Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[p4-feedback] Which ontology does Protege select for import when multiple ontologies with the same URI are present?

Patrick McCrae patrick.mccrae at informatik.uni-hamburg.de
Fri Apr 17 00:54:41 PDT 2009


Thanks for your help on this, Nick!

> I have a feeling there are some flaws with this "peek" so please send 
> me an example that breaks it if you can.
I have been observing that problems arise when multiple versions of an 
import ontology, all with the same URI, are present in the same 
directory (as would be expected from your account of the import 
procedure). Using 'Save As' I have created multiple versions of the same 
import T-Box, each one containing slight modifications. When I then open 
the importing A-Box it seems impossible to predict which of the 
different T-Box versions is actually going to get imported.

Can you confirm this oberservation? (Maybe there is some kind of hashing 
involved which scrambles up file sequences and hence defies prediction?)

Best regards -

Pat
_______________________

Patrick McCrae 
<http://cinacs.informatik.uni-hamburg.de/index.php?option=com_content&task=view&id=17&Itemid=85> 


CINACS Graduate Research Group
Department of Informatics
Hamburg University
Vogt-Kölln-Straße 30
22527 Hamburg, Germany

patrick.mccrae at informatik.uni-hamburg.de
fon: +49 . 40 . 428 83 - 23 60
fax: +49 . 40 . 428 83 - 25 15


Nick Drummond wrote:
> Hi Pat,
>
> Across sessions p4 doesn't remember where it loaded imported 
> ontologies from when they were initially added or subsequently loaded, 
> but always applies the same strategy.
> I believe p4 tries these locations in the following order:
> - any directory from which an ontology has been explicitly loaded
> - ontology repositories
> - the web
> - it will ask the user
>
> If you are loading an ontology from a local file protege will 
> automatically search the containing directory to resolve imports.
> The algorithm peeks inside each potential owl file to see if it can 
> find a URI that matches the one in the import statement.
> I have a feeling there are some flaws with this "peek" so please send 
> me an example that breaks it if you can.
>
> Nick
>
> 2009/4/7 Patrick McCrae <patrick.mccrae at informatik.uni-hamburg.de 
> <mailto:patrick.mccrae at informatik.uni-hamburg.de>>
>
>     Hi, Nick -
>
>     Could you please briefly outline what the general mechanism is
>     that Protege 4.0 uses to link file names and URIs and how its
>     behaviour depends on the system setup?
>
>     I still believe it is Protege's mechanism of linking these two
>     ontology attributes which have also been causing my import
>     problems (my current suspicion is that the link from URI to file
>     always remains unchanged - regardless of whether I create a new
>     ontology with the same URI but a different file name. Hence always
>     the old ontology version gets loaded even if newer versions are
>     available locally. Would this make sense?).
>
>     Best regards -
>
>     Pat
>     _______________________
>
>     Patrick McCrae
>     <http://cinacs.informatik.uni-hamburg.de/index.php?option=com_content&task=view&id=17&Itemid=85>
>
>
>     CINACS Graduate Research Group
>     Department of Informatics
>     Hamburg University
>     Vogt-Kölln-Straße 30
>     22527 Hamburg, Germany
>
>     patrick.mccrae at informatik.uni-hamburg.de
>     <mailto:patrick.mccrae at informatik.uni-hamburg.de>
>     fon: +49 . 40 . 428 83 - 23 60
>     fax: +49 . 40 . 428 83 - 25 15
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/p4-feedback/attachments/20090417/39b4127d/attachment.html>


More information about the p4-feedback mailing list