Hypertext Knowledge Workbench
|Hypertext Knowledge Workbench|
The Hypertext Knowledge Workbench (HKW), is an editor and browser for semantic personal knowledge models. The tool is designed to be used by a single person to manage her personal notes about any topic that seems relevant.
Screenshot from July 2009
How To Use
- All buttons have tool tips - just fly with your cursor over one and wait, to see what it does
- The main item is always shown in the middle (larger font)
- The small icon shows what kind of item it is
- N = NameItem
- blank = ContentItem
- purple arrow = Relation
- Here you can rename the relation and its inverse relation
- three little things = Statement
- You can always go to a statement by clicking the little black arrow
- On mouse-over the icon, HKW shows you the author of the item
- In the adress-bar of HKW you can go to existing or create new NameItems.
- The blue plus icons allow you to create a linked new item. If your central item is "A" and you click on the blue plus behind [has detail], you are creating a new item (let's call it "B") plus the statement (A, has detail, B). This allows you to quickly created an interlinked set of personal notes.
- You can always use the browser back and forward buttons
- Restart your browser sometimes, otherwise it keeps using more and more memory
In the statement view, you can select another source, relation or target for your statement. Note: You cannot rename a relation or item here, just select or create a new one. To rename things, first go to them.
The light gray items ('project' in the screenshot) are inferred, that is automatically calculated items. Sub-types and types are considered for the CDS inference engine. It works based on simple rules. The two most important rules are:
If (Peter, lives in, Karlsruhe) and (lives in, is subtype of, has context), then (Peter, has context, Karlsruhe). The general rule is (s,r,t) and (r, is subtype of, p) => (s,p,t).
Furthermore, if (How To Repair Coffemachines, has type, book) and (book, is subtype of, project), then How To Repair Coffemachines has also the type project (as in our example).
Note that HKW tries to minimize the display of inferred items. If there are several inferred triples (s,r,t) and (s,p,t) it shows only the most specific ones, if the other ones follow from it. QuiKey allows to query based on all inferred triples.
If HKW hangs
- Presse the white refresh-button in HKW, the one with an arrow on it. This will reload all data from the server and re-render the complete screen.
- Simply closing hte browser completely and re-opening it often works, too.
Tipp Make sure to completely close your browser often. Firefox is keeping a lot of memory after each open-close. So Firefox gets bigger and bigger in your memory. On average, a memory consumption of 110 MB for loading HKW in Firefox seems normal.
How To Report Bugs
- Please report any bugs/issues/ideas to mailto:firstname.lastname@example.org
- Make a screenshot. On windows: Ctrl-Print, then paste into another application. In some graph.css applications you can paste the clipboard image right into a new file.
- Tell us:
- What have you done
- What did you see (the screen shot helps here)
- What did you expect to see
Bug or Feature?
HKW loads incrementally, so or bigger knowledge models, you can see each part coming in, step by step.
- Demo: http://cds.fzi.de/hkw-20090505 -- very old, the version bundles in CDS Tools is much more recent.
- editing in wikisyntax
- is written in Java and GWT
- Max Völkel
- Subject Attribution: page (any URI)
- Subject Granularity: document
- Representation destinction: false
- Terminology reuse: false
- Object type: page
- Context: true
- Import RDF: true
- Import RDFS: true
- Export RDF: true