[Attempto] Cannot convert ACE Text to OWL ontology!

[Kaarel Kaljurand]

Hi,

On Wed, May 11, 2011 at 09:03, henry fonbeyin <fonbeyin at yahoo.com> wrote:
>
> I am exploring the use of ACE View. The first thing I did was to develop a
> trial ontology using the Classes, Properties and Individual tabs of the ACE
> View editor in Protege 4.1. I then generated the ACE text as shown in Fig. 1
> attached.  Now in my second experiment I want to develop my ontology purely
> from using the ACE text generated in my first exercise. I now copy the ACE
> text and create a new ontology by pasting it into the ACE Text pane. When I
> run "Update", the ontology hierarchy that was used in generating the ACE
> text is not obtained as can be seen in Fig 2. Why is it so?

my guess is that the copy-pasted ACE snippets failed to parse and
generate any OWL axioms. You can check the "Non OWL/SWRL snippets"
metric to see if you have any unparsed snippets.

The basic reason here is that an ACE text itself does not contain all
the information that is necessary to create an OWL ontology. What is
missing are the lexicon entries: mapping of words to IRIs, the fact
that the plural of "man" is "men", etc.

Now, when you add an axiom in the standard Protege view, e.g. you
assert that "SubClassOf(NGO, Organisation)", then two types of "ACE
views" are updated:

1. ACE text/snippets/etc now contain: Every NGO is an Organisation.
2. ACE lexicon view now contains the registry of all the content words
(e.g. Organisation) and their morphological mappings, e.g. singular
Organisation -> plural Organisations

When you now only copy "Every NGO is an Organisation.", relaunch
Protege and try to parse this sentence again, then this fails because
ACE View has no idea what type of word e.g. "Organisation" is, because
the lexicon is empty.

ACE View can guess the nature of the unknown words. See the checkboxes
"Use Clex" and "Guess unknown words" in the preferences. I think these
are switched on by default, but they provide only a partial solution,
e.g. they don't help with sentences that contain capitalized nouns
(which is the case with your sentences). In general, I would recommend
switching these checkboxes off to get a clear cut situation: a word is
either defined in the lexicon or undefined (in which case it should be
defined).

While you can save the ACE text by copy-pasting it from the ACE Text
view into a file, you cannot do the same with the lexicon. The only
way to save all your work is to save it as OWL: the lexicon will be
saved into annotation property axioms and the ACE snippets will be
saved into axiom annotation axioms. For example, if you first load the
following ontology (in Manchester Syntax) which just contains the word
annotation, then you can parse the sentence "Every NGO is an
Organisation."

--cut--
Prefix: : <http://www.w3.org/2002/07/owl#>
Prefix: rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>

Ontology: <http://www.semanticweb.org/ontologies/2011/4/Ontology1305199287183.owl>

AnnotationProperty: <http://attempto.ifi.uzh.ch/ace_lexicon#CN_sg>
AnnotationProperty: <http://attempto.ifi.uzh.ch/ace_lexicon#CN_pl>

Datatype: rdf:PlainLiteral

Class: <http://www.semanticweb.org/ontologies/2011/4/Ontology1305199287183.owl#Organisation>

	Annotations:
		<http://attempto.ifi.uzh.ch/ace_lexicon#CN_sg> "Organisation",
		<http://attempto.ifi.uzh.ch/ace_lexicon#CN_pl> "Organisations"


Class: <http://www.semanticweb.org/ontologies/2011/4/Ontology1305199287183.owl#NGO>

	Annotations:
		<http://attempto.ifi.uzh.ch/ace_lexicon#CN_sg> "NGO",
		<http://attempto.ifi.uzh.ch/ace_lexicon#CN_pl> "NGOs"
--cut--

Hope this helps,
Kaarel