[protege-discussion] protege-discussion Digest, Vol 18, Issue 14

[Thomas Russ]

On Jan 17, 2008, at 11:01 AM, hhuang wrote:

> A comment and another question:
>
> Thanks for the response.  The "trick," to me, was having to click the
> "save as gif" option prior to clicking the "create graph" button.  I
> did it the other way around yesterday.

Ah, yes.  The file is only created when the graph is drawn.

>  As far as whether the
> eventual dot program input is a Word file, it is shown on my Window
> XP desktop as a Word file.

Well, Windows XP is wrong. ;)

Try opening it with Word and see what you get.  (Well, it might even  
work, since it is just a text file)

> What I ended up doing yesterday was to re-save the Word file as .txt
> and use a DOS command to draw the graph (as another one of you
> advised), and save the output as a .pdf file.  The advantage of .pdf
> seems to be that I can easily print a large model on multiple pages,
> whereas the printout for a .gif file seems to be squeezing everything
> on one page that became not legible.

You should be able to do the DOS command without renaming it.  The  
dot.exe program doesn't really care what the file extension is, so  
long as it is a valid dot input file.

> Now my next question:
>
> I noticed that when a Slot of the Instance type is created and
> multiple classes are allowed, those classes are in the OR
> relationship.  For example, I created three classes: Car,
> Transmission, and Engine and a Slot called Contains.  I, then
> included the Transmission and Engine classes in the "Allowed Classes"
> pane in the Contains Slot form.  What I get would be:  a Car Contains
> Instances of Engine Or Transmission.
> Any advice, other than creating two individual slots?

Well, that's kind of what you want isn't it?  Your contains slot for  
Car is allowed to have fillers that come from either the engine or  
transmission class.  Now, you can't attach useful cardinality  
restrictions to that without creating individual slots, which you  
could make subslots of a more general contains slot.  But you would  
have to be careful to make assertions only to the more specific slots  
in order for the cardinality constraints to be useful.


Also, I would highly recommend that you look at
   Morton E. Winston, Roger Chaffin, Douglas Herrmann: A Taxonomy of  
Part-Whole Relations.Cognitive Science 11(4): 417-444 (1987)
for some guidance on different types of part-of slots and relations.