user allways in the > same structure until the source changes. Let me say, the structure of > a tree is like a graph where the rule is, that a node anywhere in the > whole graph, can only have one parent, but more than once children. My hope is that a nice application like navibar could foster the deployment of content making the semantic web grow. As requirement to format I would see: - free merging of navigation graphs - multiple navigations for the same set of pages - easy to understand and use > > During the design phase of the NNS Sitemap i decide to use a simple > and "flat" tree representation only for practical reasons. I also > thought of a navigation tool going a more semantic way. But hey... > what damn does the "semantic way" mean? Today, i think this way is > when users ask the internet a clever question and get a clever answer. Imho, things that are not the semantic web way in http://navibar.oaklett.org/sitemap.rdf are: - Lack of type for the classes - Use of rdf:Seq which makes aggregation harder than with parseType="Collection" - Using plain-literal URIs while in fact the named-resource is meant - Unnecessary naming of resources with URNs In my example: - The resources have a type (SiteMapEntry, but would probably better be "NavigationEntry") - Instead of "identifier" a "page" property points to the page (which is a resource) - parseType="Collection" instead of Seq - rdfs:label / dc:title: question of taste I guess, could live with dc:title ;-) Like your vocabulary it supports an arbitrary graph-structures including circles, and I don't think that it would be any harder to use for an application like the current version of navibar, but more likely to be of use for more generic aggregators. > > This means practicaly, the navigation tool offers the user a more > fluid structure of the content of a website. A rule is, that the > structure of the view is changable because the users view -- depending > on the question the user asks -- is also changable. If the user is > "asking" a soccer website the question: "worldcup 2006", the structure > of the view of webpages could be: > .... Another use-case, where aggregability and graph structure become important would be a tool (navibar 2 ?) to visualize an navigate through a blog-discussion on multiple servers. reto
Received on Tuesday, 10 January 2006 10:43:23 UTC